2002, Vol.20, Issues 4, Melanoma

Dermatol Clin 20 (2002) xiii xiv
Dedication
A special thanks to Alfred W. Kopf, MD
Alfred W. Kopf, MD
Fifty years ago, a young physician came to the Skin and Cancer Clinic in New York City. During his dermatology residency training, he became interested in the disease that is still lethal in dermatology: malignant melanoma. At that time, very little was known about cause, risk factors, prognosis, or treatment. Less than 50% of persons who were diagnosed with melanoma survived 5 years. But Alfred W. Kopf, MD, was committed to making a change. Over this past half-century, perhaps no one in dermatology is as widely known both nationally and internationally for his accomplishments in this area. Although a complete list of these achievements is far beyond the scope of this dedication, some are so dramatic and important that it is remiss not to review them:
Establishment of the Malignant Melanoma
Cancer Foundation Malignant Skin Cancer Screening programs. Improvement of melanoma primary prevention programs in the public education arena in conjunction with the American Academy of Dermatology and the Skin Cancer Foundation. Training of over 20 Fellows and numerous Residents and medical students. Over 400 papers, chapters, and textbooks have been written as a direct result of these teaching activities, and many of these have changed the clinical approach to melanoma patients. Perhaps Dr Kopfs most important professional achievement, however, is the universal respect of his colleagues in medicine. His contemporaries hold him in the highest regard. He has been the President of the American Academy of Dermatology, American Board of Dermatology, and American Dermatological Association. His efforts in extending quality dermatologic care to third-world countries through the International Foundation for Dermatology have been unequalled. The American Academy of Dermatology has recognized his efforts with the Gold Medal, and he has received a Lifetime Achievement Award from the World Health Organization. Dr Kopf also has worked continuously toward enhancing the future. Those whom he has trained
Cooperative Group (NYU, Harvard University, University of California, San Francisco, and University of Pennsylvania). Development of one of the earliest melanoma patient clinical databases to assess the effects of multiple variables in survival. Focus on reduction of morbidity and mortality through early screening prevention (the development of the ABCDs of early melanomas) and his assistance in the development of the American Academy of Dermatology and Skin
0733-8635/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 0 7 3 3 - 8 6 3 5 ( 0 2 ) 0 0 0 3 9 - 6
xiv
Dedication / Dermatol Clin 20 (2002) xiiixiv
now will train future generations. His students universally view him as a role model; Alfred W. Kopf personifies the meaning of mentor. In recognition of all of his achievements, all of the authors of this issue of the Dermatologic Clinics dedicate our efforts to Alfred W. Kopf, MD.
Darrell S. Rigel, MD Ronald O. Perelman Department of Dermatology New York University Medical Center 35 East 35th Street, Suite 208 New York, NY 10016, USA E-mail address: dsrigel@prodigy.net
Dermatol Clin 20 (2002) xv
Preface
Melanoma and pigmented lesions
Darrell S. Rigel, MD Guest Editor
Malignant melanoma is one of the most dangerous cancers seen by physicians. Sadly, one American dies every hour from this disease. The natural biology of melanoma offers unique opportunities successfully to battle this cancer and to make inroads into the associated mortality rate. The cause of most cases of melanoma is known: excessive exposure to ultraviolet radiation. In addition, melanoma is one of the most clear-cut cases of a cancer where early detection and treatment are the keys. Patients with in situ melanomas may have almost a 100% chance of survival, whereas those with advanced tumors may have no effective therapy available. For these reasons, factors related to both primary and secondary prevention of melanoma are of critical importance. The scope of issues related to melanoma that physicians deal with on an emergency basis, however, go far beyond prevention. Advances in early diagnosis, better definitions of surgical margins, and newer approaches for advanced lesions including
sentinel node biopsy, immunotherapy, and the development of potential vaccines, make this one of the most dynamic areas of medicine. The rapidly rising incidence of melanoma worldwide makes it particularly important to develop advances in these aspects of therapy. This edition of the Dermatologic Clinics is focused on the full spectrum of the newest aspects of melanoma from epidemiology and prevention, through diagnosis, to emerging therapies. It is hoped that this issue will give the practicing physician a greater understanding of the challenges faced in dealing with cancer and enable physicians to deliver the best possible care to melanoma patients. Darrell S. Rigel, MD Department of Dermatology New York University School of Medicine 350 5th Avenue, #7805 New York, NY 10118, USA E-mail address: dsrigel@prodigy.net
Dermatol Clin 20 (2002) 589 595
Melanoma incidence trends

Caroline Bevona, MDa, Arthur J. Sober, MDa,b,*
a
Department of Dermatology, Massachusetts General Hospital, 55 Fruit Street, Boston, MA 02114, USA b Harvard Medical School, 55 Fruit Street, Boston, MA 02114, USA
The incidence of melanoma has been increasing in most white populations throughout the world during much of the past century [1]. From the early 1960s to the mid 1980s, a 3% to 7% annual increase in agestandardized incidence was seen in many countries [2]. Melanoma mortality has also been increasing, but at a slower rate than has occurred for incidence [3]. Two factors that increase the risk of developing melanoma are (1) having a fair complexion, as seen in people of Celtic or Nordic origin, and (2) sun exposure [4 10]. These factors help to explain the varying patterns in the incidence of melanoma worldwide. Relatively high rates occur in the light-skinned populations of northern Europe, as well as Austria and Switzerland, whereas lower rates tend to be found among people indigenous to Asia, Africa, South America, and in darker-skinned whites of southern Europe (Fig. 1) [11]. The observation that incidence rates in many countries increase at latitudes closer to the equator supports a contributing role of sun exposure in the development of melanoma. Such gradients have been found within North America, England, Norway, Sweden, and New Zealand [6,12,13]. This relationship is not uniformly seen within all European countries, however, because confounding variables, such as skin color, socioeconomic status, and behavioral factors, also contribute to the risk of developing melanoma [12,14]. The observation that people living at higher latitudes in northern Europe have higher incidence rates than those of southern Europe can be explained by the protective darker skin color of the
* Corresponding author. Department of Dermatology, Massachusetts General Hospital, 55 Fruit Street, Boston, MA 02114. E-mail address: asober@partners.org (A.J. Sober).
latter [6,15]. The highest rates of melanoma in the world are found in Australia and New Zealand where light-skinned people, largely descending from northern Europeans, live close to the equator. Within these two countries, the incidence of melanoma has reached epidemic proportions and is the fourth most common registrable cancer among men and third among women. In Australia, it is also the leading cause of cancer in individuals aged 15 to 44. The age-adjusted incidence of melanoma (standardized to the Australian 1991 population) continues to increase in Australia, and in 1997 it reached a rate of 50 and 37 per 100,000 in men and women, respectively. The most recent statistics from 1999 from select states and territories suggest that the escalation of incidence in Australia as a whole has likely continued [16]. Although incidence and mortality rates have been rising for many decades, stabilizing or falling rates are starting to be seen among certain populations in some countries [17 22]. Because such trends may be concealed by examining age-standardized rates alone, both age-specific rates and an age-birth-cohort analysis have been used to gain a better understanding of age-related and generational trends. A cohort model examines age-specific incidence or mortality rates in relation to birth year and allows one to see how, at any given age, such rates vary between people of different generations. Melanoma incidence and mortality rates have been found to differ between cohorts, suggesting a difference in exposure to certain risk factors between the generations. These may include behavioral changes, such as amount of time spent in the sun, clothing styles, and differences in sunscreen use [23,24]. Stabilizing or declining incidence rates have been seen among certain segments of the population in Australia, New Zealand, Scotland, Canada, and the
590
C. Bevona, A.J. Sober / Dermatol Clin 20 (2002) 589595
Fig. 1. Age-adjusted incidence rates of melanoma in select countries. (Data from Parkin DM, Whelan SL, Ferlay J, et al. Cancer incidence in five continents, vol. 7. Lyon: International Agency for Research on Cancer, IARC Scientific Publications No. 143; 1997; with permission.)
591
United States [19,24 28]. Although the age-standardized incidence in Australia as a whole is continuing to rise, incidence rates have been falling in younger cohorts since the late 1980s. Rates continue to rise in older cohorts, in particular those over age 60 [16,25]. In New South Wales, Australias most populous state, trends in incidence were examined between 1983 and 1996. During this period, a significant decrease in incidence was seen among women ages 15 to 49 and a downward trend was found for men ages 15 to 34. A birth cohort effect was also seen with incidence rates decreasing in women and stabilizing in men born after 1950 [28]. It has been suggested that such trends among younger people may reflect improvements in primary prevention caused by the effectiveness of educational campaigns targeting skin cancer in Australia [28,29]. Alternately, it has been suggested that the recent immigration of relatively young, low-risk ethnic groups from the Pacific Islands, Asia, and the Middle East may be a contributing factor. Almost half of those immigrating from Asia to Australia have settled in New South Wales, and this group contains a higher proportion of women than men [30]. This may help explain the significant decline in incidence occurring among younger women. In New Zealand, incidence data from 1969 to 1993 showed a stabilization in rates for both men and women born after 1949 [23]. More recent data from 1995 to 1997 show an overall decline of about 20% in age-standardized incidence rates in men and women [31]. Interestingly, these data were obtained after implementation of the Cancer Registry Act of 1993, which was designed to improve data quality by making cancer reporting mandatory [32]. A sharp increase in incidence from 1993 to 1995 suggests the success of this new legislation. Whether this subsequent downward trend is meaningful or just transient remains to be determined in the years to come. In Scotland, incidence rates of women under age 65 have stabilized since 1986. Incidence rates in men continue to increase, although to a lesser extent in younger men [27]. In Canada, age-standardized incidence rates have reached a plateau in both men and women since the mid to late 1980s. When age-specific data were examined, a downward trend in incidence was seen in men under age 50 and women under age 60, with a significant decrease in 20- to 30-year-old women. A cohort effect was seen with a slight decline in incidence rates in women, and stabilization in men born after 1950 [33]. Similar to incidence, mortality rates have either leveled off or declined in the younger populations of Australia, New Zealand, Scotland, Canada, and the United States [23,27,33 35].
Trends in darker-skinned, low-risk groups worldwide are less well studied, because it is more difficult to detect meaningful patterns when the number of events is very low. Incidence rates, however, do seem to be stable or to show only slight increases in most of such populations [12,36,37]. One notable difference between whites and people native to Africa and Asia is the predilection of the latter to present with melanomas on the foot [36,38,39]. This may result in a delayed diagnosis and poorer prognosis. Recently, however, there have been studies showing improved survival rates in such populations, which have been attributed to earlier disease detection [40,41]. Although incidence rates have not reached the magnitude that they have in Australia and New Zealand, melanoma has become a growing health concern in the United States because of its dramatic increase over the past several decades. Although the incidence of most cancers has recently started to decline, that of melanoma continues to rise [42,43]. In the United States it is currently the sixth most common cancer in men, seventh in women, and comprises 3.5% of all malignancies [44,45]. It has been estimated that 51,400 new cases of invasive melanoma were diagnosed in 2001 [46]. Because the United States does not have a national cancer registry, incidence rates and trends have been estimated from data collected by the National Cancer Institutes Surveillance, Epidemiology and End Results (SEER) program founded in 1973. SEER data are collected from nine geographic regions that cover about 10% of the US population, and were expanded in 1992 to include two additional areas to increase minority representation [43]. Before 1973, incidence data can be found within tumor registries in individual states. Cancer data have been available in Connecticut from registries in local hospitals since 1935, and from the Connecticut Tumor Registry since 1941 [47]. The incidence of melanoma has been steadily increasing in Connecticut. In 1935 to 1939 the average age-adjusted incidence (standardized to the United States 1970 population) of around 1 per 100,000 had risen to about 20 in men and 14 in women by 1996 [47,48]. Although much of this growth is believed to be real, better registration practices account for at least a portion of the increase seen in the earlier years [49,50]. Consistent with trends seen in Connecticut, the SEER data have shown a steady increase in melanoma incidence in the white population (Fig. 2). The rate of increase, however, has begun to slow in both men and women. From 1973 to 1981, there had been a 6.1% annual increase in incidence, compared with 2.8% occurring from 1981 to 1998 [44]. When cohort trends were examined, rates in men increased in
592
Fig. 2. Age-adjusted incidence rates of melanoma in whites. (From Ries LAG, Eisner MP, Kosary CL, et al. SEER cancer statistics review, 1973 1998. Bethesda, MD: National Cancer Institute; 1998; with permission. Available at: http://seer.cancer.gov/ Publications/CSR1973_1998. Accessed October 25, 2001.)
successive cohorts born until 1950, after which they declined or stabilized. A different pattern was seen in women with rates flattening in cohorts born after 1950, but unexpectedly rising again in those born after 1960 [51]. It remains to be seen whether this increase will persist in the future as further data are collected. Age-adjusted mortality rates have continued to rise in men, but the rate of increase has slowed since the late 1980s. For women, they have remained stable since about 1979 [44]. When mortality was more closely examined, rates were shown to be increasing in older people, while stabilizing or falling among younger people [2,52 54]. Five-year survival rates have significantly increased in whites from 80% in 1974 to 1976 to 89% in 1992 to 1997 [45]. Consistent with trends seen worldwide, darkerskinned populations in the United States have a lower incidence of melanoma when compared with whites (Fig. 3). In 1998, melanoma incidence rates were 16 times lower in blacks than in whites. Furthermore, incidence and mortality rates of melanoma among blacks have been stable since 1973 except for a slight decrease in mortality seen among men [45]. Within the white population, the recent increase in the incidence of melanoma is largely, but not solely, caused by an increase in the number of thin lesions
diagnosed [35,42,51,55]. Since 1988, age-adjusted incidence rates for melanomas of every thickness had increased in all people, with the exception of thick tumors remaining stable among women. When age-specific incidence was examined, no significant increases were seen in men under age 40 and thick tumors had only increased in men over age 60. In women, an increase in thin tumors was found in all age groups [51]. When stage of disease was examined, most of the increase in incidence is accounted for by localized disease. Rates of localized melanoma have been sharply increasing in both men and women, with rates now beginning to stabilize for women. The incidence of regional and distant melanomas also has been increasing, but to a smaller extent [35,42,51]. Changes in the incidence of melanoma according to body site have been seen. Anatomically, the highest incidence of melanoma occurs on the trunk for men and lower leg for women. Data from 1973 to 1994 show that the largest increase by site has been seen on the trunk for both men and women. For women, the incidence of melanomas on the trunk is now equal to that of the upper extremity, which had previously been the second most common location [35]. When adjustments were made for body surface area, men born
593
Fig. 3. Average age-adjusted incidence rates of melanoma among different racial groups (1992 to 1998). (From Ries LAG, Eisner MP, Kosary CL, et al. SEER cancer statistics review, 1973 1998. Bethesda, MD: National Cancer Institute; 1998; with permission. Available at: http://seer.cancer.gov/Publications/CSR1973_1998. Accessed October 25, 2001.)
before about 1950 showed the greatest susceptibility of the head region, whereas men born later showed the trunk to be the most common site. For women of later cohorts, the incidence rates for the trunk equal that of the legs. When adjusted for body surface area, the arms are most susceptible followed by the legs with more recent cohorts showing the trunk as the second most common site. These increasing incidence rates of melanoma on the trunk in women may reflect clothing style changes in later cohorts that allow for more sun exposure of that area [56]. The question of whether the overall increase in the incidence of melanoma is caused by artifact or by a real increase in disease has been the subject of much debate [57 59]. If artifact, it may be explained by an increase in surveillance resulting from either early detection of lesions (lead-time bias) or the detection of clinically insignificant lesions (length bias) [42]. Although artifact might account for some of the rise in melanoma, evidence supports that a true increase in disease is occurring. This is supported by the fact that, in the United States, the increase predates educational campaigns targeting melanoma and such increases are also occurring in countries with little in the way of educational programs. Also, new cases are not just limited to thin lesions and localized disease, and there is no evidence that there has been a change in histologic criterion or improved counting methods, which would explain such an increase [51,60,61].
Despite the overall continuous increase in melanoma, it is apparent that different trends are occurring simultaneously among different subsets of the population. One important trend is the decrease or stabilization in the incidence and mortality rates among younger people within certain countries. This allows for some optimism in the future with respect to the morbidity and mortality of this disease for both patients and the health care system. Despite this, however, the total incidence of melanoma will continue to rise in the near future, because increasing rates are still seen in older people who comprise most of those afflicted with melanoma [56].
References
[1] Titus-Ernstoff L. An overview of the epidemiology of cutaneous melanoma. Clin Plast Surg 2000;27:305 16. [2] Armstrong BK, Kricker A. Cutaneous melanoma. Cancer Surv 1994;20:219 40. [3] MacKie RM. Incidence, risk factors and prevention of melanoma. Eur J Cancer 1998;34(suppl 3):S3 6. [4] Brozena SJ, Fenske NA, Perez IR. Epidemiology of malignant melanoma, worldwide incidence, and etiologic factors [review]. Semin Surg Oncol 1993;9: 165 7. [5] Crombie IK. Racial differences in melanoma incidence. Br J Cancer 1979;40:185 93. [6] Crombie IK. Variation of melanoma incidence with
594
C. Bevona, A.J. Sober / Dermatol Clin 20 (2002) 589595 latitude in North America and Europe. Br J Cancer 1979;40:774 81. Holman CD, Armstrong BK. Cutaneous malignant melanoma and indicators of total accumulated exposure to the sun: an analysis separating histogenetic types. J Natl Cancer Inst 1984;73:75 82. Lancaster HO, Nelson J. Sunlight as a cause of melanoma: a clinical survey. Med J Aust 1957;1:452 6. Weinstock MA. Issues in the epidemiology of melanoma. Hematol Oncol Clin North Am 1998;12:681 98. Weinstock MA. Overview of ultraviolet radiation and cancer: what is the link? How are we doing? Environ Health Perspect 1995;103(suppl 8):251 4. Parkin DM, Whelan SL, Ferlay J, et al. Cancer incidence in five continents, vol. 7. Lyon: International Agency for Research on Cancer, IARC Scientific Publications No. 143; 1997. Berwick M. Epidemiology: current trends, risk factors, and environmental concerns. In: Balch CM, Houghton AN, Sober AJ, et al, editors. Cutaneous melanoma. edition 3. St. Louis, MO: Quality Medical Pub.; 1998. p. 551 71. Builliard JL, Cox B, Elwood J. Latitude gradients in melanoma incidence and mortality in the non-Maori population of New Zealand. Cancer Causes Control 1994;5:234 40. Sober A, Lew RA, Koh H, et al. Melanoma/skin cancer update 1991. Dermatol Clin 1991;9:617 817. Fitzpatrick TB. Enigma of the pathogenesis of primary melanoma: changing incidence and mortality in Japan and the United States. J Invest Dermatol 1989; 92(suppl 5):234S 5S. Australian Institute of Health and Welfare (AIHW), Australasian Association of Cancer Registries (AARC). Cancer in Australia 1997: incidence and mortality data for 1997 and selected data for 1998 and 1999. AIHW cat. no. CAN 10. Canberra: AIHW (Cancer Series no. 15);2000. Boyle P, Maisonneuve P, Dore JF. Epidemiology of malignant melanoma. Br Med Bull 1995;51:523 47. Elwood JM. Recent developments in melanoma epidemiology, 1993. Melanoma Res 1993;3:149 56. Kricker A, Armstrong B. International trends in skin cancer, September, 1996. Available at: http:// home.vicnet.net.au/~menzies/krick.htm. Accessed October 25, 2001. MacKie RM. The epidemiology of melanoma in the West. Melanoma Res 1997;7(suppl 1):S1. Marks R. Two decades of the public health approach to skin cancer control in Australia: why, how and where are we now? Australas J Dermatol 1999;40:1 5. Severi G, Giles GG, Robertson C, et al. Mortality from cutaneous melanoma: evidence for contrasting trends between populations. Br J Cancer 2000;82:1887 91. Bulliard JL, Cox B. Cutaneous malignant melanoma in New Zealand: trends by anatomical site, 1969 1993. Int J Epidemiol 2000;29:416 23. Bulliard JL, Cox B, Semenciw R. Trends by anatomic site in the incidence of cutaneous malignant melanoma in Canada, 1969 93. Cancer Causes Control 1999;10: 407 16. Giles GG, Thursfield V. Trends in skin cancer in Australia, September, 1996. Available at: http://home. vicnet.et.au/~menzies/giles.htm. Accessed October 25, 2001. Hursthouse MW. Melanoma incidence and trends in the Nelson-Marlborough area of New Zealand. N Z Med J 1992;105:424 5. MacKie RM, Hole D, Hunter JA, et al. Cutaneous malignant melanoma in Scotland: incidence, survival, and mortality, 1979 94. The Scottish Melanoma Group. BMJ 1997;315:1117 21. Marrett LD, Nguyen HL, Armstrong BK. Trends in the incidence of cutaneous malignant melanoma in New South Wales, 1983 1996. Int J Cancer 2001;92: 457 62. Marks R. Epidemiology of melanoma. Clin Exp Dermatol 2000;25:459 63. Czarnecki D, Meehan CJ. Is the incidence of malignant melanoma decreasing in young Australians? J Am Acad Dermatol 2000;42:672 4. New Zealand Health Information Services. Cancer: registrations and deaths. Wellington: Ministry of Health; 1997. Jones WO, Harman CR, Ng AKT, et al. Incidence of malignant melanoma in Auckland, New Zealand: highest rates in the world. World J Surg 1999;23:732 35. Gaudette LA, Gao RN. Changing trends in melanoma incidence and mortality. Health Rep 1998;10:29 41. Giles GG, Armstrong BK, Burton RC, et al. Has mortality from melanoma stopped rising in Australia? Analysis of trends between 1931 and 1994. BMJ 1996;312:1121 5. Hall HI, Miller DR, Rogers JD, et al. Update on the incidence and mortality from melanoma in the United States. J Am Acad Dermatol 1999;40:35 42. Swerdlow AJ. International trends in cutaneous melanoma. Ann N Y Acad Sci 1990;609:235 51. Yamamoto A. The epidemiology of melanoma in the East. Melanoma Res 1997;7(suppl 1):S1 2. Boisseau-Garsaud AM, Garsaud P, Ossondo M, et al. Acral melanoma in the French West Indies (Martinique) [letter]. Arch Dermatol 1998;134:112 3. Kato T, Suetake T, Sugiyama Y, et al. Epidemiology and prognosis of subungual melanoma in 34 Japanese patients. Br J Dermatol 1996;134:383 7. Crowley NJ, Dodge R, Vollmer RT, et al. Malignant melanoma in black Americans: a trend toward improved survival. Arch Surg 1991;126:1359 65. Kato T, Suetake T, Tabata N, et al. Epidemiology and prognosis of plantar melanoma in 62 Japanese patients over a 28-year period. Int J Dermatol 1999;38:515 19. Dennis LK. Analysis of the melanoma epidemic, both apparent and real: data from the 1973 through 1994 Surveillance, Epidemiology, and End Results program registry. Arch Dermatol 1999;135:275 80. Ries LA, Wingo PA, Miller DS, et al. The annual report to the nation on the status of cancer, 1973 1997, with
[7]
[25]
[8] [9] [10]
[26]
[27]
[11]
[28]
[12]
[29] [30]
[13]
[31]
[32]
[14] [15]
[33] [34]
[16]
[35]
[36] [37] [38]
[17] [18] [19]
[39]
[20] [21]
[40]
[41]
[22]
[42]
[23]
[24]
[43]
C. Bevona, A.J. Sober / Dermatol Clin 20 (2002) 589595 a special section on colorectal cancer. Cancer 2000; 88:2398 424. Howe HL, Wingo PA, Thun MJ, et al. Annual report to the nation on the status of cancer (1973 through 1998), featuring cancers with recent increasing trends. J Natl Cancer Inst 2001;93:824 42. Ries LAG, Eisner MP, Kosary CL, et al. SEER cancer statistics review, 1973 1998. Bethesda, MD: National Cancer Institute; 1998. Also available at: http://seer.cancer.gov/Publications/CSR1973_1998. Accessed October 25, 2001. Greenlee RT, Hill-Harmon MB, Murray T, et al. Cancer statistics, 2001. CA Cancer J Clin 2001;51:15 36. Polednak AP. Trends in cancer incidence in Connecticut, 1935 1991. Cancer 1994;74:2863 72. Polednak AP, Cancer incidence in Connecticut 1980 1996, October, 1999. Available at: http://www.state.ct.us/dph/OPPE/pdfs/cancrpt2.pdf. Accessed October 25, 2001. Polednak AP. Trends in cancer incidence in Connecticut. Conn Med 1997;61:211 8. Roush GC, Schymura MJ, Holford TR. Patterns of invasive melanoma in the Connecticut Tumor Registry: is the long-term increase real? Cancer 1988;61: 2586 95. Jemal A, Devesa SS, Hartge P, et al. Recent trends in cutaneous melanoma incidence among whites in the United States. J Natl Cancer Inst 2001;93:678 83. La Vecchia C, Lucchini F, Negri E, et al. Recent declines in worldwide mortality from cutaneous mel-
595
[44]
[53]
[54]
[45]
[55]
[46] [47] [48]
[56]
[57]
[58] [59]
[49] [50]
[60]
[51]
[61]
[52]
anoma in youth and middle age. Int J Cancer 1999;81: 62 6. Roush GC, McKay L, Holford TR. A reversal in the long-term increase in deaths attributable to malignant melanoma. Cancer 1992;69:1714 20. Scotto J, Pitcher H, Lee JA. Indications of future decreasing trends in skin-melanoma mortality among whites in the United States. Int J Cancer 1991;49: 490 7. Mikkilineni R, Weinstock MA. Epidemiology. In: Sober AJ, Haluska FG, editors. Skin cancer. Hamilton, Ontario: BC Decker Inc; 2001. p. 1 15. Dennis LK, White E, et al. Recent cohort trends in malignant melanoma by anatomic site in the United States. Cancer Causes Control 1993;4:93 100. Burton RC, Armstrong BK. Current melanoma epidemic: a nonmetastasizing form of melanoma? [review]. World J Surg 1995;19:330 3. Rees JL. The melanoma epidemic: reality and artefact. BMJ 1996;312:137 8. Swerlick RA, Chen S. The melanoma epidemic: is increased surveillance the solution or the problem? Arch Dermatol 1996;132:881 4. Rigel DS. Malignant melanoma: incidence issues and their effect on diagnosis and treatment in the 1990s. Mayo Clin Proc 1997;72:367 71. Rigel DS, Friedman RJ, Kopf AW. The incidence of malignant melanoma in the United States: issues as we approach the 21st century. J Am Acad Dermatol 1996; 34(5 pt 1):839 47.
Dermatol Clin 20 (2002) 597 600
Risk factors for the development of primary cutaneous malignant melanoma

Rona M. MacKie, CBE, MD, DSc
Department of Dermatology, University of Glasgow, G12 8QQ, UK
The1 value of recognizing significant risk factors for melanoma clearly lies in the subsequent identification of individuals who are at greater-than-average risk of melanoma, and thereafter considering directing both primary prevention and surveillance programs to this group. It is also important in this context to educate the public on the concept of relative risk. For example, in the United States and the United Kingdom, a womans lifetime risk of developing melanoma is around 1 in 90 but her risk of developing breast cancer is around 1 in 12. A doubling or trebling of melanoma risk still makes the chances of developing melanoma less than those of developing any one of the commonest three cancers in the developed world: breast, lung, and gastrointestinal tract. A unique feature of melanoma, however, is that visual inspection is the surveillance tool leading to earlier detection and a better prognosis. Visual inspection is inexpensive, requires no sophisticated imaging equipment, is easily taught, and is painless. In addition, a major cause of melanoma is excessive exposure to UV radiation in the form of natural
E-mail address: R.M.Mackie@clinmed.gla.ac.uk (R.M. MacKie). 1 It is a pleasure and an honor to be invited to contribute to this issue. Throughout my own period of professional interest in malignant melanoma, Al Kopf has been a mentor and a shining example of an individual working persistently, diligently, and productively in clinically based research and investigation. His friendly greeting at the registration desk of numerous conferences is always a welcome sign that the meeting will be both informative and enjoyable. May he and Dottie have many years to enjoy their numerous interests and the company of their family.
sunlight [1]. It has been calculated that inappropriate sun exposure is responsible for two thirds of all melanomas. Primary prevention activities in the form of advice about sensible sun exposure can be targeted appropriately. This contrasts with commoner malignancies, again using breast cancer for comparison, in that the major causes of breast cancer other than genetically inherited susceptibility are not yet well established and the surveillance tool of mammography is expensive, requires sophisticated equipment, and can create significant anxiety. Risk factors for melanoma have in the past been divided into genetic and environmental, but it is likely that this is a false division in that multifactorial genetic features determine response to environmental events, such as severe sunburn. This factor concentrates initially on genetic factors identified to date and thereafter on environmental factors, which may be influenced by the genetic background. The major genetic feature increasing melanoma susceptibility so far identified is the presence of mutations of the CDKN2A or p16 gene found on chromosome 9 in humans [2,3]. Around 5% of melanoma patients have a family history of one or more first-degree relatives also being affected, and of these around 25% have had mutations identified in the CDKN2A gene. The presence of a mutation is more frequent if three or more members of any family are affected, and only around 10% for families if only two affected family members have been found to have mutations. A wide spectrum of specific mutations have been identified on exons 1 and 2 of CDKN2A in melanoma families worldwide, and studies comparing the nature of these mutations show evidence of a founder effect in Sweden quite distinct from that seen in Australia and the United
598
R.M. MacKie / Dermatol Clin 20 (2002) 597600
Kingdom [4]. Studies of the quantitative effect of the presence of CDKN2A mutations on melanoma risk suggest that this varies between continents, with the highest risk seen in Australia; intermediate risk in the United States; and lower in risk in Europe with the paradoxical exception of Sweden, where the risk seems to be similar to that for the United States. In mutation-carrying members of melanoma families in the United States, the lifetime risk of developing melanoma by the age of 60 is around 60%. This is obviously a very much greater risk than that of the individual who does not carry the gene mutation, and clearly identifies a group of people who should be offered lifetime surveillance and education about sun exposure. The second gene defect identified in melanoma families is a mutation in the CDK4 gene found on chromosome 12 [5,6]. Only three families worldwide have been identified with this abnormality, and quantitating the degree of increased risk is difficult. The third genetic variation associated with increased melanoma susceptibility is the presence of certain polymorphisms in the MC1R gene on chromosome 16q24.3. Polymorphisms of this gene are
associated with red hair, which has already been identified as a melanoma risk factor, but large studies from both Australia and The Netherlands indicate that certain MC1R polymorphisms carry an additional melanoma risk independent of hair color [7,8]. The proportion of melanoma patients who have possibly causative polymorphisms relative to the proportion of normal individuals with these polymorphisms in that population is not yet well established, but it may well be that the influence of MC1R polymorphisms will vary in different geographic areas around the world. A recent area of investigation is the role of genetic variation in DNA repair genes, and a preliminary study of XPD polymorphisms and melanoma predisposition suggests that there may be overrepresentation of polymorphisms in exons 6,22, and 23 [9]. If confirmed, this could be a valuable link in explaining the mechanism by which excess sun exposure is a risk factor for melanoma. The strongest risk factor yet identified for sporadic melanoma in several large studies is the presence of large numbers of banal melanocytic nevi, with the presence of dysplastic or atypical nevi as an additional and independent risk factor [10,11]. These
Fig. 1. Risk factor chart for melanoma. Category 1 indicates lower-than-average risk, category 2 indicates average risk, category 3 indicates greater-than-average risk, and category 4 indicates very high risk.
R.M. MacKie / Dermatol Clin 20 (2002) 597600
599
observations have led to studies aimed at identifying nevus susceptibility genes, and studies on twin populations indicate that nevus numbers and distribution are determined by genetic and environmental influences in a ratio of 2:1. Although CDKN2A has been implicated as a nevus susceptibility gene and a melanoma susceptibility gene, the relationship between CDKN2A mutations and the presence of large numbers of banal nevi or dysplastic nevi is complex, in that not all CDKN2A mutation positive families have excessive numbers of nevi or atypical nevi, and in families who are CDKN2A mutation positive, melanomas may develop in family members who do not have an abnormal nevus pattern [12]. Using material from a large case control study carried out in Scotland, the author has developed a melanoma risk factor chart aimed at quantitating melanoma risk in patients who do not have a family history of the disease, by far the largest group of the melanoma population [13]. In descending order of statistical significance, the major risk factors are total number of banal nevi (less than 20 or 20 and over); degree of freckling; presence of none, one or two, or three or more clinically atypical nevi; and a history of none, one or two, or three or more episodes of blistering sunburn. These variables are shown in Fig. 1, where it can be seen that with this approach it is possible to categorize the population into those with a less than average melanoma risk, the group at average risk, and moving down the columns those with steadily increasing melanoma risk. This approach has been used in a primary care setting to determine whether or not healthy individuals could self-determine their melanoma risk [14]. This has not been possible, because healthy normal individuals seem to have difficulty with accurate identification of nevi and nevus counts. If the chart is used by professionals, however, it is a good guide to the members of the population who most require sensible sun exposure advice, and possibly ongoing surveillance. A major risk factor for developing a primary melanoma is the presence of a previous primary melanoma. Once one primary melanoma has been diagnosed, the relative risk of a second primary is around 70 compared with the risk of developing a first primary melanoma. Studies on the CDKN2A gene in patients with multiple primaries have indicated that around 10% do have a currently identifiable mutation in this gene [15,16], but most primary melanoma patients at the present time do not have either a positive family history of melanoma or a detectable CDKN2A mutation. Place of birth is also a risk factor for subsequent melanoma development. Studies of migrants to Aus-
tralia, Israel, and California have shown that those who are born and spend the first decade of their lives in high sun exposure areas have for the rest of their lives a higher melanoma risk than do migrants who arrive from less sunny areas as teenagers or young adults [17 19]. A recent case control study from Norway comparing individuals living in north and south Norway has shown that those in the south have twice the melanoma risk even if they move between north and south as adults [20]. In general, animal models of melanoma are poorly developed, but a recent paper studying hepatocyte growth factor and scatter factor transgenic albino mice showed that a single dose of UV 3.5 days after birth was sufficient to induce melanoma at a UV dose 30-fold lower than that needed in adult mice [21]. The authors speculate that this may in part be caused by an immature immune system. A further risk factor for melanoma is higher socioeconomic status. The author has carried out a study on both incidence of and survival from melanoma by social class using the Carstairs index, which uses mainly measure of income and of education. It showed that the more affluent were more likely to develop melanoma, but were less likely to die from it, even after controlling for known prognostic factors, such as Breslow thickness and ulceration [22].
Summary It is now possible to identify a small number of genes in which identifiable abnormalities are associated with an increased risk of melanoma, and a large number of phenotypic factors, which are associated with a lesser risk affecting a very much larger proportion of the melanoma population. As with colon cancer, it seems likely that there is a stepwise development of invasive malignant melanoma, and that interaction between genotype and environmental stimuli may be very different at each successive stage of melanoma progression. Unraveling these steps is the current challenge.
References
[1] Armstrong BK, English DR. Cutaneous malignant melanoma. In: Shottenfield D, Fraumeni JFJ, editors. Cancer epidemiology and prevention. New York: Oxford University Press; 1996. p. 1282 1312. [2] Piepkorn M. Melanoma genetics: an update with focus on the CDKN2A/ARF tumour suppressors. J Am Acad Dermatol 2000;42:705 22. [3] Goldstein AM, Tucker MA. Genetic epidemiology
600
R.M. MacKie / Dermatol Clin 20 (2002) 597600 of cutaneous melanoma. Arch Dermatol 2001;137: 1493 7. Bishop DT, Goldstein A, Demenais F, et al. Geographic variations in the penetrance of CDKN2A mutations for melanoma. J Natl Cancer Inst 2002; 67(Suppl 2):33. Zuo L, Weger J, Yang Q, et al. Germline mutations in the p16INK4A binding domain of CDK4 in familial melanoma. Nat Genet 1996;12:97 9. Tsao H, Benoit E, Sober AJ, et al. Novel mutations in the p16/CDKN2A binding region of the cyclin dependent kinase 4 gene. Cancer Res 1998;58:109 13. Palmer JS, Duffy DL, Box NF, et al. Melanocortin 1 receptor polymorphisms and risk of melanoma: is the risk explained solely by pigmentation phenotype? Am J Hum Genet 2000;66:176 86. Van der Velden PA, Sandkuijl LA, Bergman W, et al. Melanocortin 1 receptor variant R151C modifies melanoma risk in Dutch families with melanoma. Am J Hum Genet 2001;69:774 9. Tomescu D, Kavanagh G, Ha T, Melton D. Nucleotide excision repair gene XPD polymorphisms and genetic predisposition to melanoma. Carcinogenesis 2001;22: 403 8. Sverdlow AJ, English J, MacKie RM, et al. Benign melanocytic naevi as a risk factor for melanoma. BMJ 1986;292:1555 9. Holly EA, Kelly JW, Shpall SN, Chiu SH. Number of melanocytic naevi as a major risk factor for malignant melanoma. J Am Acad Dermatol 1987;17:459 68. Newton Bishop J, Wachsmuth RC, Harland M, et al. genotype/phenotype and penetrance studies in melanoma families with germline CDKN2A mutations. J Invest Dermatol 2000;114:28 33. [13] MacKie RM, Freudenberger T, Aitchison TC. Personal risk factor chart for cutaneous melanoma. Lancet 1989; 26:164 8. [14] Jackson A, Wilkinson C, Ranger M, Pill R, August P. Can primary prevention or selective screening for melanoma be more precisely targeted through general practice? BMJ 1998;316:34 8. [15] MacKie RM, Andrew N, Lanyon WG, Connor JM. CDKN2A germline mutations in UK patients with familial and multiple primary melanomas. J Invest Dermatol 1998;111:269 72. [16] Monzon J, Liu L, Brill H, et al. CDKN2A mutations in multiple primary melanomas. N Engl J Med 1998;338: 879 87. [17] Holman CDJ, Mulroney CD, Armstrong BK. Epidemiology of preinvasive and invasive melanoma in Western Australia. Int J Cancer 1980;25:317 23. [18] Anaise D, Steinitz R, Ben Hur N. Solar radiation: a possible aetiological factor in malignant melanoma in Israel. Cancer 1978;42:299 304. [19] Mack TM. Cancer surveillance programmes in Los Angeles County. In: Epidemiology and cancer registries in the Pacific basin. National Cancer Institute monograph no 47. Washington; 1977. p. 99 101. [20] Robsahm TE, Tretl S. Cutaneous malignant melanoma in Norway. Cancer Causes Control 2001;12:569 76. [21] Noonan F, Recio JA, Takayama H, et al. Neonatal sunburn and melanoma in mice. Nature 2001;412: 271 2. [22] MacKie RM, Hole DJ. Incidence and thickness of primary tumours and survival of patients with cutaneous malignant melanoma in relation to socio economic status. BMJ 1996;312:1125 8.
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
Dermatol Clin 20 (2002) 601 606
The effect of sunscreen on melanoma risk

Darrell S. Rigel, MD
Ronald O. Perelman Department of Dermatology, New York University Medical Center, 35 East 35th Street, Suite 208, New York, NY 10016, USA
The sun provides energy for all living things on earth. Along with these beneficial effects, there are significant negative events that may result from excessive sun exposure. In terms of the integument, it is the UV band of the irradiance spectrum that seems to provide the greatest degree of destruction. Sun protection is designed to protect the skin from the damage that occurs as a result of UV exposure. Unprotected UV exposure leads to two significant areas of skin problems. Most important is the increased risk of developing skin cancer. At current rates, 1 in 5 Americans develop a skin cancer of some sort during their lifetime, with over 1,300,000 new cases appearing this year. The incidence of malignant melanoma, the most dangerous type of skin cancer, is increasing faster than any other cancer in the United States [1]. In 1935, the lifetime risk for an American developing melanoma was 1 in 1500. It is estimated that there will be 87,900 new cases (53,600 invasive and 34,300 in situ) of melanoma in the United States this year [2]. At 2002 rates, the risk for developing invasive melanoma is now 1 in 68 and is projected to increase to 1 in 50 by the year 2010. In addition, according to the World Health Organization, melanoma is increasing faster than any other malignancy worldwide. Health care costs associated with the treatment of these skin cancers are over $100 million in the United States alone [3]. Effective mechanisms that protect the skin from skin cancer causing UV rays are critical.
of sun protection usually has been behaviorally reinforced early in life. The importance of sun protection has also been recognized since earliest times. Ancient Egyptian art depicts persons wearing long robes and sitting in the shade of trees. Yet, it was not until the 1950s when suntan lotions came into usage. Their purpose was to help the person tan without burning. In the early 1970s, the first true sunscreen, para-aminobenzoic acid (PABA), became generally available. High-intensity sunscreens, however, have only been marketed over the past 15 years.
Sunblocks Sunblocks physically block the suns UV radiation to the skin. Protective clothing, umbrellas, and trees are considered sunblocks. Sunblocks can be in opaque forms and applied to the skin. When used in a topically applied form, these agents scatter, reflect, and primarily block UV light. They are partially effective for patients who have diseases related to light exposure including lupus erythematous, polymorphous light eruption, xeroderma pigmentosum, and solar urticaria. Sunblocks are also useful in persons spending extensive periods of time outdoors (eg, golfers, tennis players, lifeguards, and so forth) and to protect sensitive areas, such as the nose, ears, and cheeks. The most commonly used sunblock agents are zinc oxide, talc, titanium dioxide, and red vetenary petrolatum. Most topical sunblocks contain combinations of a subset of this list. The most commonly used individual agent is zinc oxide. Certain disadvantages exist with the use of topical physical sunblocks. They tend to be messy and can stain clothing and fiberglass recreational equipment. To increase their cosmetic and social accept-
History of sun protection Because most persons experience at least one painful sunburn during their lifetime, the importance
E-mail address: dsrigel@prodigy.net (D.S. Rigel).
602
D.S. Rigel / Dermatol Clin 20 (2002) 601606
ability, sunblocks with designer coloring are now being distributed. Newer formulations with micronized titanium dioxide have good UV blocking capabilities but are also almost invisible and more cosmetically desirable.
Sunscreens Table 1 lists common sunscreen agents and their UV protective wavelengths. PABA has been used in sunscreens in the United States since the early 1970s. One of the major characteristics of PABA that makes it effective as a sunscreen is its ability to bind to epidermal cells. This tends to make PABA-based sunscreens fairly water and perspiration resistant, but also makes them prone to staining. PABA tends to block UV radiation most effectively in the UVB zone (290 to 320 nm). The PABA esters (padimate A, padimate O, and glyceryl PABA) have the absorption characteristics of
PABA but have the additional advantage of only rarely staining. Most current PABA-containing sunscreens use the PABA esters. The PABA has a major disadvantage over other sunscreen components. There is a much higher presence of contact and photocontact allergy to PABA than to other sunscreen agents [4]. Among PABAbased chemicals, padimate A is often chosen for use in sunscreen compounds because of the PABA esters lower allergic incidence. Sunscreens with PABA esters may contain up to 0.2% to 4.5% PABA. This may account for the many allergic reactions seen with these agents. For all of these reasons, most new formulations do not include PABA.
Benzophenones Benzophenones are the second most commonly found component of sunscreens. Although their primary protective range is found in the UVA range (320 to 400 nm), a secondary protective band is noted in the UVB zone. Benzophenones were originally used alone as a PABA-free sunscreen alternative, but are now combined with other sunscreen agents to provide broad-spectrum coverage. The most commonly used benzophenone agents are oxybenzone and dioxybenzone. These ingredients are much less allergenic than PABA and do not stain. The benzophenones, however, are less water resistant than PABA. The bases that are used in benzophenone-containing sunscreens must be thicker and are less cosmetically acceptable.
Table 1 Current common sunscreen agents and their UV protective wavelengths Range of protection (nm) 260 313 290 315 290 315 260 313 Maximal effect of protection (nm) 283 311 309 297
Sunscreen PABA and PABA esters PABA Padimate O Padimate A Glycerol aminobenzoate Cinnamates Octyl methoxycinnamate Cinoxate Salicylates Homosalicylate Octyl salicylate Triethanolamine salicylate Octocrylene Etocrylene Benzophenones Oxybenzone Dioxybenzone Sulisobenzone Menthylanthranilate Dibenzoylmethanes Tert-butylmethoxydibenzoylmethane (Parsol) 4-isopropyldibenzoylmethane (Eusolex)
280 310 270 328 290 315 260 310 269 320 287 323 296 383 270 350 206 380 250 380 200 380 310 400
311 290 306 307 298 303 303 290 325 284 327 286 324 336 358
Cinnamates Cinnamates, a derivative of cinnamon, are also good protectors from the sun. Their products are chemically related to balsam of Peru, tolu balsam, coca leaves, cinnamic aldehyde, and cinnamic oil. Persons with sensitivity to these items may crossreact to sunscreens containing cinnamates. The most commonly used cinnamates are octyl methoxycinnamate and cinoxate. They are nonstaining but also have poor water-resistant qualities. Cinnamate products may require more frequent reapplication or special substantive bases.
Salicylates and anthranilates

310 400 345
Salicylates are among the original sunscreen chemicals. Homomenthyl salicylate absorbs prima-
603
rily in the UVB range and is typically added to other components to increase sun protection factor (SPF). Octyl and triethanolamine salicylates are also used. They may cause photocontact dermatitis more frequently than homomenthyl salicylate and are used less frequently. Anthranilates, such as menthyl anthranilate, with low-level broad-spectrum coverage, are also added to many sunscreens to augment protection.
Dibenzoylmethanes Recent concerns over the effects of UVA radiation to the skin have demonstrated the need for better UVA protection in sunscreens. The newest ingredients that have shown to have the best UVA protection are the dibenzoylmethanes. Because they offer no protection from UVB rays, they must be used in combination with other ingredients. Tert-butylmethoxydibenzoylmethane (Avobenzone, Parsol 1789) is approved for use in the United States. Its range of coverage is 310 to 400 nm, with a peak effectiveness at 358 nm. A second member of this family, isopropyldibenzoylmethane (Eusolex 8020), has been used in Europe for several years. Because of the high incidence of contact dermatitis reactions associated with its use, it has not been approved for incorporation into sunscreens in the United States.
SPF System The effectiveness of a sunscreen is measured by the use of the SPF system [5]. SPF is the ratio of the amount of time that a person exposed to the sun takes to sunburn while wearing a sunscreen compared with the time required to sunburn without protection. Initial sunscreens in the 1970s had SPFs of 2 to 4. These sunscreens, however, only provided protection from 50% to 75% of the UVB rays. Current highpotency sunscreens have SPFs ranging from 15 to 50, protecting from 93% to 98% of UVB. Several significant problems, however, exist with the SPF system. First, SPFs measures only UVB protection. In natural sunlight on a summer day at noon, approximately 10% of the UV energy striking the skin is in the UVA band. In general, because UVA has so much less energy, it takes an impractical amount of time to test an SPF for UVA in a manner similar to UVB. Since 1979, the Food and Drug Administration has explored different methods for measuring the effectiveness of UVA
protection that a sunscreen provides. It is hoped that there will be a resolution of this problem in the near future. Second, the SPF of a sunscreen is measured under ideal, controlled conditions. The true effectiveness of a sunscreen is a function of its usage. On average, it takes approximately 1 oz of sunscreen to cover the entire body. Using less than this amount results in lowering the true SPF. Sunscreens take about 15 to 20 minutes from the time of application until they become effective. They should be applied before the time that they are needed. Depending on some sunscreen components and bases, some mixtures may be more resistant to water. The term water resistant is defined as the sunscreen having the same SPF after a person has been immersed in water for 40 minutes as it did initially. Very water resistant is similar except that no degradation of protection is present at 80 minutes. Sunscreens with less substantivity (water resistance) may need to be reapplied more frequently to maintain their predicted SPF level of protection. A debate exists as to the value of additional protection provided by sunscreens with SPFs higher than 15. In a controlled environment, the marginal protection provided by these high SPF sunscreens is only 2% to 4%. Because most persons underapply sunscreens, however, the higher SPF sunscreens have a margin of safety hopefully giving the user at least an SPF 15 level of protection. Recent studies have shown that higher SPF sunscreens may provide more effective protection than those with SPFs of 15. Cesarini et al [6] described the first occurrence where there were fewer sunburn cells (markers for acute sun damage) in fair-skinned individuals using greater than 15 SPF sunscreens. These sunburn cells are regarded as markers for UV radiation induced damage to DNA. Because altered DNA is thought to be the effector of photocarcinogenesis, it was hypothesized that higher SPF sunscreens offered better protection from skin cancer. Further studies are needed to verify the clinical relevance of these findings. Nevertheless, SPF greater than 15 sunscreens can be recommended for persons with type I skin, with a personal or family history of skin cancer, or in those persons who do not apply sunscreen adequately.
Vehicles One of the most common objections to sunscreen usage is lack of cosmetic acceptability of the product. The ingredients used to formulate sunscreen
604
vehicles include mineral oil, petrolatum, isopropyl esters, lanolin, aliphatic alcohols, triglycerides, fatty acids, waxes, propylene glycol, emulsifiers, thickeners, preservatives, and fragrance. A sunscreen acceptable to one person or for use in a given area of the body may not be acceptable in other situations. Oil-based sunscreens may be comedogenic. Recently, water-and perspiration-resistant formulations with SPF 15 or greater that do not sting on eye contact have been developed for athletics. The development of better vehicles in the future may help to enhance sunscreen usage.
(up to 26%), contact sensitization can occur. Because the active agents in sunscreens absorb radiation, they also have the potential to cause photosensitization. Both types of reactions can occur from not only the sunscreen agents themselves, but also from components of the vehicles. The most common offending agents in vehicles and preservatives are as follows: Avocado oil t-Butyl alcohol Methylparabens Phenyldimethicone Solvent red 1 Solvent red 3 Triethanolamine stearate Benzyl alcohol Cetostearyl alcohol Sorbitan sesquiolate Imidazolidinyl urea (Germall 115) Methylisothiazolinone and methylchloroisothiazolinone (Kathon CG) Glyceryl monostearate 6-Acetoxy-2, 4-dimethyl-m-dioxane Carbowaxes Ethyl alcohol Glycerol Isopropyl alcohol Isopropyl myristate Petrolatum Stearyl alcohol The most common sunscreen agent causing an allergic reaction is PABA. It has been estimated that 3% to 7% of the US population is PABA sensitive [8]. Both contact and photoallergic reactions have been reported. Benzophenone sensitization has been estimated at 1% to 2%. Other sunscreen agents also can cause contact reactions. By far, the most common sunscreen reaction is caused by sensitive skin. It is estimated that up to one in three persons using sunscreens at some time complain about sunscreens irritating their skin. Complaints include subjective signs, such as stinging, burning, and itching, and objective findings, such as urticaria, acnegenicity, and pustulgenicity. The most effective method for alleviating these findings is to choose an alternative sunscreen from a different chemical family or with a different base. True sunscreen reactions need to be patch and photopatch tested to determine whether the reaction is merely allergic or possibly light related. Use of too low a concentration of the testing materials may result in a false-negative result.
Protection from immunologic effects of UV radiation A UV radiation to the skin results in various forms of local and systemic immune suppression. Release of preinflammatory cytokines (interleukin-1 and-6, tumor necrosis factor) and prostaglandins is inhibited. Langerhans cells in the skin are depleted and inactivated. Autoimmune diseases, such as systemic lupus erythematous, can be activated. Similar activation of herpes virus eruption is noted. Allergic contact dermatitis can be suppressed. In addition, in mice, skin cancer surveillance and rejection mechanisms are suppressed. Can sunscreen prevent UV-induced immunologic damage? Clinically, sunscreens seem to reduce outbreaks of perioral herpes simplex. In 15 studies conducted from 1981 through 1992, 5 showed no protection, 17 showed partial protection, and 3 demonstrated complete protection from general immune damage [7]. Further studies are needed to understand better the relationship between sunscreens and immunologic response.
Sunscreen reactions Many types of sunscreen reactions have been reported. People often complain of a reaction to a specific sunscreen and do not subsequently use the product. These types of reactions include

Allergic contact dermatitis Photoallergic contact dermatitis Irritant dermatitis Acne Aesthetic issues
Because sunscreens are applied topically to the skin frequently and in relatively high concentrations
605
Sunscreen protection against the development of melanoma Whether sunscreens can reduce the risk for melanoma has not yet been proved directly [9]. Inferential evidence, however, strongly supports a protective effect. Because direct protective effects against skin cancer have only been demonstrated recently in humans [10 12] some have suggested that sunscreens might not reduce the risk for melanoma. The absence of direct proof has emboldened the most extreme critics to suggest that sunscreen use might actually increase melanoma risk [13]. It is critical that a careful analysis of existing studies be undertaken before conclusions are drawn [14]. A variety of factors (some or all of which exist in prior studies) may have led to why disparate conclusions were reached in these evaluations. One of the most common flaws noted is related to the fact that persons tend to be poor historians in terms of reporting prior sunscreen usage. In addition, because sunscreen usage is generally deemed to be a healthy behavior, subjects are more likely to say that they applied it to themselves or their children than may have been the case. Also, given a latency in the development of melanoma of up to many years, the assumption that only the behaviors during the years that are included in the study influenced melanoma risk is flawed [15]. Many of the sunscreen studies use sunburn protection or nevus development as a measure of potential melanoma protection. Rather than using a surrogate end point, the best end point to test the relationship of sunscreen to melanoma is the development of melanoma itself. Also, many of the prior studies do not look at the interrelationships between melanoma protection and other factors. Data on other factors (ie, sun avoidance, seeking shade, and so forth) and clothing usage need to be evaluated in a careful multivariate analysis to determine which factors were causal. Data for clothing usage need to be defined better. The SPF of a dry, white, cotton T-shirt is approximately six, but wet approximately three [16]. If sunscreens fail through allowing some of the UV rays to hit the skin, why would summer-weight clothes (with a lower SPF) be any better? Multivariate analysis might also demonstrate that fair-skinned persons with high constitutional risk factors for nevi development were simply more likely to be high sunscreen users rather than sunscreen use itself being the source of the problem. Finally, and most importantly, high sunscreen use could simply be a surrogate for those who received the highest dose of UV exposure because of a combination of inadequate sunscreen use and more time spent outdoors.
Sunscreen usage needs to be defined better. Sunscreen use means different things to different persons and the definitions used in prior study were blurred. How much was applied? Studies show that far less sunscreen is used than is recommended [17]. Frequency of application and time of application (whether applied before exposure or after exposure has begun) [18] need to be determined when data are collected to assess critically the impact of sunscreen on melanoma prevention. The most appropriate end point for determining the effectiveness of public policies regarding sun protection is the melanoma incidence rate. In Australia, where 74% of the population regularly uses sunscreen [19], melanoma incidence and mortality rates are beginning to fall. Although other factors, such as changing demographics, may be contributory, because sunscreens are the most commonly used method of sun protection in Australia it is clear that sunscreens are a major factor in this outcome. Melanoma rates are also falling among white Hawaiians, a group that has among the highest per capita use of sunscreen in the United States [20]. Until there are prospective data to show the effect of sunscreens on melanoma risk more directly, this is some of the best evidence of their effect on melanoma incidence in large populations of individuals. How could a study be designed to determine accurately the relationship between sunscreen usage and melanoma risk? The significant limitations of previous retrospective studies regarding the efficacy of sunscreens must be recognized. Earlier studies that have significant problems, such as those that evaluated patients during times when higher SPF sunscreens were not yet available, need to be eliminated from consideration. Other studies need to be evaluated within the context of the biases inherent to all retrospective studies. Methodologic problems are the most likely explanation for the diverse and often conflicting results that occur across the various retrospective studies. A prospective study to determine sunscreen efficacy needs to be undertaken. There are significant problems, however, that must be overcome. The prevalence of melanoma in the population is relatively low. To have a study with adequate statistical power a large number of subjects are needed. In addition, there is a 10- to 50-year latency period between the time that sun exposure initially occurs and the time that melanoma appears clinically. A follow-up period of at least 10 years is needed if the study is going to use children or young adults of average risk. The difficulties of obtaining accurate sun exposure and sunscreen usage history for the period before
606
D.S. Rigel / Dermatol Clin 20 (2002) 601606 [4] Mackie BS, Mackie LE. The PABA story. Australas J Dermatol 1999;40:51 3. [5] Sayre RM, Desrochers DL, Marlowe E, Urbach F. The correlation of indoor solar simulator and natural sunlight: testing of a sunscreen preparation. Arch Dermatol 1978;114:1649 51. [6] Cesarini JP, Chardon A, Binet O, Hourseau C, Grollier JF. High-protection sunscreen formulation prevents UVB-induced sunburn cell formation. Photodermatol 1989;6:20 3. [7] Rooney JF, Bryson Y, Mannix ML, Dillon M, Wohlenberg CR, Banks S, et al. Prevention of ultraviolet-lightinduced herpes labialis by sunscreen. Lancet 1991;338: 1419 22. [8] Darvay A, White IR, Rycroft RJ, Jones AB, Hawk JL, McFadden JP. Photoallergic contact dermatitis is uncommon. Br J Dermatol 2001;145:597 601. [9] Bigby M. The sunscreen and melanoma controversy. Arch Dermatol 1999;135:1526 7. [10] Green A, Williams G, Neale R, et al. Daily sunscreen application and betacarotene supplementation in prevention of basal-cell and squamous-cell carcinomas of the skin: a randomised controlled trial. Lancet 1999;354:723 9. [11] Naylor MF, Boyd A, Smith DW, et al. High sun protection factor (SPF) sunscreens in the suppression of actinic neoplasia. Arch Dermatol 1995;131:170 5. [12] Thompson SC, Jolley D, Marks R. Reduction of solar keratoses by regular sunscreen use. N Engl J Med 1993;329:1147 51. [13] Garland CF, Garland FC, Gorham ED. Rising trends in melanoma: an hypothesis concerning sunscreen effectiveness. Ann Epidemiol 1993;3:103 10. [14] Rigel DS, Robinson JK, Naylor M. What is the evidence for a sunscreen and melanoma controversy? Arch Dermatol 2000;136:1447 9. [15] Autier P, Dore JF, Cattaruzza MS, et al. Sunscreen use, wearing clothes, and number of nevi in 6- to 7-year-old European children. J Natl Cancer Inst 1998;90:1873 80. [16] Menter JM, Hollins TD, Sayre RM, Eternadi AA, Willis I, Hughes SN. Protection against UV photocarcinogenesis by fabric materials. J Am Acad Dermatol 1994;31(5 pt 1):711 6. [17] Stokes R, Diffey B. How well are sunscreen users protected? Photodermatol Photoimmunol Photomed 1997;13:186 8. [18] Pruim B, Green A. Photobiological aspects of sunscreen re-application. Australas J Dermatol 1999;40:14 8. [19] Martin RH. Relationship between risk factors, knowledge and preventative behavior relevant to skin cancer in general practice patients in South Australia. Br J Gen Pract 1995;45:365 7. [20] Chuang TY, Charles J, Reizner GT, Elpern DJ, Farmer ER. Melanoma in Kauai, Hawaii, 1981 1990: the significance of an in situ melanoma and the incidence trend. Int J Dermatol 1999;38:101 7. [21] Lakhani S, Flavell RA. Natural sunscreen revealed. Nat Cell Biol 2001;3:272.
the studies inception must be overcome to minimize the effects of earlier behaviors. Finally, any placebocontrolled study design needs to overcome the ethical dilemma of depriving a subject of the use of sunscreen when strong inferential data already suggest that sunscreens protect from skin cancer when used appropriately.
Future trends Many other substances are currently being investigated for their potential as sunscreens [21]. Psoralen derivatives that enhance the production of melanin are being studied. Natural substances including derivatives of coral from Australias Great Barrier Reef have shown some early promise. Ultimately, oral preparations with appropriate concentration delivery to the skin to make the sunscreen effective will be developed.
Summary The usage of sunscreens has grown dramatically worldwide over the past decade. Current data suggest that a regimen of sun protection that includes protective clothing, avoiding midday sun, and regular use of broad-spectrum high SPF sunscreen (such as practiced in Australia [19]) seems to be reducing melanoma incidence rates. This is the current recommendation of the American Academy of Dermatology and it is also the recommendation that is best supported by existing data. Except for total sun avoidance, sunscreens remain the best individual method of protection from UV-induced damage to the skin. It is hoped there will be even more definitive answers to questions related to the effectiveness of sunscreens for reducing melanoma risk as better sunscreen components are developed and as evaluations are performed in the future that overcome the problems better in existing studies. References
[1] Rigel DS, Carucci JA. Malignant melanoma: prevention, early detection, and treatment in the 21st century. CA Cancer J Clin 2000;50:215 36. [2] Jemal A, Thomas A, Murray T, Thun M. Cancer statistics, 2002. CA Cancer J Clin 2002;52:23 47. [3] Tsao H, Rogers G, Sober A. An estimate of the annual direct cost of treating cutaneous melanoma. J Am Acad Dermatol 1998;38(5 pt 1):669 80.
Dermatol Clin 20 (2002) 607 616
Congenital melanocytic nevi Evaluation and management

Ashfaq A. Marghoob, MD
Department of Medicine, Dermatology Division, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021, USA
Congenital melanocytic nevi (CMN) are nevomelanocytic nevi that are present at birth. Some CMN may not be apparent at birth because they lack visible pigment; however, these tardive CMN may slowly develop pigment over time and become visible (Fig. 1) [1,2]. An example highlighting this point can be observed in patients with large CMN who have multiple satellite nevi. Satellite nevi, considered to be congenital in origin, are usually present at birth; however, some do not become clinically manifest until months to years after birth. Those nevi with congenital features in which the history of their presence since birth cannot be verified are termed congenital-nevuslike nevi (CNLN) [3]. Most CNLN are probably true CMN; however, some of them may be acquired nevi with congenital features. Unfortunately, the histologic overlap between CMN and acquired nevi also precludes accurate classification [4]. Perhaps molecular or genomic research may someday reveal the true origin of all nevomelanocytic nevi. Between 1% and 6% of infants are born with a CMN [5,6], and between 2% and 6% of the population have a CNLN [7 10]. CMN usually present as round to oval, fairly homogeneous, brown, multishaded pigmented lesions with sharply demarcated borders, often with a mammillated surface and hypertrichosis. Some CMN, however, especially the larger ones, can be heterogeneous displaying multiple colors, irregular topography, and a nodular surface (Fig. 2). Except during early infancy, when some CMN can grow rapidly, most CMN grow in proportion to the growth
E-mail address: marghooa@mskcc.org (A.A. Marghoob).
of the child [11]. With the passage of time CMN may get darker; lighter; lose pigmentation; become more heterogeneous; become more homogeneous; develop a nodular surface; or rarely, regress (Fig. 3) [12,13]. CMN are typically classified according to the size they attain, or are predicted to attain in adulthood [14,15]. CMN that are less than 1.5 cm in greatest diameter are classified as small, CMN measuring between 1.5 and 19.9 cm are classified as medium, and CMN that measure at least 20 cm in greatest diameter are classified as large [16]. The very large CMN are also known as giant nevi. CMN form during early embryogenesis between the 5th and 24th week of gestation. Although there are reported cases of familial clustering of CMN, most occur sporadically. Research is beginning to shed some light on the cause of CMN. It is theorized that during embryogenesis a morphogenic error in the neuroectoderm results in the dysregulated growth of melanoblasts (precursors to melanocytes) [17]. In addition, the melanoblasts migrate from the neural crest to the leptomeninges and integument in a dysregulated pattern [18]. Perturbations in migration, proliferation, or differentiation of the neural crest derived melanoblasts may be linked to the c-met proto-oncogene, which controls the expression of the tyrosine-kinase receptor Met. This receptor binds hepatocyte growth factor scatter factor and, if either the Met receptor or its ligand (hepatocyte growth factor scatter factor) are overexpressed, may result in the formation of CMN or neurocutaneous melanocytosis (NCM) (Fig. 4) [17,19,20]. It is believed that hepatocyte growth factor scatter factor serves partly as a chemotactic agent, which guides migration of those melanoblasts that express Met [20]. The c-met
0733-8635/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 0 7 3 3 - 8 6 3 5 ( 0 2 ) 0 0 0 3 0 - X
608
A.A. Marghoob / Dermatol Clin 20 (2002) 607616
Fig. 1. This melanocytic nevus has a diameter of 2.3 cm. It has a mamillated surface and hypertrichosis, both features commonly seen in congenital melanocytic nevi; however, this congenital-nevus-like nevus first became clinically visible at the age of 3 years.
proto-oncogene may also play a role in the development of melanoma or rhabdomyosarcomas, both of which occur at a higher than expected frequency in patients with large CMN [17,21 24]. A CMN can have medical, cosmetic, and psychologic ramifications. Behavioral and emotional problems are reported to occur in as many as 30% of patients with large CMN [25]. The psychologic burden on patients and parents may stem from the cosmetic appearance of the CMN; the anxiety associated with the knowledge that complications, such as melanoma, can develop; the discomfort associated with the often multiple-staged surgical treatments rendered; and the cosmetic appearance of resultant scars [25]. In addition, some CMN or nevi excision scars may develop
Fig. 3. Congenital melanocytic nevus on the finger that developed multiple papules on the surface. These papules represent neurotization.
tenderness, pruritus, or skin fragility, adding to the patients discomfort. Melanoma can develop within any CMN [26]; however, the risk for developing melanoma seems to correlate at least to some degree with the size of the nevus. In turn, the size of the nevus frequently correlates with the depth of penetration of the nevus cells and the total number of nevomelanocytes present within the nevus [16,27,28]. Individuals at greatest risk are those with large CMN. The nevus cells in large CMN are frequently found to penetrate deep into the dermis or beyond. The estimated lifetime risk
Fig. 2. Heterogeneous large congenital melanocytic nevus with multiple colors; areas that are flat; and areas that are raised, papular, and mammillated.
Fig. 4. Hepatocyte growth factor scatter factor Met pathway controls melanocyte proliferation and migration.
Table 1 Frequently asked questions regarding congenital melanocytic nevi CMN size 1. What is the reported lifetime risk for developing melanoma? 2. Where do melanomas develop? Small and Medium 0% 4.9% Within the CMN CNS (if ! 3 CMN or multiple satellite nevi present) Dermoepidermal junction Peripheral edge of the CMN Because melanomas tend to develop at the dermoepidermal junction, they can be recognized easily clinically and diagnosed early At or after puberty Patients with ! 3 CMN or multiple satellite nevi Clinical follow-up Monthly self skin examination Cosmetic improvement (make-up, lasers, excision) Prophylactic removal Removal can be planned at anytime up to puberty Excision (simple or serial, with or without skin grafts or tissue expanders) Clinical inspection Dermoscopy Confocal laser microscopy Baseline photographs used during follow-up examinations to help detect subtle changes Cosmetic issues because of CMN or surgical scars Psychosocial Dermatologist Plastic surgeon Psychologist MRI of brain if risk factors for NCM are present Large 4.5% 10% Within the CMN CNS Other (ie, retroperitoneal) Anywhere within the CMN but most frequently below the dermoepidermal junction (ie, dermis) Because many melanomas develop deep within the CMN, the clinical recognition of early melanoma is often difficult Before puberty CMN located in axial locations (ie, head and neck, midline back) Patients with multiple satellite CMN Clinical follow-up Monthly self skin examination Cosmetic improvement (make-up, lasers, excision, dermabrasion) Prophylactic removal Treatment should be rendered early in life Excision (simple or serial, with or without skin grafts or tissue expanders) Curettage in neonatal period? Clinical inspection Palpation Baseline photographs to help detect subtle changes Dermoscopy? Confocal laser microscopy? Cosmetic issues because of CMN or surgical scars Psychosocial Other malignancies (ie, rhabdomyosarcoma) Symptoms (ie, pruritus, tenderness, skin fragility) Dermatologist Plastic surgeon Pediatric neurologist Psychologist MRI of brain (radiologist)
3. If melanoma develops within the CMN, where does it start? 4. What is the ease of diagnosing cutaneous melanoma? 5. When do most melanomas develop? 6. Who is at increased risk for NCM? 7. What are the management options?
8. If prophylactic removal is decided, when should it be done and what methods are available?
9. Aids to the detection of early melanoma (if prophylactic excision is not an option or if portions of the nevus are not excised)?
10. What complications other than melanoma can develop?
11. What laboratory tests or consultations should be considered?
All patients should be instructed on the importance of sun avoidance and sun protection. Abbreviations: CMN; congenital melanocyte nevi; NCM, neurocutaneous melanocytosis.
609
610
Fig. 5. This patient with large congenital melanocytic nevus and multiple satellite nevi developed a melanoma within the large nevus on his right thigh. Melanomas developing in large congenital melanocytic nevi may have a bimodal age distribution with most occurring in early childhood and some occurring in late adulthood, as in this patient. (Courtesy of Alfred Kopf, MD.)
between true melanoma and its simulants can be difficult at times, and together with patient selection and referral bias may partly explain the overestimated melanoma risks previously reported, some as high as 42%. Evaluating nevus biopsy specimens for DNA content, cell cycle aberrations, clonality, gene mutations, or the overexpression or underexpression of specific proteins may someday aid in differentiating melanoma from its simulants [34]. The ability to detect mutations and protein expression has been facilitated by new microdissection techniques, immunohistochemistry, polymerase chain reaction DNA amplification, DNA-microarray technology, comparative genomic hybridization, and fluorescence in situ hybridization, all of which may prove useful in correctly classifying melanocytic neoplasms [34]. Although it is clear that patients with large CMN are at increased risk for developing melanoma, this risk may not be equal for all such patients. There may be differences in risk between large and very large CMN (Fig. 6A, B); superficial and deep penetrating CMN (Fig. 6C, D); macular and rugose CMN (see Fig. 6A, C, D); and homogeneous and heterogeneous CMN (see Fig. 6C, A, D). Continued prospective clinical follow-up will help better identify those at highest risk. Individuals with large CMN are also at risk for NCM, an entity in which the leptomeninges contain excessive amounts of melanocytes and melanin
for developing melanoma in patients with large CMN is between 4.5% and 10%, and the relative risk is between 101 and 1046 [23]. Individuals with large CMN can develop melanoma at any age (Fig. 5); however, 70% of the melanomas are diagnosed in children under the age of 10 [29]. These primary melanomas can occur within the large nevus or within the central nervous system [23]. Twenty-three percent of patients with large CMN and melanoma present with metastatic disease in which the primary site of the melanoma cannot be found [29]. It is possible that in these patients the focus of the primary melanoma is hidden somewhere within their large CMN. In fact, it is estimated that two thirds of melanomas forming within large CMN develop subepidermally (ie, nonepidermal sites), making their clinical detection challenging (Table 1) [30]. A word of caution is warranted. Biopsy of growths that develop within large CMN, particularly those found during early infancy, may have some histologic features of melanoma, yet behave in a clinically benign fashion [31 33]. The clinical and histologic differentiation
Fig. 6. (A) Heterogeneous (speckled) large congenital melanocytic nevus. (B) Very large congenital melanocytic nevus (giant nevus). (C) Relatively flat and homogeneous light brown, large congenital melanocytic nevus. (D) Raised, rugose, very large congenital melanocytic nevus on the back. From Marghoob AA, Schoenbach SP, Kopf AW, et al. Large congenital melanocytic nevi and the risk for development of malignant melanoma: a prospective study. Arch Dermatol 1996;132:170 5; with permission.
611
Fig. 7. Cumulative 5-year life-table risk for developing melanoma or neurocutaneous melanocytosis in patients with large congenital melanocytic nevi. (From Bittencourt FV, Marghoob AA, Kopf AW, et al. Large congenital melanocytic nevi and the risk for development of malignant melanoma and neurocutaneous melanocytosis. Pediatrics 2000;106:736 41; reproduced by permission of Pediatrics.)
[14,35]. Malignant degeneration of these melanocytes results in development of primary central nervous system melanomas. Benign proliferation of melanocytes within the leptomeninges, however, can also result in serious complications and even death. The 5-year cumulative risk for developing melanoma or NCM in individuals with large CMN is shown in Fig. 7 [35]. Patients with large CMN at greatest risk for developing melanoma or NCM are those with multiple satellite congenital nevi (see Fig. 5), and those in whom the large CMN is in paravertebral or axial locations, such as the back, head, or neck (see Figs. 2, and 6B, D) [14,23,24]. In addition, patients with multiple medium-sized CMN (Fig. 8) or with multiple satellite nevi but no large CMN may also be at increased risk for developing NCM [36,37]. Despite the fact that most patients with large CMN have multiple satellite nevi, no melanoma has ever been reported to have developed within one of these satellite nevi [24]. The CMN that are less than 20 cm in diameter have a reported lifetime risk for developing melanoma of between 0% and 4.9% [16,38]. Two recent studies evaluating medium-sized CMN did not reveal an increased risk for melanoma developing in nevi of this size [39,40]. There are many case reports and series of cases, however, that clearly show that melanomas can and do develop in CMN that are less than 20 cm in diameter, and these lesions require monitoring (Fig. 9) [41]. In contrast to large CMN,
melanomas developing in smaller CMN tend to develop at or after puberty (see Table 1). Further-
Fig. 8. This person with multiple medium-sized congenital melanocytic nevi and satellite nevi is at risk for neurocutaneous melanocytosis. This 31-year-old man is healthy, however, without neurologic symptoms, and has a normal MRI.
612
Fig. 9. This 70-year-old man with a small congenital melanocytic nevus developed melanoma at the edge of the congenital melanocytic nevus. The melanoma developed at the dermoepidermal junction. The nevus was confirmed to be present since birth by history and by the identification of the nevus on baby photographs.
more, these melanomas tend to develop at the dermoepidermal junction, making their clinical detection easier than for those that develop in the dermis, as is often the case in large CMN (see Table 1) [42]. For CMN less than 1.5 cm in diameter the risk for developing melanoma is considered to be low enough that prophylactic removal of these lesions may not be warranted [43]. However, if a small CMN develops suspicious changes, becomes symptomatic, or is irregular, then an excisional biopsy may be justified. Besides being potential precursors to melanoma, some but not all studies have also shown smaller CMN and CNLN to be markers identifying a subset of the general population at increased risk for developing melanoma [3,44]. The management of CMN needs to be tailored for each patient. The nevus location, nevus size, cosmetic issues regarding the nevus or resultant surgical scars, risk of anesthesia, risk of surgery, psychologic implications, risk of melanoma, and risk of NCM all need to be taken into account in this decision-making process. Surgical removal of CMN, especially large or clinically atypical CMN, may lower the risk for developing melanoma; however, this has not been
confirmed in any controlled study. Analysis of previous studies, arranged chronologically by year of publication, reveals a progressive decline in the reported risk of melanoma developing in large CMN [15]. Study design, selection bias, awareness of the importance of ultraviolet protection, and the introduction of sunscreens may, in part, explain some of this apparent decline in the melanoma risk associated with large CMN. In addition, over the last 30 years, great strides have also been made in improving the safety of general anesthesia and surgical techniques. The routine use of tissue expanders and skin grafts has made removal of larger CMN more feasible and many patients (or their parents) elect to have their CMN excised to improve the cosmetic appearance or reduce the risk of melanoma. The fact that many patients undergo complete or partial removal of their large CMN may be another variable that can explain the decline in melanoma risk observed over time [35]. Most excisions, however, cannot eliminate the risk of melanoma completely because frequently it is impossible or impractical to remove every nevus cell and melanomas can develop from remnant nevus cells [30]. Furthermore, surgical excision does not eliminate the risk of developing extracutaneous melanomas as can occur in patients with NCM. If surgical excision is selected as the treatment of choice, ideally it should address the risk of malignant transformation, achieve satisfactory cosmetic results, and maintain adequate function. Improvements and developments of tissue engineered skin, tissue expanders, surgical techniques, and pharmaceutical and biologic agents to help wound healing (ie, platelet-derived growth factor, keratinocyte growth factor, epidermal growth factor, and cytokines) may someday make large excisions even more feasible [45 47]. Although other surgical interventions, such as curettage, dermabrasion, and lasers, may improve the cosmetic appearance of CMN, they do not address adequately the risk of developing melanoma from nevus remnants located in the dermis [48 52]. Furthermore, it may be difficult to detect changes indicative of melanoma, and problematic to differentiate between melanoma and pseudomelanoma (recurrent nevus) in CMN treated in this fashion [53]. Finally, some patients and their families have unrealistic expectations regarding the cosmetic outcome after surgery. These individuals should be informed that although surgery may sometimes improve the cosmetic appearance, frequently the patient is left with scars that can also be unattractive or uncomfortable. An alternative to prophylactic excision, especially for those nevi that are uniform, light-colored, eventextured, and without nodules, is periodic clinical
613
Fig. 10. Dermoscopy of the lesion in Fig. 9 showing the dermoscopic features of a congenital melanocytic nevus with homogeneous distribution of globules (cobblestone pattern) on the right side of the image. On the left side one observes irregular black dots, blue-gray areas, and thickened reticulation, all of which are commonly seen in melanoma.
surveillance. Patient and family involvement can aid in the early detection of melanoma by performance of monthly self skin examinations. Any changes suspicious for melanoma should undergo biopsy and be treated appropriately. To avoid the false-positive diagnosis of melanoma, which can result in unnecessary surgery or chemotherapy, it is recommended that all biopsies be evaluated by a pathologist experienced in the evaluation of pigmented lesions. It is important to remember that erosions and proliferative nodules, which are often seen in large CMN within the first few weeks of the patients life, are not necessarily signs of cancer and most heal or disappear over time [31,54,55]. Baseline photographs of the nevus can be used for comparison during subsequent skin examinations to detect subtle changes within these nevi [56,57]. Changing lesions can be evaluated further by in vivo techniques, such as dermoscopy [58] or confocal laser microscopy [59]. Dermoscopy may help detect small foci of melanoma within a CMN, helping clinicians and pathologists in evaluating the most appropriate area (Fig. 10) [42,60]. This may potentially avoid sampling error resulting in the falsenegative diagnosis of malignancy. Confocal laser microscopy, which uses a near infrared laser beam, is a new experimental in vivo imaging technique that provides near histologic horizontal microscopic section images of the epidermis, dermoepidermal junction, and papillary dermis. This technique may prove to be useful in evaluating CMN suspect for having developed melanoma (Fig. 11) [61,62]. It is important, however, to appreciate that both dermoscopy and confocal microscopy can evaluate skin to the depth of the papillary dermis. These techniques can
provide useful information when evaluating smaller or superficial CMN (see Table 1) because most melanomas that develop in these nevi occur at the dermoepidermal junction. On the other hand, both dermoscopy and confocal laser microscopy have limited use in the evaluation of changes occurring in the dermis of large or deeply penetrating CMN. Lesions suggestive of melanoma on clinical inspection, dermoscopy, or confocal laser microscopy should promptly undergo biopsy. It is important to inform patients or caregivers that sunburns increase the risk of developing melanoma within nevi [63]. All patients should be instructed on proper sun avoidance and sun-protection techniques. CMN present on exposed skin can be concealed with cover-up foundation and cosmetics, many of which contain sunscreens. CMN on nonexposed skin can be protected from ultraviolet light by use of sunprotective clothing and sun-protective swimwear. Future developments in chemoprevention with improved sunscreen formulations, development of compounds that help repair ultraviolet-damaged DNA [64], and the development of compounds that can prevent ultraviolet damage may also help decrease the risk of developing melanoma [65]. Patients with
Fig. 11. (A) Confocal image of the congenital melanocytic nevus portion of the lesion in Fig. 9. The normal rete ridge pattern at the dermoepidermal junction is preserved and there are nests of round nevomelanocytes in the papillary dermis, a feature often seen in congenital melanocytic nevus. (B) Confocal image of the melanoma portion of the lesion in Fig. 9. Notice the loss of the normal rete ridge pattern; single cells predominate over nests; and pagetoid spread of melanocytes. The arrow points to one of the irregular, pagetoid, and dendritic melanocytes located in the epidermis. The combination of large dendritic melanocytes in the epidermis in conjunction with a loss of the normal architecture is frequently observed in melanoma viewed under confocal laser microscopy.
614
large CMN should be informed that the skin overlying the nevus may be fragile, and often lacks the normal skin barrier functions found in normal skin. The involved skin easily can become xerotic, which in turn can cause pruritus. Diligent use of moisturizers applied on a regular basis may help control these symptoms. For those patients at increased risk for neurocutaneous melanocytosis, obtaining a screening MRI of the brain should be considered [66 68]. Ideally, the screening MRI should be performed within the first 4 to 6 months of life before normal myelination of the brain obscures melanin deposits (James A. Barkovich, personal communication). It is important to note that as many as 23% of neurologically asymptomatic patients with large CMN may have a positive MRI scan revealing changes suggestive of NCM, but a very small percentage of them actually develop symptomatic NCM [69]. On the other hand, a normal MRI does not exclude the possibility of developing symptoms, although the chance of this is small [70]. Some physicians recommend obtaining serial MRIs in asymptomatic patients at risk for developing NCM [71]. Thorough neurologic examinations can also be used in the assessment of patients at risk for developing NCM. Serial neurologic examinations may be a substitute for MRIs in following these individuals. There are some physicians and patients who question the necessity of obtaining screening MRIs because there are no treatments available for asymptomatic NCM. Knowledge of a positive MRI, however, may affect the management decisions regarding the cutaneous nevi, and a negative MRI may help alleviate some anxieties. Patients with symptomatic NCM, on the other hand, should have MRIs performed to determine whether they may be candidates for medical intervention (ie, ventriculoperitoneal shunt placement; surgical excision; radiation therapy; or experimental therapy with interferon, chemotherapy, or retinoids) to help alleviate their symptoms [72]. In conclusion, the care of patients with CMN often requires a multidisciplinary approach involving pediatricians, family physicians, internists, dermatologists, psychologists, plastic surgeons, neurologists, and radiologists. The cosmetic and psychosocial issues combined with the knowledge of the increased risk of developing melanoma or NCM is a huge burden that many of these patients and their families have to carry. It is important for physicians to help these patients and families come to terms with these issues. Furthermore, patients and families should be reminded that although melanoma, NCM, or other complications can develop, most affected individuals do not develop any complications. There are many
healthy, happy, functional adults with large, small, and multiple CMN alive today (see Fig. 8). Lastly, patient advocacy groups can be of great help to patients and their family members. (Nevus Outreach, Inc. is one such nevus support group. They can be found on the web at www.nevus.org or by phone at 877-426-3887.)
References
[1] Rhodes AR. Benign neoplasias and hyperplasias of melanocytes. In: Freedberg IM, Eisen AZ, Wolff K, et al, editors. Dermatology in general medicine. New York: McGraw-Hill; 1999. p. 1018 59. [2] Mizushima J, Nogita T, Higaki Y, et al. Melanocytes in the dermis: do dermal melanocytes of acquired dermal melanocytosis exist from birth? Br J Dermatol 1998; 139:349 50. [3] Kopf AW, Levine LJ, Rigel DS, et al. Congenital -au-lait spots in nevus-like nevi, nevi spili, and cafe patients with malignant melanoma. J Dermatol Surg Oncol 1985;11:275 80. [4] Cribier BJ, Santinelli F, Grosshans E. Lack of clinicalpathological correlation in the diagnosis of congenital naevi. Br J Dermatol 1999;141:1004 9. [5] Walton RG, Jacobs AH, Cox AJ. Pigmented lesions in newborn infants. Br J Dermatol 1976;95:389 96. [6] Lorenz S, Maier C, Segerer H, et al. Skin changes in newborn infants in the first 5 days of life. Hautarzt 2000;51:396 400. [7] Rivers JK, MacLennan R, Kelly JW, et al. The Eastern Australian childhood nevus study: prevalence of atypical nevi, congenital nevus-like nevi, and other pigmented lesions. J Am Acad Dermatol 1995;32: 957 63. [8] Fung WK, Lo KK. Prevalence of skin disease among school children and adolescents in a Student Health Service Center in Hong Kong. Pediatr Dermatol 2000; 17:440 6. [9] Kopf AW, Levine LJ, Rigel DS, et al. Prevalence of -au-lait congenital-nevus-like nevi, nevi spili, and cafe spots. Arch Dermatol 1985;121:766 9. [10] Sigg C, Pelloni F, Schnyder UW. Frequency of con -au-lait spots and their genital nevi, nevi spili and cafe relation to nevus count and skin complexion in 939 children. Dermatologica 1990;180:118 23. [11] Rhodes AR, Albert LS, Weinstock MA. Congenital nevomelanocytic nevi: proportionate area expansion during infancy and early childhood. J Am Acad Dermatol 1996;34:51 62. [12] Dawson HA, Atherton DJ, Mayou B. A prospective study of congenital melanocytic naevi: progress report and evaluation after 6 years. Br J Dermatol 1996;134: 617 23. [13] Tokura Y, Yamanaka K, Wakita H, et al. Halo congenital nevus undergoing spontaneous regression: involvement of T-cell immunity in the involution and presence
A.A. Marghoob / Dermatol Clin 20 (2002) 607616 of circulating anti-nevus cell IgM antibodies. Arch Dermatol 1994;130:1036 41. DeDavid M, Orlow SJ, Provost N, et al. Neurocutaneous melanosis: clinical features of large congenital melanocytic nevi in patients with manifest central nervous system melanosis. J Am Acad Dermatol 1996;35: 529 38. Marghoob AA, Schoenbach SP, Kopf AW, et al. Large congenital melanocytic nevi and the risk for the development of malignant melanoma: a prospective study. Arch Dermatol 1996;132:170 5. Marghoob A, Kopf A, Bittencourt F. Moles present at birth: their medical significance. Skin Cancer Foundation Journal 1999;17:36 98. Takayama H, Nagashima Y, Hara M, et al. Immunohistochemical detection of the c-met proto-oncogene product in the congenital melanocytic nevus of an infant with neurocutaneous melanosis. J Am Acad Dermatol 2001;44:538 40. Cramer SF. The melanocytic differentiation pathway in congenital melanocytic nevi: theoretical considerations. Pediatr Pathol Lab Med 1988;8:253 65. Kos L, Aronzon A, Takayama H, et al. Hepatocyte growth factor/scatter factor-Met signaling in neural crest-derived melanocyte development. Pigment Cell Res 1999;12:13 21. Takayama H, LaRochelle WJ, Anver M, et al. Scatter factor/hepatocyte growth factor as a regulator of skeletal muscle and neural crest development. Proc Natl Acad Sci U S A 1996;93:5866 71. Otsuka T, Takayama H, Sharp R, et al. C-met autocrine activation induces development of malignant melanoma and acquisition of the metastatic phenotype. Cancer Res 1998;58:5157 67. Natali PG, Nicotra MR, DiRenzo MF, et al. Expression of the c-met/HGF receptor in human melanocytic neoplasms: demonstration of the relationship to malignant melanoma tumour progression. Br J Cancer 1993;68: 746 50. Marghoob A, Bittencourt F, Kopf A, Bart R. Large congenital melanocytic nevi. Curr Probl Dermatol 2000; 146 52. DeDavid M, Orlow SJ, Provost N, et al. A study of large congenital melanocytic nevi and associated malignant melanomas: review of cases in the New York University Registry and the world literature. J Am Acad Dermatol 1997;36:409 16. Koot HM, DeWaard-van der Spek F, Peer CD, et al. Psychosocial sequelae in 29 children with giant congenital melanocytic naevi. Clin Exp Dermatol 2000;25: 589 93. Ceballos P, Ruiz-Maldonado R, Mihm M. Melanoma in children. N Engl J Med 1995;332:656 62. Kanzler MH, Marz-Gernhard S. Primary cutaneous malignant melanoma and its precursor lesions: diagnostic and therapeutic overview. J Am Acad Dermatol 2001;45:260 76. Barnhill RL, Fleischli M. Histologic features of congenital melanocytic nevi in infants 1 year of age or younger. J Am Acad Dermatol 1995;33:780 5.
615
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26] [27]
[28]
[29] Marghoob AA, Schoenbach SP, Kopf AW, et al. Large congenital melanocytic nevi and the risk of developing malignant melanoma: a prospective study and review of the world literature. J Invest Dermatol 1995;104:563. [30] Rhodes AR, Wood WC, Sober AJ, et al. Nonepidermal origin of malignant melanoma associated with giant congenital nevocellular nevus. Plast Reconstr Surg 1981;67:782 90. [31] Mancianti M, Clark WH, Hayes FN, et al. Malignant melanoma simulants arising in congenital melanocytic nevi do not show experimental evidence for a malignant phenotype. Am J Pathol 1990;136:817 29. [32] Leech S, Bell H, Lewis-Jones S, et al. Giant congenital pigmented naevi with proliferating nodules occurring in two neonates. Br J Dermatol 1999;141:32 3. [33] Kovach R, Hochberg J, Ardenghy M, et al. Diagnosing and treating congenital melanocytic nevus simulating malignant melanoma. W V Med J 1996;92:191 3. [34] Bastian BC, LeBoit PE, Pinkel D. Genomic approaches to skin cancer diagnosis. Arch Dermatol 2001;137: 1507 11. [35] Bittencourt FV, Marghoob AA, Kopf AW, et al. Large congenital melanocytic nevi and the risk for development of malignant melanoma and neurocutaneous melanocytosis. Pediatrics 2000;106:736 741. [36] Kodonaga JN, Frieden IJ. Neurocutaneous melanosis: definition and review of the literature. J Am Acad Dermatol 1991;24:747 755. [37] Reyes-Mugica M, Chou P, Byrd S, et al. Nevomelanocytic proliferations in the central nervous system of children. Cancer 1993;73:227 85. [38] Rhodes AR, Melski JW. Small congenital nevocellular nevi and the risk of cutaneous melanoma. J Pediatr 1982;100:219 24. [39] Sahin S, Levin L, Kopf A, et al. Risk of melanoma in medium-sized congenital melanocytic nevi: a followup study. J Am Acad Dermatol 1998;39:428 33. [40] Swerdlow A, English J, Qiao Z. The risk of melanoma in patients with congenital nevi: a cohort study. J Am Acad Dermatol 1995;32:595 9. [41] Illig C, Weidener F, Hundeiker H, et al. Congenital nevi less than or equal to 10 cm as precursors to melanoma: 52 cases, a review, and a new conception. Arch Dermatol 1985;121:1274 81. [42] Betti R, Inselvini E, Vergani R, et al. Small congenital nevi associated with melanoma: case reports and considerations. J Dermatol 2000;27:583 90. [43] Bono A, Bartoli C, Zurrida SM, et al. Lets stop worrying about pigmented skin lesions in children. Eur J Cancer 1994;30A:417. [44] Tucker MA, Halpern A, Holly EA, et al. Clinically recognized dysplastic nevi: a central risk factor for cutaneous melanoma. JAMA 1997;277:1439 44. [45] Eaglstein W, Falanga V. Tissue engineering for skin: an update. J Am Acad Dermatol 1998;39:1007 10. [46] Cohen MA, Eaglstein WH. Recombinant human platelet-derived growth factor gel speeds healing of acute full-thickness punch biopsy wounds. J Am Acad Dermatol 2001;45:857 62.
616
A.A. Marghoob / Dermatol Clin 20 (2002) 607616 [61] Charles CA, Lee G, Marghoob AA, et al. In vivo confocal scanning laser microscopy of congenital melanocytic nevi that are suspicious for having developed malignant melanoma. Melanoma Res 2001;11:S183. [62] Busam KJ, Charles C, Lee G, et al. Morphologic features of melanocytes, pigmented keratinocytes, and melanophages by in vivo confocal scanning laser microscopy. Mod Pathol 2001;14:862 8. [63] Carli P, Massi D, Santucci M, et al. Cutaneous melanoma histologically associated with a nevus and melanoma de novo have a different profile of risk: results from a case-control study. J Am Acad Dermatol 1999; 40:549 57. [64] Yarosh D, Klein J, OConnor A, et al. Effect of topically applied T4 endonuclease V in liposomes on skin cancer in xeroderma pigmentosum: a randomized study. Lancet 2001;357:926 9. [65] Salti GI, Kichina JV, DasGupta TK, et al. Betulinic acid reduces ultraviolet-C-induced DNA breakage in congenital melanocytic naeval cells: evidence for a potential role as a chemopreventive agent. Melanoma Res 2001;11:99 104. [66] Frieden IJ, Williams ML, Barkovich AJ. Giant congenital melanocytic nevi: brain magnetic resonance findings in neurologically asymptomatic children. J Am Acad Dermatol 1994;31:423 9. [67] Barkovich AJ, Frieden IJ, Williams ML. MR of neurocutaneous melanosis. AJNR Am J Neuroradiol 1994; 15:859 67. [68] Kinsler VA, Aylett SE, Coley SC, et al. Central nervous system imaging and congenital melanocytic naevi. Arch Dis Child 2001;84:152 5. [69] Foster RD, Williams ML, Barkovich AJ, et al. Giant congenital melanocytic nevi: the significance of neurocutaneous melanosis in neurologically asymptomatic children. Plast Reconstr Surg 2001;107:933 41. [70] Ruiz-Maldonado R, del Rosario BM, Hidalgo-Galvan LR, et al. Giant congenital melanocytic nevi, neurocutaneous melanosis and neurological alterations. Dermatology 1997;195:125 8. [71] Gondo K, Kia R, Tokunaga Y, et al. Age-related changes of the MR appearance of CNS involvement in neurocutaneous melanosis complex. Pediatr Radiol 2000;30:866 8. [72] Yoshioka S, Miyayama H, Ishihara A, et al. Neurocutaneous melanosis: a case report. No To Shinkei 1994;46:279 84.
[47] Koveker GB. Growth factors in clinical practice. Int J Clin Pract 2000;54:590 3. [48] Casanova D, Bardot J, Andrac-Meyer L, et al. Early curettage of giant congenital naevi in children. Br J Dermatol 1998;138:341 5. [49] Bohn J, Svensson H, Aberg M. Dermabrasion of large congenital melanocytic naevi in neonates. Scand J Plast Reconstr Surg Hand Surg 2000;34:321 6. [50] Arons MS. Successful treatment of a giant congenital melanocytic naevus with high-energy pulsed CO2 laser. Br J Plast Surg 1998;51:570 1. [51] Nelson JS, Kelly KM. Q-switched laser treatment of a congenital melanocytic nevus. Dermatol Surg 1999; 25:274 6. [52] Duke D, Byers HR, Sober AJ, et al. Treatment of benign and atypical nevi with the normal-mode ruby laser and the Q-switched ruby laser: clinical improvement but failure to completely eliminate nevomelanocytes. Arch Dermatol 1999;135:290 6. [53] Dummer R, Kempf W, Burg G. Pseudo-melanoma after laser therapy. Dermatology 1998;197:71 3. [54] Giam YC, Williams ML, Leboit PE, et al. Neonatal erosions and ulcerations in giant congenital melanocytic nevi. Pediatr Dermatol 1999;16:354 8. [55] Borbujo J, Jara M, Cortes L, et al. A newborn with nodular ulcerated lesion on a giant congenital nevus. Pediatr Dermatol 2000;17:299 301. [56] Rhodes AR. Intervention strategy to prevent lethal cutaneous melanoma: use of dermatologic photography to aid surveillance of high-risk persons. J Am Acad Dermatol 1998;39:262 7. [57] Oliveria SA, Nehal KS, Marghoob AA, et al. Use of photography and dermoscopy in dermatology clinics: a survey of dermatology residency programs in the United States. Melanoma Res 2001;11:S147. [58] Bafounta M, Beauchet A, Aegerter P, et al. Is dermoscopy (epiluminescence microscopy) useful for the diagnosis of melanoma? Results of a meta-analysis using techniques adapted to the evaluation of diagnostic tests. Arch Dermatol 2001;137:1343 50. [59] Langley R, Rajadhyaksha M, Dwyer P, et al. Confocal scanning laser microscopy of benign and malignant melanocytic skin lesions in vivo. J Am Acad Dermatol 2001;45:365 76. [60] Bauer J, Metzler G, Rassner G, et al. Dermatoscopy turns histopathologists attention to the suspicious area in melanocytic lesions. Arch Dermatol 2001;137: 1338 40.
Dermatol Clin 20 (2002) 617 628
The dilemma of the dysplastic nevus

Thomas G. Salopek, MD
Division of Dermatology and Cutaneous Sciences, University of Alberta, 2-125 Clinical Sciences Building, Edmonton, Alberta, Canada, T6G 2G3
There are few areas in dermatology that provoke as much controversy and discussion as that of dysplastic nevus. There are two entrenched camps with diametrically opposed ideas concerning the validity of this lesion as a distinct entity from banal melanocytic nevus, or its potential for progression to melanoma. Even the name raises considerable debate. In 1978, Clark and colleagues [1] at the University of Pennsylvania described for the first time a clinicopathologic entity, which identified patients at increased risk for melanoma. They originally coined the term B-K mole syndrome, based on the first initials of the surnames of the probands. Since then numerous names have been attributed to this lesion and its corresponding syndrome, including dysplastic nevus (and syndrome) [2]; atypical mole (and syndrome) [3]; familial atypical multiple mole-melanoma syndrome [4]; and Clarks nevus (and syndrome) [5]. In addition, the following histologic descriptors have been used: pagetoid melanocytic proliferation [6], nevus with architectural disorder [7], atypical melanocytic proliferation [8], and intraepithelial melanocytic neoplasia [9]. Because of variable definitions for dysplasia, the National Institute of Health (NIH) issued a Consensus Statement paper in 1992, recommending the following: the term dysplastic nevus be replaced with atypical moles, that histologically they be diagnosed as nevi with architectural disorder, and that the syndrome for melanoma-prone families be called familial atypical mole and melanoma (FAMM) syndrome [7]. These recommendations have never been accepted fully by the medical community, and the term that has persisted the longest and is most frequently used by clinicians, pathologists, and re-
searchers is dysplastic nevus (and syndrome), and it is the term that is used throughout this article. In 1995, the author and colleagues extensively reviewed the topic of dysplastic nevi (atypical moles) [10]. Since then there has been a wealth of new information that enhances the understanding of this lesion in terms of its genetics and potential molecular basis. Distinct and reliable clinical and histologic features have been delineated. The role for dermoscopy, photographic surveillance, genetic mapping and counseling, chemoprevention, and nevi removal has also been addressed. This article summarizes some of these advances and discusses how these patients can be managed.
Genetics of dysplastic nevi It has been well documented in multiple retrospective and prospective studies that patients with dysplastic nevi are at increased risk for developing melanoma (particularly if there is a family history of melanoma); as such, the lesion represents a useful means of identifying patients at risk for this cancer [10,11]. Because dysplastic nevi are strongly associated with melanoma, it is logical to presume that it might be linked to one of the putative melanoma susceptibility genes [12]. There is mounting evidence that melanoma is a potentially hereditary disorder whose propensity to develop is dependent on the inheritance of one of four postulated susceptibility genes (Table 1) [11]. The strongest evidence for a melanoma susceptibility gene has focused on a gene located on chromosome 9 [13 17]. The locus, 9p21, encodes for a tumor suppressor gene, p16, also known as cyclin-dependent kinase 2A (CDKN2A). CDKN2A results in the arrest
E-mail address: tsalopek@ualberta.ca (T.G. Salopek).
618 Table 1 Melanoma susceptibility genes Name CMM1 CMM2 CMM3 CMM4
T.G. Salopek / Dermatol Clin 20 (2002) 617628
Chromosome 1p36 9p21 12q14 9p21
Gene product Unknown CDKN2A (p16) CDK4 Possibly p19
Presumed function Unknown Tumor suppressor gene Oncogene Tumor suppressor gene
From Greene MH. The genetics of hereditary melanoma and nevi. 1998 update. Cancer 1999;86(Suppl 11):2464 77; with permission.
of cells in G1 phase because of binding and inhibition of CDK4 and 6 [18]. Mutations in CDKN2A result in uncontrolled cell division. It has been postulated that melanoma and dysplastic nevi represent pleiotropic manifestations of the same gene [16,19]; therefore, one would expect a link between the dysplastic nevus phenotype and CDKN2A. Unfortunately, studies have been inconsistent in this regard. Considering that melanocytic nevi represent a heterogenous group and their development is caused by a complex and poorly understood interaction of genetic and environmental factors, it should not come as a surprise that conflicting results have been reported. There is little doubt that genetic factors influence the development of nevi. It is known that there is a stronger correlation of total nevi numbers among monozygotic than dizygotic twins [20]. Crijns et al [21] reported a fourfold increase in incidence of dysplastic nevi in relatives of probands with dysplastic nevi and melanoma as compared with controls. Newton et al [22] reported similar findings, specifically that 39% of relatives with dysplastic nevus syndrome also had the syndrome, whereas only 15% of melanoma patients and 2% of controls had this phenotype. One of the first studies on the role of CDKN2A mutations and the development of melanoma demonstrated a poor correlation between mutations in the gene and the presence of dysplastic nevi [23]. Similarly, Dutch researchers reported high 9p21-linked Log of the odds (LOD) scores in patients with florid dysplastic nevus syndrome and melanoma (ie, FAMM syndrome); however, inclusion of less severe cases of dysplastic nevus syndrome reduced the LOD scores [24]. This finding and the fact that many family members from their melanoma-prone families had dysplastic nevi without CDKN2A mutations suggested to the authors that dysplastic nevi are not caused by the CDKN2A melanoma gene [24]. Lastly, Spanish researchers failed to demonstrate a linkage between dysplastic nevi and CDKN2A mutations [25]. The importance of this negative finding is diminished by the fact that the study examined a single, small melanoma-prone family, and by their
observations in a subsequent study that CDKN2A mutations seem to play a small role in the development of melanoma in Spanish melanoma families [26]. Using microdissection techniques, allelic deletions of p16 (9p21) have been shown to be more frequent in dysplastic nevi than benign nevi, suggesting that deletion of p16 may play a role in the development of these lesions [27]. In segregation and linkage analysis, Goldstein et al [28] found gene-covariate interaction with dysplastic nevi and CDKN2A in American melanoma-prone families. The odds ratio for dysplastic nevi, however, was paradoxically greatest in subjects without CDKN2A mutations. In contrast, Bishop et al [29] noted in five British families with dysplastic nevus syndrome and documented germline mutations in CDKN2A that family members with the syndrome were three times more likely to be mutant gene carriers than their relatives who did not have the syndrome, supporting the notion that CDKN2A is nevogenic. Despite this, there was considerable overlap between gene carriers and noncarriers such that the dysplastic nevus phenotype did not differentiate mutant carriers adequately from those with a normal gene. The authors conclude that it is prudent to consider all family members from melanoma-prone families as potential carriers, and at increased risk for melanoma [30]. In five Swedish melanoma kindreds, there was highly significant correlation between the presence of CDKN2A mutations and development of melanoma [31]. Similarly, carriers of CDKN2A mutations were much more likely to exhibit the dysplastic nevus syndrome phenotype than individuals without mutations [31]. In summary, genetic factors seem to be important in the expression of the dysplastic nevus syndrome phenotype, and CDKN2A mutations may be causally involved in some but not all forms of familial melanoma. Other genes including modifying genes that co-segregate with CDKN2A and increase the penetrance of the latter may have a role in the expression of this phenotype [29]. The importance of the mutations in this or other genes for sporadic, dysplastic nevi has yet to be explored.
619
Biology The basis for dysplastic nevus as a distinct clinicopathologic entity has its origins from its association with familial melanoma. Individuals from such melanoma-prone families exhibit a high propensity for developing melanoma in addition to having numerous melanocytic tumors that are both histologically and clinically distinct (ie, dysplastic nevi) [1,4]. Although most researchers and clinicians have little reservation in accepting the concept of dysplastic nevi in melanoma-prone families, the importance of this lesion in the absence of familial melanoma (ie, sporadic dysplastic nevi) is not clear. To complicate matters, most dysplastic nevi in either setting rarely eventuate into melanoma, and most melanomas seem to arise de novo, and not in association with a pre-existing melanocytic nevus [32,33]. If dysplastic nevi are potential precursors to melanoma, one would expect to find genomic and posttranscriptional changes that place the lesion somewhere along the tumor progression pathway from benign nevus to melanoma. Numerous studies have shown this to be the case, with distinct karyotypic, immunologic, biochemical, and biologic differences noted [34 47]. It has been difficult to differentiate whether the changes seen in dysplastic nevi are of causal importance, or are simply an epiphenomenon and are merely an expression of genetic mutations. Rather than summarizing the aforementioned findings, the author focuses on the genetic changes that occur and are likely the basis for this condition. Early researchers of the dysplastic nevus syndrome had noticed karyotypic changes in cultured normal fibroblasts or lymphocytes obtained from such patients [36,40]. Unfortunately, limited molecular biology techniques and inability to culture nevocytes curtailed progress in this area for years. Despite these limitations, it was proposed that dysplastic nevus syndrome might represent a chromosomal instability disorder [36]. Lynch et al [48] were one of the first groups to demonstrate chromosomal instability in patients with FAMM syndrome. Studying two extended FAMM kindreds, they observed marked chromosome instability as evidenced by increased frequency of chromosomal rearrangements (ie, translocations, deletions, and inversions) in cell cultures from dysplastic nevi and normal skin fibroblasts [48]. The fact that changes where seen in fibroblasts and nevocytes suggested to the authors that FAMM syndrome may represent a basic genetic instability disorder with possible susceptibility to other cancers. Except for a
possible link with pancreatic cancer [49,50], this does not seem to be the case [51,52]. It is likely that environmental factors (eg, solar exposure) are required to produce enough genetic mutations to allow for progression to malignancy. As to why only melanocytic and not nonmelanocytic skin tumors develop remains a mystery; however, the relative resistance of melanocytes to ultraviolet radiation (UVR) induced apoptosis may have some role in this regard [53]. Two reasons have been identified as possible sources of the chromosomal instability seen in this condition. One theory holds that it is due to UV hypersensitivity, whereas the other theory proposes inherent chromosomal instability caused by a process known as microsatellite instability. The theories are not mutually exclusive and it is possible that each abnormality influences the other. Preliminary reports from the early 1980s suggested a slight increase in UVR sensitivity in lymphoblasts and fibroblasts from melanoma patients of high-risk families [54,55]. There was also evidence of increased mutability secondary to exposure to x-ray radiation and a carcinogen (ie, a UV-mimetic chemical substance) [56,57]. Using double- and single-stranded DNA viruses as probes, it is possible to investigate repair processes in cells [58]. Moriwaki et al [59] examined the frequency of mutations after UVR exposure (using germicidal lamps [ie, UVC]) in the plasmid shuttle vector pSP189, transfected and subsequently recovered and assayed from 31 lymphoblastoid cell lines from six familial melanoma kindreds, as compared with 16 normal control cell lines. They reported a markedly increased frequency of mutations of UV-irradiated plasmids isolated from cell lines of 13 of 13 patients with melanoma and dysplastic nevi, and five of eight cell lines from patients with dysplastic nevi alone. Similar results were reported by Abrahams et al [58] using cultured skin fibroblasts from dysplastic nevus syndrome patients transfected with UVC-irradiated herpes simplex virus type 1. The ability of the host cell to repair and reactivate UVR-damaged herpes simplex virus is an indirect measure of the DNA repair capacity of the cell. Although there was considerable heterogeneity in host cell reactivation of the viral probe, in general, cell lines from dysplastic nevus syndrome patients exhibited decreased viral reactivation as compared with controls. Lastly, Noz et al [60] using a single cell gel electrophoresis assay (comet assay) to study environmentally relevant UVB-induced DNA damage directly in pigment cells noted a substantial sensitivity of nevus cells from dysplastic nevi to UVB irradiation as compared with normal melanocytes and common melanocytic nevi. Other researchers
620
have been unable to confirm these observations using different methods for assessing DNA repair systems [61,62]. Recently it has been proposed that microsatellite instability may have a role in the malignant transformation of cells, including melanocytes [63]. Microsatellites are repetitive DNA sequence arranged in tandem from one to four DNA bases long scattered throughout the human genome. They are stably inherited and are highly conserved from one generation to the next, and in different cells from the same individual. Although their exact function is uncertain, they are thought to be either binding sites for DNA topoisomerases or are tumor promoters [63]. Microsatellite instability refers to polymorphism of these repetitive nucleotide motifs, and is presumed to arise from accumulated frameshift mutations that occur because of uncorrected misalignments between template and daughter DNA strands during replication. The latter arises secondary to mismatch repair defects resulting in a general mutator phenotype. Hussein et al [63] have recently demonstrated microsatellite instability in melanomas and dysplastic nevi but not benign nevi. In a subsequent paper, these researchers report a progressive decrease in the expression of mismatch repair proteins in melanocytic tumors from benign nevi, dysplastic nevi, and melanoma, which may account for the microsatellite instability that they had noted earlier [64].
Clinical features A dysplastic nevus is a mole that shares features common to melanoma. It was originally defined as a mole larger than 5 mm in diameter, variegated in pigmentation, and irregular in outline [65]. Since the initial description, additional refinements have been made to the clinical recognition of a dysplastic nevus. Kelly et al [66] define a dysplastic nevus as a mole with a macular component and showing at least three of five criteria: (1) ill-defined or (2) irregular border, (3) irregularly distributed pigmentation, (4) background of erythema, and (5) size greater than 5 mm. Tucker et al [67] have used similar clinical features, notably the mole should be greater than 5 mm in size and minimally elevated. In addition, two of three other features should be seen, including variable pigmentation, irregular or asymmetric outline, and indistinct borders. It is clear from the literature that most clinicians are capable of recognizing a dysplastic nevus; however, what constitutes the dysplastic nevus syndrome remains a highly contentious issue. The NIH con-
sensus paper lists the following criteria for a diagnosis of FAMM syndrome: (1) occurrence of melanoma in one or more first- or second-degree relatives; (2) large number of moles (often greater than 50), several of which are atypical; and (3) moles that demonstrate distinct histologic features [7]. In contrast, the Dutch Working Group has delineated five criteria for diagnosis of clinically dyplastic nevi: (1) greater than or equal to 5 mm in diameter, (2) vague border, (3) asymmetric shape, (4) irregular pigmentation, and (5) red hue [68]. The presence of three or more of these five criteria is sufficient for a diagnosis of dysplastic nevus. Dysplastic nevus syndrome is diagnosed if a patient has melanoma and one or more clinically dysplastic nevi. The Dutch group reserves the term dysplasia for histologic diagnosis, and limits the term atypical nevus to clinical practice. An increasing number of researchers are beginning to accept the scoring system devised by Newton [12,69]. This scoring system considers five parameters: (1) one hundred or more nevi 2 mm or larger, (2) two or more atypical nevi, (3) one or more nevi on the anterior scalp, (4) one nevus on the buttocks, or two or more on the dorsum of the feet, and (5) one or more iris nevi. Each feature generates one point in the scoring system. Individuals with a score of three or higher are considered to have the dysplastic nevus syndrome phenotype. The scoring system has been validated, and it has been shown that nonspecialist health care professionals easily can be instructed on the use of the scoring system with the intent of reliably identifying individuals at increased risk for developing melanoma [70]. Based on the results of their recent study examining the penetrance of CDKN2A mutations in patients from melanoma families, they have revised their scoring system such that they have forgone the need for scalp and iris nevi, and the phenotype is said to be present if a total score of two or higher is obtained [29]. This new scoring system, although simpler than the original, has yet to be validated.
Histology The primary reason for controversy concerning the dysplastic nevus has been a lack of consensus regarding the histologic criteria required to make this diagnosis [71]. Ackerman [72] has been the most vociferous proponent against the validity of the dysplastic nevus as a distinct entity, and has expressed his disappointment that hundreds of authors of thousands of articles about melanocytic dysplasia, dysplastic nevus, and dysplastic nevus syndrome have been unable, in more than 20 years, to agree
621
about the definition of those terms. In recent years, there have been several well-designed studies that refute these claims. An interobserver study by the pathology subgroup of the EORTC Malignant Melanoma Cooperative Group found that it is possible to diagnose dysplastic nevi reliably by histopathologic examination [73]. The authors define 20 histologic criteria that permit the differentiation of dysplastic nevi from other pigmented lesions. Using these criteria there was a high degree of concordance of diagnosis, with the most reproducible criteria for diagnosing dysplastic nevus being the presence of irregular nests, lymphohistiocytic infiltrate, marked junctional proliferation, and large nuclei. The presence of at least three of the four features resulted in a sensitivity, specificity, positive, and negative predicative values of 0.86, 0.91, 0.96, and 0.73, respectively. Similarly, the Dysplastic Nevus Panel found that melanocytic dysplasia can be graded reproducibly using predetermined histologic criteria: size of nucleus, variability in shape and size of the nucleus, chromatin staining, and features of the nucleolus and cytoplasm [74]. In this study, a random sample of 112 melanocytic tumors (including banal and dysplastic nevi, and melanoma) were graded by a panel of five dermatopathologists and two melanoma specialists using an arbitrary five-point scale to represent the spectrum of dysplasia (1 for no dysplasia and 5 for melanoma). Interrater reliability as measured by intraclass and Pearson correlation coefficients were 0.67 (95% CI, 0.59 to 0.73) and 0.67 to 0.84, respectively. Individual panel members were within one grade (on the five-point scale) of the mean of the study panel 88% of the time, and rarely by two or more grades (3%). The authors conclude that it is possible to diagnosis melanocytic dysplasia by well-defined and agreed on criteria, and that its severity graded with reasonable although imperfect reliability [74]. To diagnose dysplastic melanocytic nevi consistently, and with considerable reliability, both architectural disorder and cytologic atypia should be present. In a study of 166 dysplastic nevi, Shea et al [75] found a positive correlation between the degrees of architectural disorder and cytologic atypia. Although several of the features demonstrated significant interdependence, it was not uncommon to find cases that exhibited a high score for one parameter and a low score for the other, and vice versa. As a consequence, quantification of both parameters should be done to classify these lesions appropriately as mild, moderate, or severely dysplastic. Pozo et al [76] similarly reported the importance of cytologic and architectural features in grading dysplastic nevi.
In their retrospective study of 123 dysplastic nevi (40 associated with the dysplastic nevus syndrome), they identified by multivariate analyses three nuclear variables (pleomorphism, chromatin pattern, and nucleolus prominence) and two architectural features (junctional symmetry and presence and location of suprabasilar melanocytes) that were useful in stratifying dysplastic into either severe or mild-moderate groups (diagnostic accuracy 99.5%). The aforementioned histologic criteria did not permit differentiation of sporadically occurring dysplastic nevi from those arising in individuals with the dysplastic nevus syndrome; as such the diagnosis of the latter remains a clinicopathologic one. This should not come as a surprise considering the prevalence of dysplastic changes in even clinically normal-appearing nevi [77]. The purpose of histologic evaluation is not to confirm a diagnosis of dysplastic nevus syndrome but to exclude melanoma [78]. Critics of the dysplastic nevus have argued that if the lesion exists as a distinct entity, then there should be a strong correlation between clinical atypia and histologic dysplasia. Several studies have in fact confirmed that there is considerable correlation between the clinical and histologic appearance [79 81]. Conversely, others have found dysplastic changes to be so exceedingly common that they have concluded that the lesion does not exist as a distinct entity [77,82 84]. Dismissing the dysplastic nevus because the gold standard (ie, histologic evaluation) shows considerable overlap with other melanocytic tumors is analogous to refuting the existence of atopic dermatitis because it shares pathologic features with other dermatitides. Researchers have repeatedly stated that the diagnosis is based on clinical and pathologic findings.
Management Management of dysplastic nevi has centered on the following issues: regular follow-up facilitated by the use of diagnostic aids for the early detection of melanoma (eg, dermoscopy, total body photography, and digital imaging devices); biopsy of any suspicious lesions; prevention against malignant transformation through sun avoidance and protection; and nevi reduction to prevent melanoma or simply for cosmetic reasons. Follow-up Patients with dysplastic nevi should be seen on a regular basis, the frequency of follow-up dependent
622
on the risk for developing melanoma. Various scoring systems have been devised to predict the likelihood of this occurrence [69,85 87]. Suffice it to say, the stronger the family history of melanoma and dysplastic nevi, the higher the risk of melanoma. Screening should be offered not only to patients with dysplastic nevi, but also all first-degree relatives and selected second-degree relatives. Wachsmuth et al [30] have suggested based on their study of the prevalence of mutations of CDKN2A in melanoma families that because of marked overlap in the phenotypic expression of the dysplastic nevus syndrome between gene carriers and noncarriers, all family members should be viewed as potential carriers. Similarly, readily identifiable phenotypic traits such as a significant number of banal and atypical nevi, sun sensitivity, or hair and eye color, are useful in stratifying patients according to their probability for developing melanoma [67,88,89]. Patients from melanoma-prone families should have their skin examined initially every 3 to 6 months, until both the patient and physician are comfortable that the patients nevi are stable in appearance. Assessing for the latter has been shown to be greatly facilitated by the use of total body photographs [66,90,91]. Not only is this reassuring to the patient and physician, it ultimately results in less nevi being excised for histologic assessment and leads to the early detection of thin, potentially curable melanomas [66]. If the moles seem to be static, the interval between visits can be increased to every 6 to 12 months. A change in appearance does not necessarily imply malignant transformation. Halpern et al [92] have shown that dysplastic nevi are dynamic throughout life, and that 42% either exhibit decreased atypia or disappear altogether. During follow-up visits the physician should educate the patient of the warning signs of melanoma, and reiterate the need for routine self-examination (eg, once every 2 to 3 months) and the means to safeguard against this and other skin cancers. Because melanoma is rare in childhood [93], even in melanomaprone families [52], screening can be started in the late teens or the early twenties, and should continue indefinitely (largely dependent on the number of risk factors for melanoma). Aids to the detection of melanoma The entire skin surface (including intertriginous areas and scalp) should be examined carefully and any suspicious lesions should be scrutinized by dermoscopy when possible. Dermoscopy has been shown to increase the diagnostic accuracy for melanoma, particularly if the clinician is experienced with
its use [94 98]. Although there are no diagnostic or pathognomonic dermoscopic features to distinguish dysplastic nevi from melanoma reliably and with absolute certainty, there are features that should heighten ones level of suspicion for melanoma, which in turn should prompt the clinician to perform a biopsy to exclude the latter [99]. Are we doing a disservice to patients by not performing dermoscopy on suspicious pigmented lesions as has been suggested [100]? The answer to this question is not as straightforward as it may seem. Although this author and most melanoma specialists rely on dermoscopy to differentiate benign from malignant pigmented lesions, only approximately 15% of dermatologists in the United States routinely use this device, whereas throughout Europe it is almost the standard of care when assessing pigmented lesions [100]. Reasons given for not using dermoscopy include with typical melanomas, dermoscopy adds little to the clinical diagnosis [101]; the diagnostic accuracy for melanoma with dermoscopy is not 100% (it is in the order of 80% to 90%) [99]; and approximately 10% of melanomas are featureless in terms of dermoscopic structures, and dermoscopy does not help in their identification [102]. Furthermore, special training is generally required; inexperienced users have a tendency to overdiagnose melanoma, leading to more rather than less biopsies being performed [95,103]. It has been argued that if a lesion is atypical enough to warrant evaluation by dermoscopy, it is probably equally suspicious enough to warrant histologic assessment. Lastly, medicolegal concerns and financial remuneration issues further complicate this topic. Several studies have shown that melanoma surveillance can be enhanced through the use of conventional total body photography, which permits the detection of change in a patients nevi [90]. Not only does photography lead to the identification of early melanomas, it also results in a substantial reduction in the number of lesions that are excised [66]. Unfortunately, medical payers (in the United States and Canada) are reluctant to reimburse this tool. As a consequence baseline cutaneous photography is generally unavailable outside academic institutions or specialized pigmented lesion clinics. Furthermore, conventional photography is extremely time consuming and requires considerable effort for the acquisition, storage, and retrieval of slides (or prints) required to make comparison assessments on follow-up visits. Computerized imaging systems (eg, MoleMax) can minimize some of these problems, which may in turn improve the surveillance of patients with multiple dysplastic nevi [104]. Some digital imaging systems
623
are now capable of analyzing lesions dermoscopically and provide an automated diagnosis in terms of the probability of whether or not a lesion is malignant. In the foreseeable future, it is likely that the diagnosis rendered by the machine will equal if not exceed the best clinicians in diagnosing melanoma [105,106]. Genetic counseling It is estimated that approximately 5% to 10% of melanoma patients have one or more family members with melanoma. Although many of these family clusters are caused by chance alone, and arise secondary to shared sun exposure, detailed genetic studies suggest that perhaps one in five such families have a distinct family predisposition for developing melanoma because of the inheritance of a melanoma susceptibility gene. Most evidence suggests that the propensity for developing melanoma is transmitted in an autosomal-dominant manner [107], with affected individuals having an 80% to 100% lifetime risk for developing melanoma [52]. It logically follows that it is prudent to be able to identify such individuals so that steps can be taken either to prevent melanoma, or at least ensure its early detection when the cure rate is highest. Genetic testing has been proposed as one means to achieve this goal. Despite the apparent advantages to genetic testing there are several disadvantages, the most obvious being that a negative test result does not guarantee against developing melanoma, and patients may be reassured falsely that they are not at risk. Genetic testing for cancer risk also carries psychologic implications and raises ethical and legal issues. The Melanoma Genetics Consortium is of the opinion that there is no role for genetic counseling and testing for CDKN2A and CDK4 mutations outside research protocols [108]. Sun protection Because the appearance of moles is dependent on genetic and environmental factors [109,110], patients should be advised of the need to minimize solar exposure through sun avoidance during peak hours of sunlight intensity (ie, between 10 AM and 4 PM); the use of sun-protective clothing; and broad-spectrum sunscreens of SPF 30 or higher. Gallagher et al [111] have recently demonstrated that regular sunscreen use in children can act as a chemopreventative agent, resulting in the development of fewer nevi. The beneficial effects seem to be most prominent in individuals who have numerous freckles (ie, those who are susceptible to sun-induced skin injury). These findings, however,
do not prove that sunscreens are beneficial at preventing melanoma. There have been 11 case-controlled studies examining the relationship between melanoma and sunscreen use [112]. Six studies found paradoxically higher sunscreen use in the melanoma group, whereas three studies found no difference, and two studies reported lower sunscreen use in melanoma patients than in controls [112]. Other approaches of chemoprevention for skin cancer are only at the experimental stages, and await clinical trials [113]. Removal of dysplastic nevi Traditionally, dysplastic nevi were viewed as precursor lesions to melanoma, and it was common to recommend that atypical-appearing moles be removed prophylactically. Because dysplastic nevi rarely eventuate into melanoma, the notion of them being precursor lesions has diminished greatly; instead, they are viewed primarily as markers of increased risk for developing melanoma. Because the identification of dysplastic nevi can be made readily and reliably on clinical grounds [66,67,70], there is no justification for removing moles simply to confirm the diagnosis. Histologic evaluation of a dysplastic nevus does not help identify dysplastic nevus syndrome patients [78], nor does it stratify patients into melanoma risk categories [76]. The main objective of histology is not to confirm a diagnosis of dysplastic nevus syndrome but rather to rule out melanoma [78]. There are several well-conducted studies that demonstrate the futility of removing dysplastic nevi preventively [66,114]. Kelly et al [66] in a prospective cohort study of 278 adults with five or more atypical nevi detected 20 new melanomas in 16 patients over the study period, 11 of which were detected because of change in pigmented lesions as compared with baseline photographs, and the other 9 of which were detected by patients or their partners. Thirteen of the 20 melanomas arose from new lesions, and three from previously existing dysplastic nevi. The authors surmise that if all dysplastic nevi in their patients were removed (almost 6000 moles), only three incident melanomas would have been prevented. They conclude that prophylactic excision of dysplastic nevi is an unsatisfactory alternative to follow-up, because it does not provide sufficient risk-reduction to justify the cost and morbidity associated with this procedure. Cohen et al [114] detailed their experience of wholesale excision of nevi in high-risk patients to assess the frequency of undetected melanoma. In their study, 78 stage I and II melanoma patients underwent prophy-
624
lactic excision of 1630 melanocytic nevi, which resulted in the detection of 12 unsuspected melanomas. Restated, for every melanoma detected they needed to remove approximately 136 nevi. Despite these observations, an argument can be made for prophylactically removing nevi if baseline total body photography is not available [90]; the number of dysplastic nevi are few; or are located in areas that are difficult to visualize (eg, perineum). In the long run it is more cost effective simply to remove a couple of atypical nevi than ask the patient to return for frequent follow-up visits. Ideally, suspicious pigmented lesions should be removed by excisional biopsy because this provides the pathologist with optimal tissue for histologic diagnosis [115]. The usefulness of deep-shave excision has been recently revisited [116,117]. It has been shown for macular, melanocytic nevi that the razor blade biopsy technique provides adequate tissue for histologic assessment 88% of the time [117]. In only 12% of the cases are nevi incompletely excised on depth. The authors conclude that deep razor blade excision is a highly useful and inexpensive method for removal of macular, melanocytic nevi that results in adequate specimens for histologic assessment, and is particularly useful in patients with multiple nevi, such as those with dysplastic nevus syndrome [117]. Similarly, Breuninger et al [116] have shown that shave excision resulted in fewer complications, and the scars were generally considered cosmetically more desirable than those obtained with excision and primary closure, however, the incidence of recurrences was three times higher with shave removal. Because of difficulties in pathologically distinguishing early melanoma in situ from severely dysplastic nevi, pathologists occasionally recommend re-excision if melanocytes extend to or are in close proximity to the surgical margin of the specimen. The need for re-excision in this setting is controversial. NIH Consensus Statement suggested re-excision with margins of 0.2 to 0.5 mm when there is involvement of the margin [7]. Conversely, it can be argued that because most melanomas arise de novo, and the fact that most dysplastic nevi do not progress to melanoma, re-excision of most lesions is probably unnecessary unless there is substantial cytologic or architectural atypia [118]. Cohen et al [119] evaluated the usefulness of re-excision in 189 dysplastic nevi. In their study, approximately 25% of re-excision specimens had residual nevus cells. Interestingly, the proportion of specimens with residual nevi was more common in those removed by punch biopsy rather than by shave or elliptical excision. Only 1 of the 189 lesions on re-excision demonstrated melanoma.
Nevi reduction for cosmetic reasons (and possibly prophylactic reasons) Current management of dysplastic nevi consists of surgical removal with adequate margins to ensure the submission of an optimal specimen for histologic evaluation [120]. Unfortunately, this may result in multiple unsightly scars. As a consequence various attempts have been tried to ablate these lesion with the least amount of disfigurement. Treatments tried have included topical 5% fluorouracil [121], topical tretinoin with or without hydrocortisone [122 125], oral isotretinoin [126], and laser ablation [127]. None of these modalities result in the complete ablation of the nevus clinically or histologically. Although lasers seem the most promising of these therapies, its role in the treatment of melanocytic nevi (including dysplastic nevi) remains controversial [120]. Most lasers seem only to ablate melanocytes within the epidermis and papillary dermis [120]. In addition, there is the theoretical concern that lasers may result in the malignant transformation of partially regressed nevi.
References
[1] Clark Jr. WH, Reimer RR, Greene M, et al. Origin of familial malignant melanomas from heritable melanocytic lesions: the B-K mole syndrome. Arch Dermatol 1978;114:732 8. [2] Elder DE, Goldman LI, Goldman SC, et al. Dysplastic nevus syndrome: a phenotypic association of sporadic cutaneous melanoma. Cancer 1980;46: 1787 94. [3] Dowd P, Everall J. Atypical multiple mole melanoma syndrome. Br J Dermatol 1979;101(suppl 17):32 3. [4] Lynch HT, Frichot 3rd BC, Lynch JF. Familial atypical multiple mole-melanoma syndrome. J Med Genet 1978;15:352 6. [5] Ackerman AB, Milde P. Naming acquired melanocytic nevi: common and dysplastic, normal and atypical, or Unna, Miescher, Spitz, and Clark? Am J Dermatopathol 1992;14:447 53. [6] Urso C, Giannini A, Bartolini M, et al. Histological analysis of intraepidermal proliferations of atypical melanocytes. Am J Dermatopathol 1990;12:150 5. [7] NIH Consensus Development Conference. Diagnosis and treatment of early melanoma. Presented at the NIH Consensus Development Conference. January 27 29, 1992;10(1):1 25. [8] Sina B, Wood C. Atypical melanocytic proliferations. Am J Dermatopathol 1991;13:317 9. [9] Frankel KA. Intraepithelial melanocytic neoplasia: a classification by pattern analysis of proliferations of atypical melanocytes. Am J Dermatopathol 1987; 9:80 1.
T.G. Salopek / Dermatol Clin 20 (2002) 617628 [10] Slade J, Marghoob AA, Salopek TG, et al. Atypical mole syndrome: risk factor for cutaneous malignant melanoma and implications for management. J Am Acad Dermatol 1995;32:479 94. [11] Greene MH. Genetics of cutaneous melanoma and nevi. Mayo Clin Proc 1997;72:467 74. [12] Hayward N. New developments in melanoma genetics. Curr Oncol Rep 2000;2:300 6. [13] Gruis NA, Sandkuijl LA, Weber JL, et al. Linkage analysis in Dutch familial atypical multiple mole-melanoma (FAMMM) syndrome families: effect of naevus count. Melanoma Res 1993;3:271 7. [14] Holland EA, Schmid H, Kefford RF, et al. CDKN2A (P16(INK4a)) and CDK4 mutation analysis in 131 Australian melanoma probands: effect of family history and multiple primary melanomas. Genes Chromosomes Cancer 1999;25:339 48. [15] MacKie RM, Andrew N, Lanyon WG, et al. CDKN2A germline mutations in U.K. patients with familial melanoma and multiple primary melanomas. J Invest Dermatol 1998;111:269 72. [16] Newton Bishop JA, Bataille V, Pinney E, et al. Family studies in melanoma: identification of the atypical mole syndrome (AMS) phenotype. Melanoma Res 1994;4:199 206. [17] Soufir N, Avril MF, Chompret A, et al. Prevalence of p16 and CDK4 germline mutations in 48 melanomaprone families in France. The French Familial Melanoma Study Group. Hum Mol Genet 1998;7:209 16. [18] Serrano M, Hannon GJ, Beach D. A new regulatory motif in cell-cycle control causing specific inhibition of cyclin D/CDK4. Nature 1993;366:704 7. [19] Bale SJ, Chakravarti A, Greene MH. Cutaneous malignant melanoma and familial dysplastic nevi: evidence for autosomal dominance and pleiotropy. Am J Hum Genet 1986;38:188 96. [20] Easton DF, Cox GM, Macdonald AM, et al. Genetic susceptibility to naevi: a twin study. Br J Cancer 1991; 64:1164 7. [21] Crijns MB, Vink J, Van Hees CL, et al. Dysplastic nevi: occurrence in first- and second-degree relatives of patients with sporadic dysplastic nevus syndrome. Arch Dermatol 1991;127:1346 51. [22] Newton JA, Bataille V, Griffiths K, et al. How common is the atypical mole syndrome phenotype in apparently sporadic melanoma? J Am Acad Dermatol 1993;29:989 96. [23] Hussussian CJ, Struewing JP, Goldstein AM, et al. Germline p16 mutations in familial melanoma. Nat Genet 1994;8:15 21. [24] Gruis NA, Van der Velden PA, Bergman W, et al. Genetics of familial atypical multiple mole-melanoma (FAMMM) syndrome in The Netherlands: how far have we come? Bull Cancer 1998;85:627 30. [25] Puig S, Ruiz A, Castel T, et al. Inherited susceptibility to several cancers but absence of linkage between dysplastic nevus syndrome and CDKN2A in a melanoma family with a mutation in the CDKN2A (P16INK4A) gene. Hum Genet 1997;101:359 64.
625
[26] Ruiz A, Puig S, Malvehy J, et al. CDKN2A mutations in Spanish cutaneous malignant melanoma families and patients with multiple melanomas and other neoplasia. J Med Genet 1999;36:490 3. [27] Park WS, Vortmeyer AO, Pack S, et al. Allelic deletion at chromosome 9p21(p16) and 17p13(p53) in microdissected sporadic dysplastic nevus. Hum Pathol 1998;29:127 30. [28] Goldstein AM, Martinez M, Tucker MA, et al. Gene-covariate interaction between dysplastic nevi and the CDKN2A gene in American melanomaprone families. Cancer Epidemiol Biomarkers Prev 2000;9:889 94. [29] Bishop JA, Wachsmuth RC, Harland M, et al. Genotype/phenotype and penetrance studies in melanoma families with germline CDKN2A mutations. J Invest Dermatol 2000;114:28 33. [30] Wachsmuth RC, Harland M, Bishop JA. The atypicalmole syndrome and predisposition to melanoma. N Engl J Med 1998;339:348 9. [31] Hashemi J, Linder S, Platz A, et al. Melanoma development in relation to non-functional p16/INK4A protein and dysplastic naevus syndrome in Swedish melanoma kindreds. Melanoma Res 1999;9:21 30. [32] Clark Jr. WH, Ackerman AB. An exchange of views regarding the dysplastic nevus controversy. Semin Dermatol 1989;8:229 50. [33] Ackerman AB, Mihara I. Dysplasia, dysplastic melanocytes, dysplastic nevi, the dysplastic nevus syndrome, and the relation between dysplastic nevi and malignant melanomas. Hum Pathol 1985;16:87 91. [34] Ahmed AA, Nordlind K, Hedblad M, et al. Interleukin (IL)-1 alpha- and -1 beta-, IL-6-, and tumor necrosis factor-alpha-like immunoreactivities in human common and dysplastic nevocellular nevi and malignant melanoma. Am J Dermatopathol 1995;17: 222 9. [35] Bergman W, Ruiter DJ, Scheffer E, et al. Melanocytic atypia in dysplastic nevi: immunohistochemical and cytophotometrical analysis. Cancer 1988;61:1660 6. [36] Caporaso N, Greene MH, Tsai S, et al. Cytogenetics in hereditary malignant melanoma and dysplastic nevus syndrome: is dysplastic nevus syndrome a chromosome instability disorder? Cancer Genet Cytogenet 1987;24:299 314. [37] Elder DE, Herlyn M. Antigens associated with tumor progression in melanocytic neoplasia. Pigment Cell Res Suppl 1992;2:136 43. [38] Ewanowich C, Brynes RK, Medeiros L, et al. Cyclin D1 expression in dysplastic nevi: an immunohistochemical study. Arch Pathol Lab Med 2001;125: 208 10. [39] Fleming MG, Howe SF, Candel AG. Immunohistochemical localization of cytokines in nevi. Am J Dermatopathol 1992;14:496 503. [40] Jaspers NG, Roza-de Jongh EJ, Donselaar IG, et al. Sister chromatid exchanges, hyperdiploidy and chromosomal rearrangements studied in cells from melanoma-prone individuals belonging to families with
626
T.G. Salopek / Dermatol Clin 20 (2002) 617628 the dysplastic nevus syndrome. Cancer Genet Cytogenet 1987;24:33 43. Kim JC, Murphy GF. Dysplastic melanocytic nevi and prognostically indeterminate nevomelanomatoid proliferations. Clin Lab Med 2000;20:691 712. Lazzaro B, Strassburg A. Tumor antigen expression in compound dysplastic nevi and superficial spreading melanoma defined by a panel of nevomelanoma monoclonal antibodies. Hybridoma 1996;15: 141 6. Meier F, Satyamoorthy K, Nesbit M, et al. Molecular events in melanoma development and progression. Front Biosci 1998;3:D1005 10. Moretti S, Martini L, Berti E, et al. Adhesion molecule profile and malignancy of melanocytic lesions. Melanoma Res 1993;3:235 9. Nanney LB, Coffey Jr. RJ, Ellis DL. Expression and distribution of transforming growth factor-alpha within melanocytic lesions. J Invest Dermatol 1994;103: 707 14. Platz A, Ringborg U, Grafstrom E, et al. Immunohistochemical analysis of the N-ras p21 and the p53 proteins in naevi, primary tumours and metastases of human cutaneous malignant melanoma: increased immunopositivity in hereditary melanoma. Melanoma Res 1995;5:101 6. Wang Y, Rao U, Mascari R, et al. Molecular analysis of melanoma precursor lesions. Cell Growth Differ 1996;7:1733 40. Lynch HT, Fusaro RM, Sandberg AA, et al. Chromosome instability and the FAMMM syndrome. Cancer Genet Cytogenet 1993;71:27 39. Bergman W, Watson P, de Jong J, et al. Systemic cancer and the FAMMM syndrome. Br J Cancer 1990;61:932 6. Lynch HT, Fusaro RM, Pester J, et al. Tumour spectrum in the FAMMM syndrome. Br J Cancer 1981; 44:553 60. Greene MH, Tucker MA, Clark Jr. WH, et al. Hereditary melanoma and the dysplastic nevus syndrome: the risk of cancers other than melanoma. J Am Acad Dermatol 1987;16:792 7. Tucker MA, Fraser MC, Goldstein AM, et al. Risk of melanoma and other cancers in melanoma-prone families. J Invest Dermatol 1993;100:350S 5S. Gilchrest BA, Eller MS, Geller AC, et al. The pathogenesis of melanoma induced by ultraviolet radiation. N Engl J Med 1999;340:1341 8. Ramsay RG, Chen P, Imray FP, et al. Familial melanoma associated with dominant ultraviolet radiation sensitivity. Cancer Res 1982;42:2909 12. Smith PJ, Greene MH, Devlin DA, et al. Abnormal sensitivity to UV-radiation in cultured skin fibroblasts from patients with hereditary cutaneous malignant melanoma and dysplastic nevus syndrome. Int J Cancer 1982;30:39 45. Howell JN, Greene MH, Corner RC, et al. Fibroblasts from patients with hereditary cutaneous malignant melanoma are abnormally sensitive to the mutagenic effect of simulated sunlight and 4-nitroquinoline 1-oxide. Proc Natl Acad Sci U S A 1984;81:1179 83. Sanford KK, Tarone RE, Parshad R, et al. Hypersensitivity to G2 chromatid radiation damage in familial dysplastic naevus syndrome. Lancet 1987; 2:1111 6. Abrahams PJ, Houweling A, Cornelissen-Steijger PD, et al. Impaired DNA repair capacity in skin fibroblasts from various hereditary cancer-prone syndromes. Mutat Res 1998;407:189 201. Moriwaki SI, Tarone RE, Tucker MA, et al. Hypermutability of UV-treated plasmids in dysplastic nevus/familial melanoma cell lines. Cancer Res 1997; 57:4637 41. Noz KC, Bauwens M, van Buul PP, et al. Comet assay demonstrates a higher ultraviolet B sensitivity to DNA damage in dysplastic nevus cells than in common melanocytic nevus cells and foreskin melanocytes. J Invest Dermatol 1996;106:1198 202. Runger TM, Epe B, Moller K, et al. Repair of directly and indirectly UV-induced DNA lesions and of DNA double-strand breaks in cells from skin cancer-prone patients with the disorders dysplastic nevus syndrome or basal cell nevus syndrome. Recent Results Cancer Res 1997;143:337 51. Thielmann HW, Popanda O, Edler L, et al. DNA repair synthesis following irradiation with 254-nm and 312-nm ultraviolet light is not diminished in fibroblasts from patients with dysplastic nevus syndrome. J Cancer Res Clin Oncol 1995;121:327 37. Hussein MR, Sun M, Tuthill RJ, et al. Comprehensive analysis of 112 melanocytic skin lesions demonstrates microsatellite instability in melanomas and dysplastic nevi, but not in benign nevi. J Cutan Pathol 2001;28: 343 50. Hussein MR, Roggero E, Sudilovsky EC, et al. Alterations of mismatch repair protein expression in benign melanocytic nevi, melanocytic dysplastic nevi, and cutaneous malignant melanomas. Am J Dermatopathol 2001;23:308 14. Elder DE, Green MH, Guerry 4th D, et al. The dysplastic nevus syndrome: our definition. Am J Dermatopathol 1982;4:455 60. Kelly JW, Yeatman JM, Regalia C, et al. A high incidence of melanoma found in patients with multiple dysplastic naevi by photographic surveillance. Med J Aust 1997;167:191 4. Tucker MA, Halpern A, Holly EA, et al. Clinically recognized dysplastic nevi. A central risk factor for cutaneous melanoma. JAMA 1997;277:1439 44. Bergman W, van Voorst Vader PC, Ruiter DJ. Dysplastic nevi and the risk of melanoma: a guideline for patient care. (Nederlandse Melanoom Werkgroep van de Vereniging voor Integrale Kankercentra). Ned Tijdschr Geneeskd 1997;141:2010 4. Newton JA. Familial melanoma. Clin Exp Dermatol 1993;18:5 11. Bishop JA, Bradburn M, Bergman W, et al. Teaching non-specialist health care professionals how to iden-
[41]
[57]
[42]
[58]
[43]
[59]
[44]
[60]
[45]
[61]
[46]
[62]
[47]
[48]
[63]
[49]
[50]
[64]
[51]
[65]
[52]
[66]
[53]
[67]
[54]
[68]
[55]
[69] [70]
[56]
T.G. Salopek / Dermatol Clin 20 (2002) 617628 tify the atypical mole syndrome phenotype: a multinational study. Br J Dermatol 2000;142:331 7. Roth ME, Grant-Kels JM, Ackerman AB, et al. The histopathology of dysplastic nevi: continued controversy. Am J Dermatopathol 1991;13:38 51. Ackerman AB. Dysplastic nevus. Am J Surg Pathol 2000;24:757 8. de Wit PE, vant Hof-Grootenboer B, Ruiter DJ, et al. Validity of the histopathological criteria used for diagnosing dysplastic naevi: an interobserver study by the pathology subgroup of the EORTC Malignant Melanoma Cooperative Group. Eur J Cancer 1993; 29A:831 9. Weinstock MA, Barnhill RL, Rhodes AR, et al. Reliability of the histopathologic diagnosis of melanocytic dysplasia. The Dysplastic Nevus Panel. Arch Dermatol 1997;133:953 8. Shea CR, Vollmer RT, Prieto VG. Correlating architectural disorder and cytologic atypia in Clark (dysplastic) melanocytic nevi. Hum Pathol 1999;30:500 5. Pozo L, Naase M, Cerio R, et al. Critical analysis of histologic criteria for grading atypical (dysplastic) melanocytic nevi. Am J Clin Pathol 2001;115:194 204. Piepkorn M, Meyer LJ, Goldgar D, et al. The dysplastic melanocytic nevus: a prevalent lesion that correlates poorly with clinical phenotype. J Am Acad Dermatol 1989;20:407 15. Mooi WJ. The dysplastic naevus. J Clin Pathol 1997; 50:711 5. Barnhill RL, Roush GC. Correlation of clinical and histopathologic features in clinically atypical melanocytic nevi. Cancer 1991;67:3157 64. Black WC, Hunt WC. Histologic correlations with the clinical diagnosis of dysplastic nevus. Am J Surg Pathol 1990;14:44 52. Kelly JW, Crutcher WA, Sagebiel RW. Clinical diagnosis of dysplastic melanocytic nevi: a clinicopathologic correlation. J Am Acad Dermatol 1986;14: 1044 52. Annessi G, Cattaruzza MS, Abeni D, et al. Correlation between clinical atypia and histologic dysplasia in acquired melanocytic nevi. J Am Acad Dermatol 2001;45:77 85. Klein LJ, Barr RJ. Histologic atypia in clinically benign nevi: a prospective study. J Am Acad Dermatol 1990;22:275 82. Toussaint S, Kamino H. Dysplastic changes in different types of melanocytic nevi: a unifying concept. J Cutan Pathol 1999;26:84 90. Kraemer KH, Greene MH. Dysplastic nevus syndrome: familial and sporadic precursors of cutaneous melanoma. Dermatol Clin 1985;3:225 37. MacKie RM, McHenry P, Hole D. Accelerated detection with prospective surveillance for cutaneous malignant melanoma in high-risk groups. Lancet 1993; 341:1618 20. Rigel DS, Rivers JK, Friedman RJ, et al. Risk gradient for malignant melanoma in individuals with dysplastic naevi. Lancet 1988;1:352 3.
627
[71]
[72] [73]
[74]
[75]
[76]
[77]
[78] [79]
[80]
[81]
[82]
[83]
[84]
[85]
[86]
[87]
[88] Koh HK. Cutaneous melanoma. N Engl J Med 1991; 325:171 82. [89] Marghoob AA, Slade J, Salopek TG, et al. Basal cell and squamous cell carcinomas are important risk factors for cutaneous malignant melanoma: screening implications. Cancer 1995;75(suppl 2): 707 14. [90] Rhodes AR. Intervention strategy to prevent lethal cutaneous melanoma: use of dermatologic photography to aid surveillance of high-risk persons. J Am Acad Dermatol 1998;39:262 7. [91] Slue W, Kopf AW, Rivers JK. Total-body photographs of dysplastic nevi. Arch Dermatol 1988;124: 1239 43. [92] Halpern AC, Guerry 4th D, Elder DE, et al. Natural history of dysplastic nevi. J Am Acad Dermatol 1993; 29:51 7. [93] Scalzo DA, Hida CA, Toth G, et al. Childhood melanoma: a clinicopathological study of 22 cases. Melanoma Res 1997;7:63 8. [94] Binder M, Puespoeck-Schwarz M, Steiner A, et al. Epiluminescence microscopy of small pigmented skin lesions: short-term formal training improves the diagnostic performance of dermatologists. J Am Acad Dermatol 1997;36:197 202. [95] Binder M, Schwarz M, Winkler A, et al. Epiluminescence microscopy: a useful tool for the diagnosis of pigmented skin lesions for formally trained dermatologists. Arch Dermatol 1995;131:286 91. [96] Nachbar F, Stolz W, Merkle T, et al. The ABCD rule of dermatoscopy: high prospective value in the diagnosis of doubtful melanocytic skin lesions. J Am Acad Dermatol 1994;30:551 9. [97] Nilles M, Boedeker RH, Schill WB. Surface microscopy of naevi and melanoma: clues to melanoma. Br J Dermatol 1994;130:349 55. [98] Steiner A, Binder M, Schemper M, et al. Statistical evaluation of epiluminescence microscopy criteria for melanocytic pigmented skin lesions. J Am Acad Dermatol 1993;29:581 8. [99] Salopek TG, Kopf AW, Stefanato CM, et al. Differentiation of atypical moles (dysplastic nevi) from early melanomas by dermoscopy. Dermatol Clin 2001;19:337 45. [100] Johr R, Izakovic J. Should you be using epiluminescence microscopy? Skin & Aging 2000;8:28 38. [101] Mayer J. Systematic review of the diagnostic accuracy of dermatoscopy in detecting malignant melanoma. Med J Aust 1997;167:206 10. [102] Menzies SW, Ingvar C, McCarthy WH. A sensitivity and specificity analysis of the surface microscopy features of invasive melanoma. Melanoma Res 1996; 6:55 62. [103] Rao BK, Marghoob AA, Stolz W, et al. Can early malignant melanoma be differentiated from atypical melanocytic nevi by in vivo techniques? Part I. Clinical and dermoscopic characteristics. Skin Research and Technology 1997;3:8 14. [104] Kittler H, Pehamberger H, Wolff K, et al. Follow-up
628
T.G. Salopek / Dermatol Clin 20 (2002) 617628 of melanocytic skin lesions with digital epiluminescence microscopy: patterns of modifications observed in early melanoma, atypical nevi, and common nevi. J Am Acad Dermatol 2000;43:467 76. Andreassi L, Perotti R, Rubegni P, et al. Digital dermoscopy analysis for the differentiation of atypical nevi and early melanoma: a new quantitative semiology. Arch Dermatol 1999;135:1459 65. Menzies SW. Automated epiluminescence microscopy: human vs machine in the diagnosis of melanoma. Arch Dermatol 1999;135:1538 40. Greene MH. The genetics of hereditary melanoma and nevi. 1998 update. Cancer 1999;86(suppl 11): 2464 77. Kefford RF, Newton Bishop JA, Bergman W, et al. Counseling and DNA testing for individuals perceived to be genetically predisposed to melanoma: a consensus statement of the Melanoma Genetics Consortium. J Clin Oncol 1999;17:3245 51. Bataille V, Grulich A, Sasieni P, et al. The association between naevi and melanoma in populations with different levels of sun exposure: a joint case-control study of melanoma in the UK and Australia. Br J Cancer 1998;77:505 10. Kelly JW, Rivers JK, MacLennan R, Harrison S, Lewis AE, Tate BJ. Sunlight: a major factor associated with the development of melanocytic nevi in Australian schoolchildren. J Am Acad Dermatol 1994; 30:40 8. Gallagher RP, Rivers JK, Lee TK, et al. Broad-spectrum sunscreen use and the development of new nevi in white children: a randomized controlled trial. JAMA 2000;283:2955 60. Bigby M. The sunscreen and melanoma controversy. Arch Dermatol 1999;135:1526 7. Bickers DR, Athar M. Novel approaches to chemoprevention of skin cancer. J Dermatol 2000;27: 691 5. Cohen MH, Cohen BJ, Shotkin JD, et al. Surgical prophylaxis of malignant melanoma. Ann Surg 1991; 213:308 14. Salopek TG, Slade J, Marghoob AA, et al. Management of cutaneous malignant melanoma by dermatologists of the American Academy of Dermatology. I. Survey of biopsy practices of pigmented lesions suspected as melanoma. J Am Acad Dermatol 1995;33: 441 50. Breuninger H, Garbe C, Rassner G. Shave excision of melanocytic nevi of the skin: indications, technique, results. Hautarzt 2000;51:575 80. Gambichler T, Senger E, Rapp S, et al. Deep shave excision of macular melanocytic nevi with the razor blade biopsy technique. Dermatol Surg 2000;26: 662 6. Metcalf JS, Maize JC. Clarks nevus. Semin Cutan Med Surg 1999;18:43 6. Cohen LM, Hodge SJ, Owen LG, et al. Atypical melanocytic nevi: clinical and histopathologic predictors of residual tumor at reexcision. J Am Acad Dermatol 1992;27:701 6. Stratigos AJ, Dover JS, Arndt KA. Laser treatment of pigmented lesions 2000: how far have we gone? Arch Dermatol 2000;136:915 21. Bondi EE, Clark Jr. WH, Elder D, et al. Topical chemotherapy of dysplastic melanocytic nevi with 5% fluorouracil. Arch Dermatol 1981;117:89 92. Edwards L, Jaffe P. The effect of topical tretinoin on dysplastic nevi: a preliminary trial. Arch Dermatol 1990;126:494 9. Halpern AC, Schuchter LM, Elder DE, et al. Effects of topical tretinoin on dysplastic nevi. J Clin Oncol 1994;12:1028 35. Meyskens Jr. FL, Edwards L, Levine NS. Role of topical tretinoin in melanoma and dysplastic nevi. J Am Acad Dermatol 1986;15:822 5. Stam-Posthuma JJ, Vink J, le Cessie S, et al. Effect of topical tretinoin under occlusion on atypical naevi. Melanoma Res 1998;8:539 48. Edwards L, Meyskens F, Levine N. Effect of oral isotretinoin on dysplastic nevi. J Am Acad Dermatol 1989;20:257 60. Duke D, Byers HR, Sober AJ, et al. Treatment of benign and atypical nevi with the normal-mode ruby laser and the Q-switched ruby laser: clinical improvement but failure to completely eliminate nevomelanocytes. Arch Dermatol 1999;135:290 6.
[105]
[116]
[117]
[106]
[107]
[118] [119]
[108]
[120]
[109]
[121]
[122]
[110]
[123]
[111]
[124]
[125]
[112] [113]
[126]
[114]
[127]
[115]
Dermatol Clin 20 (2002) 629 640
Screening for melanoma

Alan C. Geller, MPH, RN*
Department of Dermatology, School of Medicine, Boston University, 720 Harrison Ave DOB 801A Boston, MA 02118, USA Department of Epidemiology and Biostatistics, School of Public Health, Boston University, 115 Albany St, T Building, Boston, MA 02118, USA
New developments in melanoma screening pose exciting challenges for cancer control planners. Most recently (2000 and 2001), two influential bodies in the United States neither recommended populationbased screening nor provided Medicare reimbursement for screening, but proposed new initiatives for screening elderly, at-risk men. Concurrently, a population-based, community screening effort just underway in Queensland, Australia, promises to answer many key questions in the decade to come. This article reviews major studies of the past two decades, highlights the need to screen previously unscreened populations, and concludes with suggestions for reducing the rising rates of melanoma mortality.
of screening for melanoma, however, recent evidencebased guidelines have not endorsed routine screening. Applications of screening definitions to melanoma Distinguishing screening from early detection, education, and case finding can be challenging in this visible tumor. Precise definitions of the word screening can apply to any process of visually examining the skin for early detection: a narrower use distinguishes screening from surveillance; case-finding (when searching for cancer within a routine physical examination by a physician); early detection; and opportunistic surveillance. Furthermore, screening can take place in a variety of settings: within the physician examination at his or her office, at health fairs, at workplaces, or in mass efforts where persons select themselves to attend [1 4]. Trying to apply standard textbook definitions of cancer screening to melanoma underscores the difficulties of making rigid distinctions between screening and education. As an external tumor, melanoma should be discovered more readily than other types of cancer: melanoma writes its message in the skin with its own ink and is there for all to see [5]. Education can alert the public about ways to recognize melanoma, especially stressing the ABCD rule of melanoma [6]. Publicity should also promote awareness of risk factors, prompt medical attention for suspect lesions, and improve skin self-examination (SSE) rates for high-risk persons. Because patients may be unaware of melanoma on the back and other areas that are difficult for self-inspection, visual examinations by physicians, nurses, family members, and others can aid early detection. Professional education can be facilitated by melanomas unique visual nature and
Rationale for melanoma screening Screening and early detection programs could save many lives otherwise lost to melanoma. This cancer is external and visible, risk factors are well documented, and screening tests are safe and acceptable to the public. Furthermore, early melanoma can be cured by simple surgical excision. Melanoma screening runs the spectrum from an average-risk person conducting casual self-examination to high-risk persons undergoing careful, regular, and systematic monthly examination of the entire skin by a skin cancer specialist [1 3]. Without randomized trials testing the efficacy
* 720 Harrison Avenue, Doctors Office Building 801A, Boston, MA 02118. E-mail address: ageller@bu.edu (A.C. Geller).
630
A.C. Geller / Dermatol Clin 20 (2002) 629640
teaching through pictorial displays, atlases, and webbased education. All these strategies must be targeted to the level of knowledge and practices of health professionals and public within a given community. Theoretical aspects of melanoma screening Cancer screening helps most when (1) the disease is highly prevalent and causes considerable morbidity and mortality; (2) the natural history of the disease is known; (3) early treatment can prevent morbidity and mortality; and (4) an acceptable, safe, and a relatively inexpensive screening test exists [4]. Using these criteria, many cancer control experts agree that cutaneous melanoma screening has theoretic appeal but also a number of barriers: Facilitators Natural history of the disease is well known Rising incidence and mortality rates Easily identifiable risk factors Many at-risk individuals in health care system Physicians can examine hard-to-see areas Barriers Lack of randomized clinical trial Prohibitive costs for screening trial Lack of fully trained medical force Difficulties in maintaining control population In the United States, the disease is increasingly prevalent, with rising incidence and mortality rates [7]. Surgical treatment of early thin melanoma can lead to cure, whereas metastatic melanoma remains generally incurable. The screening examination (a visual examination by a qualified observer) can take only a few minutes, is safe and acceptable, and is regarded by many as reliable in diagnostic situations. Such examinations can find a sizable minority of melanomas on the back and posterior legs (which cannot be viewed easily by the person with the lesion). Moreover, skin screening examinations also offer opportunities for personalized health education at a teachable moment [8,9]. Some biologic and clinical considerations, however, may dampen the effect of screening. A fraction of melanomas may be amelanotic, or clinically unrecognizable, reducing the sensitivity of skin examinations. The radial growth phase in some melanomas may be absent (eg, nodular melanoma) or of too short duration for detection by periodic screening. Screen-
ing for melanoma leads to many false-positives, and the discovery of other skin cancers (basal cell carcinoma and squamous cell cancer), which uncommonly lead to death. Whether early discovery of such nonmelanoma skin cancers can reduce morbidity is unproved. Screening can introduce a host of biases, such as lead-time bias or length bias, or uncover pseudodisease (thin nonmetastasizing form of melanoma) [10 13].
Defining criteria for successful control programs Reduction in melanoma mortality rates is the major goal for melanoma control. It should be noted that no randomized trial or definitive data about mortality reduction from screening exist; current recommendations for skin cancer screening vary greatly. Ideally, evaluating melanoma screening would involve a randomized trial demonstrating sustained reductions in melanoma mortality rates in a defined population (compared with a control population); however, such a trial would pose significant logistical challenges, requiring screening and follow-up of nearly a million persons for years (and perhaps even a decade or more), costing millions of dollars, and training a large medical force [1,2,12]. Intermediate end points other than mortality rates may include documenting fewer late-stage or thick lesions in a defined population. Trials must be able to monitor such measures accurately; national Surveillance, Epidemiology, and End Result Registry registries would need to be used because some state-based registries still do not record tumor thickness more than 10 years after an initial report [14]. One such study (using melanoma thickness) from west Scotland (an area with previously high melanoma mortality) found a trend toward reduction in both thick tumors ( P < .05) and mortality (not statistically tested) in women after the campaign. MacKie and Hole [15] noted that mortality was increasing for men and women before 1988, but then fell for women from 1988 to 1990. Other types of studies offer initial data to evaluate screening programs. Case-control studies, such as Berwick et als [16] study on self-screening for melanoma, can provide critical information. Still other evaluations tracking intermediate outcome measures (after initiating screening or education) provide initial data but require careful interpretation in light of possible biases. In an analysis of trends in Australia and New Zealand, Burton and Armstrong [17] emphasized the importance of measuring the reduction of thick tumors, because there has been a large increase in the incidence of very thin melanomas.
631
Other outcomes for suggesting benefit from melanoma control programs include reversing years of potential life lost per death for melanoma, and improving quality of life by detecting thinner, more curable melanoma (and thereby reducing levels of physical and psychosocial morbidity seen with recurrent melanoma) [18]. Economic savings represent yet another potential measure of evaluating a skin cancer control program. Tsao et al [19] provided a baseline estimate of melanoma-related costs in the United States. The annual direct cost of treating newly diagnosed melanoma in 1997 was estimated to be $563 million. Stage I and II disease each comprised about 5% of the total cost; stage III and IV disease consumed 34% and 55% of the total cost, respectively. About 90% of the total annual direct cost of treating melanoma in 1997 was attributable to less than 20% of patients with advanced disease.
Screening recommendations for melanoma Recommendations (1989 to 2001) from many influential organizations are reviewed next. No randomized trials have examined the efficacy of different strategies for the early detection of melanoma, although a new trial is currently being conducted in Australia [20]. Recommendations from the early 1990s to mid-decade were diverse but more recent recommendations have not endorsed population-based screening. It is worth noting that increased attention to evidence-based recommendations has resulted in changes in the US Preventive Services Task Force series of three recommendations from 1989 to 2001. The 1989 report recommended screening for high-risk populations (family or personal history of skin cancer, clinical evidence of precursor lesions, and those with increased occupational or recreational exposure to sunlight). The 1996 US Preventive Services Task Force found insufficient evidence to recommend for or against either routine screening for skin cancer by primary care providers [21]. Most recently, the Third United States Preventive Services Task Force (2001) concluded evidence is lacking that the skin examination by clinicians is effective in reducing mortality or morbidity from skin cancer [22]. Other eminent organizations have made screening recommendations and calls for increased attention to clinical skill-building. In successive years, the International Union Against Cancer recommended no skin cancer screening (1990) [23], whereas Healthy People 2000 recommended an increase (to at least 40%) in the proportion of persons aged 50 years and older visiting a primary care provider in the preceding year who
should receive skin examinations during one such visit (1991) [24]. The National Institutes of Health Consensus Conference on the Diagnosis and Treatment of Early Melanoma (1992) suggested that physicians and nurses involved in primary care specialties (pediatrics, internal medicine, family practice, and obstetrics and gynecology) should communicate skin cancer information to patients and their families and identify members of high-risk groups [25]. The National Conference to Develop a National Skin Cancer Agenda (1995) concluded that all graduates of medical and nursing schools should learn to perform a skilled examination for skin cancer in adults. The same conference also called for physicians to examine exposed areas of the skin as part of a total physical examination or during focused examination to evaluate specific lesions. They concluded that all physicians who screen for skin cancer should develop a standardized office protocol to ensure proper follow-up and biopsy (if necessary) of suspect lesions [26]. However, the Institute of Medicine report (2000) concluded that there is insufficient evidence to support positive or negative conclusions about the adoption of a new program of clinical screening of asymptomatic Medicare beneficiaries [27].
Identification of high-risk populations With evidence pointing away from endorsement of population-based screening programs, an important challenge to national programs for skin cancer education and screening is the identification of individuals who are at greater-than-average risk of developing melanoma and who might benefit most from screening. Because it allows effective targeting, this identification process would improve the success and acceptability of skin cancer educational and screening programs, while decreasing the societal cost. Melanoma risk factors In general, the most efficient screening programs target individuals at high risk for the disease (or death from the disease). For melanoma, special emphasis could include light-skinned persons who suntan poorly or burn easily (relative risk [RR] 2 to 3); those with a family history of melanoma (RR 2 to 8); higherthan-average numbers of nevi (RR 1 to 64); dysplastic nevi or atypical moles (RR 7 to 70); and excessive sun exposure (RR 3 to 5) [28]. Targeting high-risk persons should improve the predictive value of the visual examination as a
632
screening tool. There may be special value in screening by known risk factors, or easy-to-identify demographic variables, or a combination thereof. Groups at greatest risk of melanoma deaths, such as middleaged and older men, persons of low socioeconomic status, or the uninsured, are discussed next. Targeting by demographic variables Middle-aged and older men It is well documented that men, particularly those aged 50 years or older, have higher incidence and mortality rates for melanoma [29 32]. Of particular relevance to mortality rates, during the past decade, the US incidence of thick tumors ( > 4 mm) increased significantly only in men aged 60 years and older [33]. Nearly 50% of all melanoma deaths in the United States are in white men ages 50 years of age and above [34]; a recent population-based study using the Florida Tumor Registry found men to be more than twice as likely to have late-stage melanoma [35]. Birth-cohort adjusted incidence rates in the US show even more striking differences between men and women [36]. This may reflect less careful lifetime sun exposure, less sun protection, or exposure of a larger proportion of the body among men, although as yet unrecognized biologic factors may also contribute. In the future, as the population ages and melanoma rates continue to increase the relatively high morbidity and mortality of older men is expected to be even more striking [36]. Similar to the US, 50% of deaths from melanoma in New South Wales, Australia, occur in men over 50 years of age [37], prompting a call for improved opportunistic screening by Australian general practitioners at the time of medical visits for other reasons [38]. In Australia, the mortality for men aged 60 years and over has also increased, with a plateau in the rest of the population [37]. Higher mortality in US men may be attributable to their greater likelihood of presenting with regional and distant disease (14% of men versus 10% of women) ( P < .0001). Balch et al [39] found an overall survival disadvantage for men compared with women by analyzing more than 6500 melanoma patients at the University of Alabama at Birmingham and the Sydney Melanoma Unit and using a subsequent meta-analysis of 8500 patients worldwide. Men had thicker lesions (median 1.4 versus 1 mm) and more ulcerated and fewer extremity lesions than did women. In particular, men with back lesions had worse survival rates as compared with women ( P = 0.0163). Another retrospective study of 674 patients in a tertiary referral center in New South Wales, Australia, found that men
were more likely than women to be diagnosed with thick lesions (exceeding 3 mm) [40]. Social class Melanoma incidence rates correlate with higher social class status, attributed by some to more opportunities for recreational sun exposure. Five studies, however, have found disproportionate mortality among persons of lower socioeconomic status [41 44]. One study of melanoma cases and deaths in Massachusetts found case-fatality rates to be 50% higher in persons of low socioeconomic status [41]. Data from more than 28,000 cancer patients in Florida found that the uninsured were more likely than persons with insurance to be diagnosed with late-stage melanoma (odds ratio [OR] = 2.59, P = .004). Patients insured by Medicaid compared with commercial indemnity insurance were more likely diagnosed at a late stage of melanoma (OR = 4.69, P < .001) [45]. Special screenings of World War II veterans, particularly those from specific regions, may have some merit because it has been suggested that exposure to high levels of solar radiation in young adulthood, in the World War II Pacific and European Theaters, was associated with a higher risk of melanoma mortality [46].
Current venues in melanoma screening and education Dermatologist-delivered skin cancer screening (United States) The annual American Academy of Dermatology (AAD) screening programs have provided free screening and education to more than 1 million Americans, with 1.2 million screenings being performed since 1985 [47]. In the spring of each year, local and national media publicize risk factors and warning signs for melanoma and skin cancer [48]. The yield and predictive value of confirmed melanoma according to demographics and risk factors recorded during screening have been demonstrated. Middle-aged and older men ( ! 50 years) comprised 25.2% of AAD screenees but 44.4% of confirmed cases. The overall yield of melanoma (number of confirmed cases per number of screenees) was 1.5 per 1000 (363 of 242,374) in contrast to 2.6 per 1000 among men greater than or equal to 50 years. The yield was further improved for men greater than or equal to 50 years who reported either a changing mole (4.6 per 1000) or skin type I and II (3.8 per 1000)
633
Fig. 1. Yield of melanoma (363 cases) by age, sex, and risk factor in American Academy of Dermatology screening programs (1992 to 1994). N = 242,374. Yield per 1000.
(Fig. 1). The predictive value of a screening diagnosis of melanoma was more than twice as high for men greater than or equal to 50 years with either a changing mole or skin type I and II compared with all other participants [49]. The AAD programs have detected thinner lesions than have been recorded in population-based registries; however, these programs are conducted among self-selected populations and subject to bias [48]. More than 98% of confirmed melanomas in AAD screenings (1992 to 1994) were localized and more than 90% were in situ or thin lesions. The 8.3% of AAD cases with advanced melanoma is a lower proportion than that reported by the 1990 Surveillance, Epidemiology, and End Result Registry [48]. Of persons with a confirmed melanoma, 39% indicated that without the free program, they would not have considered having a physician examine their skin. Evaluation of AAD programs found that the predictive value positive of a visual examination by a dermatologist for melanoma was 27% to 31% in Massachusetts screenings (1986 to 1989) and 17% in national AAD screenings (1992 to 1994) [48,50]. Rampen et al [51] linked information on 1551 persons with a negative screening result with populationbased registries (during 42 months of follow-up), and they found 15 persons with new skin cancers.
Of these, three were present at the 1990 screening, and most likely were missed; 12 were considered to be new lesions. No melanomas were found among the missed cases. Available data suggest that the cost-effectiveness of such screenings is comparable with that of other cancer screenings, including breast cancer [52]. Freedberg et al [52] found that AAD skin cancer screening increased both life expectancy and quality-adjusted life expectancy. Although the AAD screen could not be compared directly with other cancer screening cost-effectiveness studies, a onetime skin cancer screening was generally comparable in cost-effectiveness with screening for breast cancer in women aged 55 to 65 years. There were strong benefits for men 50 years of age and older. Screening programs also provide essential information, both before and during the screening process. It is estimated that millions have heard about warning signs and received educational materials, particularly at a time when the screenee is receptive to education about skin cancer [9,47,48]. A single study of Massachusetts screenees attempted to quantify the educational effect of screenings. In an analysis of 643 Massachusetts residents with positive screening diagnoses, rates of self-screening improved from 60% before screening to 84% after
634
screening; physician examinations improved from 39% to 55% [53]. Although the AAD screening program has provided a major public health service to high-risk Americans with little access to dermatologic care and service, future efforts could seek to improve rates of complete examinations and provide care to the uninsured. Rigel et al [54] reported that screenees having complete skin examinations were 6.4 times more likely to have a melanoma detected than those having partial examinations. Academy programs must make special efforts to assist individuals without insurance. Jonna et al [55] found noncompliance to seek recommended follow-up care to be associated with lack of health insurance for men too young to participate in Medicare and suggested that approaches to definitive work-up for uninsured patients be developed alongside free skin cancer screening programs. Dermatologist-delivered skin cancer screening (international) In general, international screening efforts have yielded similar results to those found in the United States. Rampen et al [56] yielded similar findings in a screening of 2564 self-selected participants in The Netherlands. Nine melanomas were pathologically confirmed. Hoffman et al [57] reported on screening activities in Bochum, Germany, that resulted in 14 melanomas being diagnosed in a screening population of 1467. Yield in melanoma screening among 520 residents of British Columbia, Canada, for the period 1994 and 1995 were virtually identical to US studies [58]. A French study in the occupational setting demonstrated the efficiency of this maneuver [59]. Rossi et al [60] reported on programs with intensive prescreening outreach in Italy. Screening organizers produced 90,000 leaflets for a population of 243,000, leading to 2050 individuals attending screening programs. The sensitivity and predictive value of the visual examination were 93% and 7%, respectively. Screening in the primary care setting Screening data Despite increased publicity and increased attention to screening during the 1990s, recent Federal data indicate that the proportion of Americans reporting that they had ever received a screening examination remained static at 21% from 1992 to 1998 [61,62]. Screening rates among the more than 30,000 people completing the National Health Interview Survey Cancer Control Supplement (1992, 1998) were lowest
among younger persons, people with less than a high school degree, and those living close to the poverty line [62]. Other national sources point to low screening uptake. In an analysis of the National Ambulatory Medical Care Survey of office-based physician visits during 1997, Oliveria et al [63] found that skin cancer screening activities were much lower than other cancer screening activities. Screening examinations were performed in only 16% of instances. Most respondents in a telephone sample of Rhode Island residents reported that their health care provider never or rarely looked at the areas of their skin in which melanoma is most likely to arise [64]. Rationale Melanoma patients typically have contacts with physicians in the year before diagnosis, suggesting many lesions might be diagnosed earlier. Routine examinations provide an excellent opportunity for melanoma detection. In one study of 216 melanoma cases, 87% had regular physicians and 63% had seen those physicians in the year before diagnosis, but only 24% of patients had examined their own skin before diagnosis and 20% reported physician skin examinations (although data on patient recollection of physician skin examinations may be an underestimate) [65]. Fewer than 20% of Americans have a dermatologist [47], whereas approximately 85% see a physician every 2 years [66]. In a single study, physiciandetected melanoma, compared with patient or family detection, was associated with an increase in the probability of detecting thinner ( < 0.75 mm) melanomas (RR 4.2; 95% confidence interval, 1.4 to 11.1; P = .01) [67]. An Australian study of the costeffectiveness of every 5-year screening for melanoma by family practice physicians for men over the age of 50 found a cost-effectiveness of Australia $6900 per year of life saved for men [68]. Encouraging providers to educate their patients about risk is a key because the public generally has a very low level of awareness about melanoma. In a nationally representative study of 1001 adults, Miller et al [69] found that only 34% of Americans knew that melanoma was some form of skin cancer, and only 26% of those who were aware could identify its specific signs. Education should emphasize heightened awareness of the most common risk factors, information about the relative ease of treatment, the need for prompt medical attention, and the high chance of cure for an early lesion. The media can help disseminate this information, either in written form (newspapers and magazines); television; or radio. May is skin cancer prevention month in the
635
United States, which presents an ideal opportunity for increased media attention and public education [47]. Performance of physicians in screening and education Early detection by physicians should be exceedingly important, but their ability to detect melanoma has not been well studied. Screening performance has very important implications if physicians are to be involved in randomized trials testing the efficacy of this procedure or if they are to serve as gatekeepers in large managed care organizations. Little is known about physician skin examination practice; data are needed on the performance or quality of the skin cancer examination during the routine medical visit, on the teaching of skin cancer education in American medical schools, and on the proportion of physicians who assess risk factors. Health care providers should be encouraged to refer suspicious lesions to dermatologists, because recognition of melanoma by nondermatologists is not high [70]. Several interventions have been designed to improve general practitioners and primary care physicians proficiency in their diagnosis of skin cancer. Successful incorporation of skin cancer screening may first require training for medical students, physicians-in-training, and nondermatologist physicians, because few medical professionals have specific education in early detection of melanoma. In a study at one US medical school, Geller et al [71] found that 28% of graduating students had not observed, 40% had not been trained, and 35% had not practiced a skin cancer examination. In another study, 89% of second-year students were capable of identifying a clinical photograph of a melanoma; however, only one fourth-year student of 285 detected a melanoma on a standardized patient [72]. Whited and Grichnik [73] estimated physicians assessments for detecting the presence or absence of melanoma have a specificity of 96% to 99%, whereas sensitivity ranges widely from 50% to 97%. A survey of 666 Victoria, Australia, general practitioners (along with a color photographic quiz of 12 examples of benign and malignant lesions) found them to be better at recognizing the 12 skin tumors as suspicious or benign than they were at providing the correct diagnosis. Benign lesions were diagnosed more proficiently (75%) than were skin cancers (61%) or dysplastic lesions (61%) [74]. Dermatologist versus nondermatologist performance in screening examinations A number of studies have compared dermatologist and nondermatologist physician performance in the
clinical examination. One well-designed prospective study of the accuracy of total-body skin examination found that skin cancer specialists decisions about biopsy were more sensitive and far more specific than those of general practitioners. The sensitivity of totalbody skin examination for detecting suspicious lesions was 0.72 for the general practitioners versus 0.97 for the four skin specialists [74]. Cassileth et al [70] showed that only 12% of nondermatologists could correctly identify at least five of six melanomas, as compared with 69% of dermatologists. Gerbert et al [75] examined (1) the readiness of primary care physicians optimally to triage lesions suspicious for skin cancer, (2) the difference in their abilities from those of dermatologists, and (3) whether accurate diagnosis after viewing slide images transfers to accurate diagnosis after viewing lesions on patients. Dermatologist scores were nearly double those of primary care residents, and primary care residents performance was positively associated with previous experience in dermatology. Primary care residents failed 50% of the time to diagnose correctly nonmelanoma skin cancer and malignant melanoma. The authors concluded that primary care residents were not ready to assume a gatekeeper role for lesions suspicious of skin cancer [75]. Targeted screening A precedent for screening high-risk persons has been developed with the identification, screening, and education of persons with multiple nevi or atypical moles and dysplastic nevi, particularly those within the familial melanoma setting [76,77]. Screening, surveillance, and educational programs have been recommended for members of hereditary melanoma kindreds with dysplastic nevi, who are very high risk for melanoma. Masri et al [76] found thinner melanomas in surveillance incident cases compared with the index case, 0.52 mm compared with 1.44 mm, respectively. Marghoob et al [78] suggested that targeting certain individuals, including those with atypical mole syndrome, could result in higher yields of melanoma detection. Skin self-examination The AAD educational programs also encourage Americans to perform SSE for the early detection of cancer. Self-screening or examination of moles is also an important means of detecting potentially serious changes that might indicate carcinogenesis. In a major case-control trial to investigate whether SSE is associated with a decreased risk of lethal melanoma,
636
Berwick et al [16] compared 650 Connecticut cases with 549 controls. SSE, although practiced by only 15% of all subjects, was associated with a reduced risk of melanoma (OR = 0.66; 95% CI, 0.44 to 0.99). The data indicated further that SSE may reduce the risk of advanced disease among melanoma patients (unadjusted risk ratio = 0.58; 95% CI, 0.31 to 1.11). Education to shift public perception of the need for SSE may be required because a single representative sample in the United States found that persons performing SSE were most likely to be more than 25 years of age, white, female, with an education beyond high school [79]. Community studies in Australia have also reported lower prevalence of SSE in men and persons with lower socioeconomic status [80]. Robinson et al [79] also found that self-perceived high risk of the development of melanoma and discussions with physicians or nurses about skin cancer prevention promoted performance of SSE. Careful instruction for SSE resulted in relatively high sensitivity ranges of 0.57 to 0.79 and specificity of 0.88 to 0.97 for patient counts of moles on the trunk or total body [81].
public. Pretrial evaluation determined that 22% of the Queensland adult population had been screened. The study will entail 3 years of screening or normal care followed by 12 years tracking deaths. Screening components include community education (booklets and pamphlets); letters of invitation; local media; publicity led by local cricket stars; education of doctors; and screening services, often led by specially trained physicians in skin cancer screening clinics. In the pilot phase, 44 towns with 560,000 adults have been divided into 22 screening towns (290,000) versus 22 control towns (270,000). Study organizers set goals and successfully screened 60% of adults in the 44 towns [20]. Physician education has consisted of providing educational resources, clinical audits, seminars, urging whole-body examinations, and improving diagnostic skills. Plans are underway to expand the program to more towns. Others have argued that more emphasis should be placed on education rather than screening. Melia et al [82] concluded that more reasonable strategies might entail improving skin self-awareness rather than screening. Targeting specific populations and use of risk-assessment strategies Because risk is disproportionately higher among certain populations, specific interventions, particularly with middle-aged and older men, might be worthwhile. Venues for screening and education for middle-aged and older men include the following: Physicians office Public health clinics American Academy of Dermatology national screenings Recreational settings Workplaces National media efforts Family education by spouse and significant others Corporate sponsorship in retail outlets frequented by men In an Australian study, men over 50 years of age, including many with multiple risk factors, responded to personalized health messages about melanoma [83]. Jackson et al [84] identified a group at high risk for melanoma and found good agreement when the patients risk scores were compared with results of the clinical skin examination. They concluded that a risk score was potentially useful in targeting early detection in general practice.
New strategies This section suggests research and investigation into the following strategies: population-based screening programs, development of risk-assessment strategies, and training of medical personnel. Population-based screening programs As discussed previously, the litmus test for screening acceptance is a randomized control trial. Population-based screening programs would be fraught with many difficulties. For example, Elwood [12] examined the options for conducting a randomized trial of screening for melanoma and found that about 21,000 people need to be screened to prevent one death. There has been no such study in the United States; US cancer control planners must look to results from the ongoing skin cancer screening study in Queensland, Australia. Background data reveal that the incidence rate in Queensland is 49 per 100,000; 2100 cases of melanoma were diagnosed and 220 deaths were confirmed in 1999 [20]. The Queensland lifetime risk is 1 per 16 in men and 1 per 24 in women [20]. In response to the burgeoning epidemic, Queensland cancer authorities have recently launched a population-based skin cancer screening program. Their efforts build on the four-pronged Sun Smart program with vital components for education regarding early detection, for both physicians and the
637
Although it is unproved, skin cancer screening using a risk assessment measure with examination or referral of high-risk patients may be promising for addressing the excess burden of disease in older persons. Because most elderly individuals consult a physician at least yearly, case finding by clinicians focusing on the elderly may reach vulnerable individuals. The aforementioned Institute of Medicine and United States Preventive Services Task Force reports called for studies to help the clinician identify patients, especially the elderly, at high risk for melanoma [22,27], and noted that evidence does support benefits of early melanoma detection and treatment as a part of usual medical care and that clinicians and patients should continue to be alert to the common signs of skin cancer, with a particular emphasis on older white males and on melanoma [22,27]. The report concluded that the major challenge related to the Medicare population is reaching the group at highest risk of death from skin cancer, specifically older fair-skinned men. Training of physicians and nurses In a randomized trial, Gerbert et al [85] tested a brief multicomponent educational intervention designed to improve the skin cancer diagnosis and evaluation planning of primary care residents to a level equivalent to that of dermatologists. The intervention consisted of face-to-face individualized feedback sessions, an interactive seminar, and provision of educational materials and tools. Results showed that primary care residents who participated in the intervention performed significantly better than the control group on overall diagnosis and evaluation planning of all three cancerous lesion categories. The findings suggest that a targeted and multicomponent educational intervention aimed at primary care residents could proficiently triage cancerous lesions. The authors suggested other less labor-intensive avenues of training residents in evaluating skin lesions, such as computer-based applications as a teaching tool. Further evaluation of the cost-effectiveness and feasibility of this type of training needs to be done. Other studies of interventions aimed at health professionals confirm the need for multicomponent and continued training. Dolan et al [86] evaluated a brief intervention consisting of two 1-hour educational seminars on skin cancer control. Physicians from the intervention group were more likely to advise at-risk patients to watch their moles, but they did not significantly improve their performance of other skin cancer control practices or their ability to name and make treatment decisions about skin lesions. An
evaluation of a continuing education module designed for nurses by McCormick et al [87] found improvement in their knowledge and self-efficacy to screen and educate patients about skin cancer. Nurses reported, however, that organizational variables, such as time, support, and resources, played an important role in their ability to carry out screening activities. Strategies for creating organizational change to increase the likelihood that screening and patient education will be available for those at risk must be considered as part of equipping health professionals in skin cancer control and prevention. Recently, using the consensus clinical diagnosis of dermatologists as the gold standard, Oliveria et al [88] found nurse practitioner (N = 5) sensitivity ranged from 50% to 100% and detection specificity was 99% to 100%. They concluded that nurse practitioners could be trained to identify and triage suspicious lesions accurately.
References
[1] Koh HK, Geller AC, Miller DR, Grossbart TA, Lew RA. Prevention and early detection strategies for melanoma and skin cancer. Arch Dermatol 1996;132:436 43. [2] Koh HK, Geller AC, Lew RA. Melanoma. In: Kramer BS, Gohagan JK, Prorok PC. Cancer screening theory and practice. New York: Marcel Dekker; 1999. p. 379 408. [3] Koh H, Lew R, Prout M. Screening for melanoma skin/ cancer: theoretic and practical considerations. J Am Acad Dermatol 1989;20:159 75. [4] Miller A. Screening for cancer. Orlando, FL: Academic Press; 1985. [5] Davis N. Modern concepts of melanoma and its management. Ann Plast Surg 1978;1:628 9. [6] Friedman RJ, Rigel DS, Silverman MK, Kopf AW, Vossaert KA. Malignant melanoma in the 1990s: the continued importance of early detection and the role of physician examination and self-examination of the skin. CA Cancer J Clin 1991;41:201 26. [7] Howe HL, Wingo PA, Thun MJ, et al. Annual report to the nation on the status of cancer (1973 through 1998), featuring cancers with recent increasing trends. J Natl Cancer Inst 2001;93:824 42 [also accessed at www. seer.cancer. gov.]. [8] Koh HK. Cutaneous melanoma. N Engl J Med 1991; 325:171 82. [9] Rigel DS. Is the ounce of screening and prevention for skin cancer worth the pound of cure? CA Cancer J Clin 1998;48:236 8. [10] Burton R, Armstrong B. Recent incidence trends imply a nonmetastasizing form of invasive melanoma. Melanoma Res 1994;4:107 13. [11] Swerlick RA, Chen S. The melanoma epidemic: is increased surveillance the solution or the problem? Arch Dermatol 1996;132:881 4.
638
A.C. Geller / Dermatol Clin 20 (2002) 629640 the presentation of patients with thick primary melanomas. Med J Aust 1991;154:583 7. Cooke KR, McNoe BM. Targeting early detection of malignant melanoma of the skin. N Z Med J 1991;103: 551 3. Wagener DK. Patterns of melanoma deaths in the United States. Ann N Y Acad Sci 1990;609:252 6. Jemal A, Devesa SS, Hartge P, Tucker MA. Recent trends in cutaneous melanoma incidence among whites in the United States. J Natl Cancer Inst 2001; 93:678 83. National Center for Health Statistics. Vital statistics mortality data multiple cause of death detail (machine-readable public use data tape). Hyattsville, MD: US Department of Health and Human Services, Public Health Service, CDC; 1997. Van Durme DJ, Ferrante JM, Pal N, et al. Demographic predictors of melanoma stage at diagnosis. Arch Fam Med 2000;9:606 11. Dennis LK. Increasing risk of melanoma with increasing age. JAMA 1999;282:1037 38. Hanrahan PF, Hersey P, DEste CA. Factors involved in presentation of older people with thick melanoma. Med J Aust 1998;169:410 4. Kelly JW. Melanoma in the elderly: a neglected public health challenge. Med J Aust 1998;169:403 4. Balch C, Soong S, Shaw H. An analysis of prognostic factors in 8500 patients with cutaneous melanoma. In: Balch C, Houghton A, Milton G, editors. Cutaneous melanoma. Philadelphia: Lippincott; 1992. p. 165 87. Bonnett A, Roder D, Esterman A. Epidemiological features of melanoma in South Australia: implications for cancer control. Med J Aust 1989;151:502 9. Geller AC, Miller DR, Lew RA, Clapp RW, Wenneker MB, Koh HK. Cutaneous melanoma mortality among the socioeconomically disadvantaged in Massachusetts. Am J Public Health 1996;86:538 43. MacKie RM, Hole DJ. Incidence and thickness of primary tumours and survival of patients with cutaneous malignant melanoma in relation to socioeconomic status [see comments]. BMJ 1996;312:1125 8. Shaw HM. Cutaneous malignant melanoma: occupation and prognosis. Med J Aust 1981;1:37 8. Vagero D, Persson G. Risks, survival and trends of malignant melanoma among white and blue collar workers in Sweden. Soc Sci Med 1984;19:475 8. Roetzheim RG, Pal N, Tennant C, et al. Effects of health insurance and race on early detection of cancer. J Natl Cancer Inst 1999;91:1409 15. Page WF, Whiteman D, Murphy M. A comparison of melanoma mortality among WWII veterans of the Pacific and European theaters. Ann Epidemiol 2000;10: 192 5. American Academy of Dermatology.Schaumburg IL. Koh HK, Norton LA, Geller AC, et al. Evaluation of the American Academy of Dermatologys National Skin Cancer Early Detection and Screening Program. J Am Acad Dermatol 1996;34:971 8. Geller AC, Miller DR, Halpern A. Targeting strategies
[12] Elwood JM. Screening for melanoma and options for its evaluation. J Med Screen 1994;1:22 38. [13] Andersen WK, Silvers DN. Melanoma? It cant be melanoma. A subset of melanomas that defines clinical recognition. JAMA 1991;266:3463 5. [14] Koh HK, Adame N, Geller AC, et al. Cancer registry data on melanomas [letter]. N Engl J Med 1990;323: 921 2. [15] MacKie RM, Hole D. Audit of public education campaign to encourage earlier detection of malignant melanoma. BMJ 1992;304:1012 5. [16] Berwick M, Begg CB, Fine JA, Roush GC, Barnhill RL. Screening for cutaneous melanoma by skin selfexamination. J Natl Cancer Inst 1996;88:17 23. [17] Burton RC, Armstrong BK. Non-metastasizing melanoma? J Surg Oncol 1998;167:180 1. [18] Albert VA, Koh HK, Geller AC, Miller DR, Prout MN, Lew RA. Years of potential life lost: another indicator of the impact of cutaneous malignant melanoma on society. J Am Acad Dermatol 1990;23:308 10. [19] Tsao H, Rogers G, Sober A. An estimate of the annual direct cost of treating cutaneous melanoma. J Am Acad Dermatol 1998;38:669 80. [20] Aitken J. Skin cancer screening in Australia. Presented at the National Institutes of Health, Bethesda, MD, July 2001. [21] US Preventive Services Task Force. Guide to clinical preventive services. 2nd edition. Baltimore: Williams Wilkins; 1996. [22] United States Preventive Services Task Force. Screening for Skin Cancer. Am J Prev Med Special Supplement April 2001 [23] Marks R, Hill D. The public health approach to melanoma control: prevention and early detection. Geneva, Switzerland: International Union Against Cancer; 1992. [24] Healthy people. National health promotion and disease prevention objectives. Washington, DC: US Department of Health and Human Services Public Health Service; 1991. [25] NIH Consensus Conference. Diagnosis and treatment of early melanoma. JAMA 1992;268:1314 9. [26] Goldsmith L, Koh HK, Bewerse B, Reilley B, Wyatt SW, Bergfeld W, et al. Proceeding from the conference to develop a national skin cancer agenda (executive summary). American Academy of Dermatology and Centers for Disease Control and Prevention. April 8 10, 1995. J Am Acad Dermatol 1996;34:822 23. [27] Institute of Medicine. Extending Medicare coverage for prevention and other services. Washington, DC: National Academy Press; 2000. [28] Rhodes AR, Weinstock MA, Fitzpatrick TB, et al. Risk factors for cutaneous melanoma: a practical method of recognizing predisposed individuals. JAMA 1987;258: 3146 54. [29] Anonymous. Boston University School of Medicine. Death rates of malignant melanoma among white men United States, 1973 1988. MMWR Morb Mortal Wkly Rep 1992;41:20 1. [30] Hersey P, Sillar RW, Howe CG, et al. Factors related to
[31]
[32] [33]
[34]
[35]
[36] [37]
[38] [39]
[40]
[41]
[42]
[43] [44]
[45]
[46]
[47] [48]
[49]
A.C. Geller / Dermatol Clin 20 (2002) 629640 for skin cancer detection in the United States [abstract]. Presented at the World Congress on Melanoma. Venice, Italy, February, 2001. Koh HK, Geller AC, Miller DR, Caruso A, Gage I, Lew RA. Who is being screened for melanoma/skin cancer? J Am Acad Dermatol 1991;24:271 7. Rampen FHJ, Casparie-van Velsen IJ, van Huystee BE, et al. False-negative findings in skin cancer and melanoma screening. J Am Acad Dermatol 1995;33: 59 63. Freedberg KA, Geller AC, Miller DR, Lew RA, Koh HK. Screening for malignant melanoma: a cost-effectiveness analysis. J Am Acad Dermatol 1999;41: 738 45. Geller AC, Halpern AC, Sun T, et al. Participant satisfaction and value in American Academy of Dermatology and American Cancer Society skin cancer screening programs in Massachusetts. J Am Acad Dermatol 1999;40:563 6. Rigel DS, Friedman RJ, Kopf AW, et al. Importance of complete cutaneous examination for the detection of malignant melanoma. J Am Acad Dermatol 1986; 14(5):857 60. Jonna BP, Delfino RJ, Newman WG, Tope WD. Positive predictive value for presumptive diagnoses of skin cancer and compliance among patients attending a community screening program. Prev Med 1998;27: 611 6. Rampen FHJ, van Huysteee EWL, Kiemeney LALM. Melanoma/skin cancer screening clinics: experience in the Netherlands. J Am Acad Dermatol 1991;25:776 7. Hoffman K, Dirschka T, Schatz H, et al. A local education campaign on early diagnosis of malignant melanoma. Eur J Epidemiol 1993;9:591 8. Engleberg D, Gallagher RP, Rivers JK. Follow-up and evaluation of skin cancer screening in British Columbia. J Am Acad Dermatol 1999;41:37 42. Guibert P, Mollat F, Ligen M, Dreno B. Melanoma screening: report of a survey in occupational medicine. Arch Dermatol 2000;136:199 202. Rossi CR, Vecchiato A, Bezze G, et al. Early detection of melanoma: an educational campaign in Padova, Italy. Melanoma Res 2000;10:181 7. Hall HI, May DS, Lew RA, Koh HK, Nadel M. Sun protection behaviors of the US white population. Prev Med 1997;26:401 7. National Center for Health Statistics. National Health Interview Survey of topics related to cancer epidemiology, 1998. Hyattsville, MD: National Center for Health Statistics; 1998. Oliveria SA, Christos PJ, Halpern AC. Skin cancer screening and prevention in the primary care setting: national ambulatory medical care survey 1997. J Gen Intern Med 2001;16:297 301. Weinstock MA, Martin RA, Risica PM, et al. Thorough skin examinations for the early detection of melanoma. Am J Prev Med 1999;17:169 75. Koh HK, Miller DR, Geller AC, Clapp RW, Mercer MG, Lew RA. Who discovers melanoma? Patterns
639
[66]
[50]
[67]
[51]
[68]
[52]
[69]
[53]
[70]
[54]
[71]
[55]
[72]
[73] [74]
[56]
[57]
[75]
[58]
[59]
[76]
[60]
[77]
[61]
[62]
[78]
[63]
[79]
[64]
[80]
[65]
[81]
from a population-based survey. J Am Acad Dermatol 1992;26:914 9. Health US. Socioeconomic status and health chartbook. US Department of Health and Human Services, DHHS Publication Number (PHS) 98. 1998. Epstein DS, Lange JR, Gruber SB, et al. Is physician detection associated with thinner melanomas? JAMA 1999;281:640 3. Girgis A, Clarke P, Burton RC, Sanson-Fisher RW. Screening for melanoma by primary health care physicians: a cost-effectiveness analysis. J Med Screen 1996; 3:47 53. Miller DR, Geller AC, Wyatt SW, et al. Melanoma awareness and self-examination practices: results of a United States survey. J Am Acad Dermatol 1996;34: 962 70. Cassileth BR, Clark Jr. WH, Lusk EJ, Frederick BE, Thompson CJ, Walsh WP. How well do physicians recognize melanoma and other problem lesions? J Am Acad Dermatol 1986;14:555 60. Geller AC, Venna S, Prout MN, Miller DR, Demierre MF, Koh HK, et al. Should the skin cancer examination be taught in medical school? Arch Dermatol 2002;138(in press, September 2002). Robinson JK, McGahie WC. Skin cancer detection in clinical practice examination with standardized patients. J Am Acad Dermatol 1996;34:709 11. Whited JD, Grichnik JM. Does this patient have a mole or a melanoma? JAMA 1998;279:696 701. Burton RC, Howe C, Adamson L, et al. General practitioner screening for melanoma: sensitivity, specificity, and effect of training. J Med Screen 1998;5: 156 61. Gerbert B, Maurer T, Berger T, Pantilat S, McPhee SJ, Wolff M, et al. Primary care physicians as gatekeepers in managed care: primary care physicians and dermatologists skills at secondary prevention of skin cancer. Arch Dermatol 1996;132:1030 8. Masri GD, Clark Jr. WH, Guerry D, et al. Screening and surveillance of patients at high risk for malignant melanoma result in detection of earlier disease. J Am Acad Dermatol 1990;22:1042 8. Vasen HF, Bergman W, van Haeringen A, Scheffer E, van Slooten EA. The familial dysplastic nevus syndrome: natural history and the impact of screening on prognosis. A study of nine families in the Netherlands. Eur J Cancer Clin Oncol 1989;25:337 41. Marghoob AA, Kopf AW, Rigel DS, et al. Risk of cutaneous malignant melanoma in patients with classic atypical-mole syndrome. Arch Dermatol 1994;130: 993 8. Robinson JK, Rigel DS, Amonette RA. What promotes skin self-examination? J Am Acad Dermatol 1998;38 (part 1):752 7. Baade PD, Balanda KP, Stanton WR, Gillespie AM, Lowe JB. Community perceptions about the important signs of early melanoma. J Am Acad Dermatol 1997; 36:33 9. Lawson DD, Moore DH, Schneider JS, et al. Nevus
640
A.C. Geller / Dermatol Clin 20 (2002) 629640 [85] Gerbert B, Bronstone A, Wolff M, Maurer T, Berger T, Pantilat S, et al. Improving primary care residents proficiency in the diagnosis of skin cancer. J Gen Intern Med 1998;13:91 7. [86] Dolan NC, Martin GJ, Robinson JK, Rademaker AW. Skin cancer control practices among physicians in a university general medicine practice. J Gen Intern Med 1995;10:515 9. [87] McCormick L, Masse L, Cummings S, Burke C. Evaluation of a skin cancer prevention module for nurses: change in knowledge, self-efficacy, and attitudes. Am J Health Promot 1999;13:282 9. [88] Oliveria SA, Nehal KS, Christos PJ, et al. Using nurse practitioners for skin cancer screening: a pilot study. Am J Prev Med 2001;21:214 7.
counting as a risk factor for melanoma: comparison of self-count with count by physician. J Am Acad Dermatol 1994;31:438 44. [82] Melia J, Harland C, Moss S, et al. Feasibility of targeted early detection for melanoma: a population-based screening study. Br J Cancer 2000;82: 1605 9. [83] Hanrahan PF, Menzies SW, DEste CA, et al. Participation of older males in a study on photography as an aid to early detection of melanoma. Aust N Z J Public Health 2000;24:615 8. [84] Jackson A, Wilkinson C, Pill R. Moles and melanoma: whos at risk, who knows, and who cares? A strategy to inform those at high risk. Br J Gen Pract 1999;49: 199 203.
Dermatol Clin 20 (2002) 641 646
Dermoscopy: a review
Caron M. Grin, MD*, Kent P. Friedman, MD, Jane M. Grant-Kels, MD
Department of Dermatology, University of Connecticut Health Center, 263 Farmington Avenue, Farmington, CT 06032, USA
Dermoscopy is a noninvasive, in vivo technique to examine structures that lie beneath the skin surface. This technique has many synonyms including epiluminescence microscopy, skin-surface microscopy, incident light microscopy, and dermatoscopy. With the use of oil immersion, illumination, and magnification, one can observe subsurface structures that are not visible to the naked eye. Various instruments have been used for microscopy including high-magnification stereomicroscopes, lower-magnification handheld instruments, and digital imaging equipment.
History Skin surface microscopy dates back to 1663, according to Stolz et al [1], when Johan Kolhaus first looked at nail fold vessels with a microscopy. Unna published a paper in 1893 entitled Diaskopie, describing how immersion oil was used together with a microscope for skin surface microscopy. Saphier then published several papers from 1920 to 1921 in which the term dermatoskopie was used for the first time. In the 1950s, Goldman [2] applied this technique to the evaluation of melanocytic nevi and melanoma. It was not until 1971, however, that Mackie [3] reported on the improvement of the preoperative diagnosis of pigmented lesions with the use of surface microscopy. In the 1980s, several authors including Fritsch and Pechlaner [4], Pehamberger et al [5], and Soyer et al [6], published articles on dermoscopy. In 1989, a consensus conference was held by the Committee on Analytical Morphology in Hamburg, Germany, to
develop standardized terminology for this technique. The results of this Consensus were then published by Bahmer et al [7] in 1990. Throughout the 1990s, several groups developed different diagnostic methods for analyzing dermoscopic images [8 12]. With the knowledge of these methods and the availability of the hand-held dermatoscope, the technique of dermoscopy became readily available for use by dermatologists in clinical practice around the world. This past year, in an attempt further to refine this technique, the Consensus Net Meeting on Dermoscopy was held. The goals of this second Consensus Meeting were further to refine the definitions of dermoscopic structures and to prove the validity of a two-step procedure for dermoscopic classification of pigmented skin lesions [13]. From July until November 2000, 40 participants from various countries (Australia, United States, Germany, Italy, Austria, Argentina, Slovenia, Spain, Mexico, Japan, Switzerland, Turkey, and Sweden) evaluated dermoscopic images transmitted over the Internet. A training set of 20 cases was reviewed by each member first as a tutorial. Then each member evaluated 108 pigmented skin lesions for the presence of various dermoscopic features. Each of the following four methods of analysis were used to make a diagnosis: (1) pattern analysis, (2) ABCD rule, (3) Menzies method, and (4) seven-point checklist. The results of the Consensus Net Meeting on Dermoscopy were then presented at the First World Congress of Dermoscopy in Rome in February, 2001.
Dermoscopic structures Dermoscopy has given a greater understanding of clinical morphology. Subsurface structures in the epidermis, dermoepidermal junction, and papillary
* Corresponding author. E-mail address: grin@nso1.uchc.edu (C.M. Grin).
642
C.M. Grin et al. / Dermatol Clin 20 (2002) 641646
dermis can be seen with dermoscopy that are not visible to the naked eye. Below is the current working terminology for dermoscopic structures based on the recent Consensus Net Meeting on Dermoscopy [13]. As with all definitions, as more study and research are done in this area, these too will be subject to further change. Pigment network is a regular honeycomb grid of brown to black lines over a diffuse lighter background. Branched streaks result from a disturbed pigment network [1]. They are broadened, branched lines, which result from the remnants of a broken pigment network. According to Stolz et al [1], if these branched streaks extend contiguously to skin, these structures are called pseudopods. Dots and globules are sharply circumscribed, usually round or oval, variably sized structures that may be brown, black, or gray in color. Dots have been defined as less than 0.1mm. Streaks are linear structures, regularly or irregularly distributed at the edge of the lesion and are not clearly associated with the pigment network. They may be light brown to black and found within the lesion or at the periphery. These may be referred to as radial streaming or pseudopods by different authors. The distinction is that in radial streaming, there are radially oriented linear extensions at the periphery of the lesion, whereas pseudopods are bulbous, fingerlike projections at the periphery. Structureless areas are devoid of dermoscopic structures. These can be brown, black, or gray hyperpigmented areas and are referred to as blotches. If they are areas of decreased pigmentation, they are referred to as hypopigmented areas. Blue-gray areas have multiple blue-gray dots, which gives a speckled or peppered appearance. White scar-like depigmentation is a fairly welldefined pure white area. Blue-white veil is a bluish white ground-glass film overlying an area of diffuse pigmentation. This dermoscopic feature is not a feature of hemangiomas, basal cell carcinomas, or blue nevi. Milia-like cysts are white-yellow round areas. They have also been referred to as horny pseudocysts. Comedolike openings are brown-yellow or brown-black, circumscribed, crater-like structures. They can be round, oval, or irregularly shaped. They have been referred to as pseudofollicular openings.Fissures are a variation of
these comedolike structures corresponding to wedge-shaped invaginations of the epidermis Leaf-like areas are brown to blue-gray pigmented areas, usually at the periphery, which together form a maple leaf pattern Arborizing vessels are treelike branching vessels. Blue-gray ovoid nests are well circumscribed, bluegray, oval, or elongated areas Red-blue lacunae are sharply demarcated, round to oval, reddish or bluish structures. Central white patch is a relatively well circumscribed white area with pigment network at the periphery. There is a close correlation between these dermoscopic structures and what one sees under the microscope. This has been documented by various studies using step-sectioned specimens and dermoscopic images to correlate histopathologic changes with specific dermoscopic structures. Table 1 is a summary of the findings from these studies [6,10,14 18].
First step: differentiation of melanocytic from nonmelanocytic lesions For the Consensus Net Meeting on Dermoscopy, an algorithm modified by Stolz, Kreusch, and Menzies [13] was used to differentiate melanocytic from nonmelanocytic pigmented lesions. This is the first step in analyzing pigmented lesions of the skin. For a lesion to be identified as a melanocytic lesion, there must be a pigment network; aggregated globules; branched streaks; homogenous blue pigmentation; or a parallel pattern (the latter pattern seen on palms and soles). If one does not find any of the dermoscopic features characteristic of melanocytic lesions, then one must look for other dermoscopic features seen in nonmelanocytic lesions. For example, the diagnosis of seborrheic keratoses can be made by the presence of milia-like cysts, comedolike openings, or fissures and ridges. Basal cell carcinomas are characterized by arborizing vessels, mapleleaf like areas, and sometimes blue-gray nests or globules. The presence of red-blue lacunae or lagoons is pathognomonic for hemangiomas.
Second step: differentiation of benign from malignant melanocytic lesions Once it has been determined in step 1 that a pigmented lesion is probably melanocytic, then step
C.M. Grin et al. / Dermatol Clin 20 (2002) 641646 Table 1 Working terminology for dermoscopic structures Dermoscopic feature Pigment network Histopathologic correlate
643
Network lines Network holes Regular network Network irregularity; Broadened pigment network Absence of network Black dots Brown globules
Blotches Dark pigment obscuring network pattern Streaks Pseudopods Radial streaming Hypopigmented areas White scarlike areas Blue area Grey-blue area Whitish (blue whitish) veil
Elongated and pigmented rete ridges single melanocytic hyperplasia at the dermoepidermal junction small nests of melanocytes at tips of rete ridges Pigmented rete ridges Dermal papillae; on face can correspond to follicles Regular pattern of elongated hyperpigmented rete ridges evenly dispersed single and nests of melanocytes at tips of rete ridges Hyperpigmented and broadened irregular rete ridges with irregularly dispersed increased number of single and aggregated melanocytes at the dermoepidermal junction Effacement of rete ridge pattern Small aggregates of melanocytes or melanin in the stratum corneum Nests of melanocytes containing melanin in the superficial dermis or at the dermoepidermal junction; clusters of melanophages in the papillary dermis Abundant amounts of melanin diffusely throughout the epidermis (all levels) or the upper dermis Confluent pigmented junctional nests of melanocytes at the periphery of the lesion Radially arranged pigmented junctional nests of melanocytes at the periphery of the lesion Area of epidermis with less melanin pigmentation Epidermis and dermis focally devoid of melanin; associated with fibrosis in papillary dermis Melanin or hemosiderin in mid and deep dermis within melanocytes, melanophages or extracellularly As above plus fibrosis in papillary dermis Epidermal hyperplasia with compact orthokeratosis with or without hypergranulosis overlying large amounts of dermal melanin within melanocytes, melanophages, or extracellularly Vascular dilation Vascular dilation Histopathologic correlate Dilated blood vessels in papillary dermis; may be surrounded by hyperplastic epidermis (angioma, angiokeratoma) Hemorrhage in stratum corneum Aggregates of heavily melanized basaloid keratinocytes within the dermis (pigmented basal cell carcinoma) Intraepidermal horn globules or pseudocysts underneath the surface Follicular keratin plugs (open to the surface) Fibrosis within dermis attached to overlyling epidermis surrounded by hyperpigmented elongated rete ridges (dermatofibroma)
Erythema Telangiectasia Dermoscopic nonmelanocytic features Red-blue-purple lacunae Homogeneous dark red-black area without pigment network Maple leaf like areas Milia-like cysts Comedolike openings Central white scarlike patch with pigment network at periphery
2 is used to determine whether the lesion is benign or malignant. There are several methods that can be used for this second step. The first method used in the differential diagnosis of melanocytic lesions is called pattern analysis. It was described by Pehamberger et al [5] in 1987. This method requires a detailed, qualitative assessment of multiple dermoscopic criteria. Although this technique has been reported to be
the most sensitive and specific, with the highest diagnostic accuracy [8] for differentiating malignant melanoma from benign melanocytic lesions, it requires a significant degree of formal training. Binder et al [19] reported that with pattern analysis, the improvement in diagnostic accuracy was only increased with dermoscopy in expert dermatologists. In fact, diagnostic accuracy was lower with the use of der-
644
moscopy in nonexperts. Several groups have proposed other algorithms, which are simpler to learn and use [9,11,12,20,21].
Diagnostic algorithms ABCD rule of Stolz In 1994, Stolz et al [1,12] introduced the ABCD rule of dermatoscopy. In a study of 194 pigmented skin lesions, they reported a sensitivity of 92% and specificity of 91% and concluded that this tool was useful in the differential diagnosis of benign and malignant pigmented lesions. With this relatively straightforward algorithm, a semiquantitative score is derived. Depending on the score, the pigmented lesion is considered most likely benign, malignant, or suspicious for melanoma. The ABCD rule of dermatoscopy is relatively easy to learn and has been shown to be reproducible. It uses the following criteria: A stands for asymmetry, B for border, C for color, and D for differential structure. To determine asymmetry, the lesion is bisected by two perpendicular 90-degree axes to produce the lowest asymmetry score. Asymmetry is calculated based on contour and the distribution of colors and structure. If the lesion is asymmetric in both axes, then the score is 2. If the lesion is not asymmetric in either axis, then the score is 0. To calculate the border score, the lesion is divided into eighths. Each section in which there is an abrupt, sharp cutoff of the pigment pattern at the periphery is given a score of 1. If there is no abrupt cutoff at the periphery in any of the eight sections, then the border score is 0. If all sections show an abrupt cutoff, then the score is 8. With regard to color, the following six colors are counted: (1) light brown, (2) dark brown, (3) black, (4) red, (5) slate-blue, and (6) white. If all colors are present, then the score is 6. Differential structures include the following: pigment network; globules; dots ( < 0.1 mm); branched streaks; and structureless areas ( > 10% of total lesion). If all the structures are present, the total score for differential structures is 5. Once individual scores for A, B, C, and D are obtained, the total dermatoscopy score can be calculated according to the following formula: Asymmetry score 1.3 Border score 0.1 Color score 0.5 Differential structures 0.5
These numbers are then added together. If the total dermatoscopy score is less than or equal to 4.75, the lesion is probably benign, whereas if the score is greater than 5.45, the lesion is most likely malignant. Pigmented lesions with a score between 4.75 and 5.45 are suspicious for melanoma. False-positive scores may be seen in nevi with a globular pattern; nevi with papillary, lentiginous, or congenital components; or Spitz nevi [1]. Seven features for melanoma Dal Pozzo et al [9] developed an algorithm based on a few dermoscopic criteria that were predictive for melanoma. The seven features included (1) regression-erythema, (2) radial streaming, (3) gray-blue veil, (4) irregularly distributed pseudopods, (5) lack of homogeneity, (6) irregular pigment network, and (7) sharp margin at the periphery. The first four features carry a score of 2, whereas the latter three criteria have a score of 1. A total score of 2 or more is associated with malignancy, whereas a score less than 2 is considered benign. Menzies method: surface microscopic features for invasive melanoma Menzies et al [11,22], in 1996, developed another method for the analysis of invasive melanomas, which was geared to the inexperienced clinician. From a study of invasive melanomas and nonmelanoma lesions, 11 features were identified for their high specificity (positive features) and for their low sensitivity (negative features). The negative features, unlikely to be seen in melanoma, included symmetry of pigmentation pattern (does not require symmetry of shape) and presence of a single color. The nine positive features for melanoma include (1) blue-white veil; (2) multiple brown dots; (3) pseudopods; (4) radial streaming; (5) scarlike depigmentation; (6) peripheral black dots or globules; (7) multiple colors ( ! 5); (8) multiple blue-gray dots; and (9) broadened network. According to Menzies, the blue-white veil is an irregular confluent blue pigmentation with an overlying ground-glass white film. Radial streaming are finger-like projections at the periphery of the lesion, whereas pseudopods are bulbous projections at the periphery. The colors that are used in the Menzies method are tan, dark brown, black, red, blue, and gray. White is not scored. For a pigmented lesion to be diagnosed as a melanoma, neither of the two negative features can be present and there must be at least one of the positive features present.
645
Seven-point checklist In 1997, Argenziano et al [8,20] simplified pattern analysis. They developed the ELM seven-point checklist based on those features that were frequently seen in malignant melanoma. This algorithm has fewer criteria than classic pattern analysis and provides a quantitative scoring system for this method of evaluation. It has been shown to be useful for nonexpert dermatologists. The seven-point checklist is divided into major and minor criteria. The major criteria are an atypical pigment network; gray-blue areas (blue whitish veil); and atypical vascular pattern. The minor criteria include irregular streaks (radial streaming and pseudopods); irregular pigmentation (blotches); irregular dot or globules; regression structures (white scarlike areas); and blue areas (blue-gray areas and multiple gray-blue dots). The major criteria, if present, each receive a score of 2. The minor criteria each receive a score of 1. If the total score is greater than or equal to 3, a diagnosis of melanoma is made.
Dedication In 1989, as a melanoma fellow at New York University Medical Center, the first author was first introduced to epiluminescence microscopy. Dr Al Kopf, the authors mentor, explained that this was a technique frequently used in pigmented lesion clinics in Europe to aid in the differential diagnosis of melanoma. He was interested in learning and studying this technique. And so it was that we spent the following year looking at hundreds of dermoscopic images kindly lent to us from colleagues abroad. Eleven years later, in February 2001, we both took part in the Consensus Conference on Dermoscopy held in Rome. Although we have come a long way in understanding this technique better, there is still much work to be done. In the future, we look toward the possibility of machine vision, the automated diagnosis of malignant melanoma. This article is a summary of the technique, terminology, histopathologic correlates, and diagnostic algorithms. It is dedicated to Dr Al Kopf, to whom the author is eternally grateful for his kindness, wisdom, and knowledge, and his continued support.
Summary Many studies have evaluated the sensitivity, specificity, and diagnostic accuracy of dermoscopy in the diagnosis of pigmented skin lesions. It was not until just very recently, however, that an evidence-based approach to the usefulness of this technique was employed. Bafounta et al [23] performed a metaanalysis to evaluate dermoscopy as a diagnostic test. In this meta-analysis, based on eight published studies, the accuracy of dermoscopic diagnosis was compared with the clinical diagnosis. The authors concluded that for experienced users, dermoscopy had higher discriminatory power than the naked-eye examination in detecting melanoma. The question that still remains is the importance of this technique in the general dermatology setting. Soyer [24] is currently developing a clinical trial by the Internet for all interested clinicians, with varying expertise, to attempt to answer this question. Although research continues to advance in the field of dermoscopy, several new imaging techniques are being explored. Confocal microscopy allows for high-resolution, in vivo imaging of cellular detail, and can provide additional information in the evaluation of melanocytic lesions [25]. Active research is being done with digital dermoscopy and computerassisted diagnosis [26]. These techniques will likely contribute to the ability to diagnose malignant melanoma in the future. References
[1] Stolz W, Braun-Falco O, Bilek P, et al, editors. Color atlas of dermatoscopy. Oxford, England: Blackwell Science Inc.; 1994. p. 3. [2] Goldman L. Some investigative studies of pigmented nevi with cutaneous microscopy. J Invest Dermatol 1951;16:407 27. [3] Mackie R. An aid to the preoperative assessment of pigmented lesions of the skin. Br J Dermatol 1971;85: 232 8. [4] Fritsch P, Pechlaner R. The pigment network: a new tool for the diagnosis of pigmented lesions. J Invest Dermatol 1980;74:458 9. [5] Pehamberger H, Steiner A, Wolff K. In vivo epiluminescence microscopy of pigmented skin lesions. I. Pattern analysis of pigmented skin lesions. J Am Acad Dermatol 1987;17:571 83. [6] Soyer HP, Smolle J, Hodl S, et al. Surface microscopy: a new approach to the diagnosis of cutaneous pigmented tumors. Am J Dermatopathol 1989;11:1 10. [7] Bahmer F, Fritch P, Kreusch J, et al. Terminology in surface microscopy. Consensus Meeting of the Committee on Analytical Morphology, Hamburg, Germany, November 17, 1989. J Am Acad Dermatol 1990;23: 1159 62. [8] Argenziano G, Fabbrocini G, Carli P, et al. Epiluminescence microscopy for the diagnosis of doubtful melanocytic skin lesions. Arch Dermatol 1998;134: 1563 70.
646
C.M. Grin et al. / Dermatol Clin 20 (2002) 641646 correlates of dermoscopic criteria. Dermatol Clin 2001;19:259 68. Yadav S, Vossaert KA, Kopf AW, et al. Histopathologic correlates of structures seen on dermoscopy (epiluminescence microscopy). Am J Dermatopathol 1993; 15:297 305. Binder M, Schwarz M, Winkler A, et al. Epiluminescence microscopy: a useful tool for the diagnosis of pigmented skin lesions for formally trained dermatologists. Arch Dermatol 1995;131:286 91. Argenziano G, Soyer HP, De Giorgi V, et al. Dermoscopy: a tutorial. Milan: EDRA Medical Publishing and New Media; 2000. Nachbar F, Stolz W, Merkle T, et al. The ABCD rule of dermatoscopy. J Am Acad Dermatol 1994;30: 551 9. Menzies SW, Ingvar C, Crotty KA, et al. Frequency and morphologic characteristics of invasive melanomas lacking specific surface microscopic features. Arch Dermatol 1996;132:1178 82. Bafounta ML, Beauchet A, Aegerter P, et al. Is dermoscopy (epiluminescence microscopy) useful for the diagnosis of melanoma? Arch Dermatol 2001;137: 1343 50. Soyer HP. Is dermoscopy useful for the diagnosis of melanoma? [editorial]. Arch Dermatol 2001;137: 1361 3. Langley RGB, Rajadhyaksha M, Dwyer P, et al. Confocal scanning laser microscopy in pigmented skin lesions J Am Acad Dermatol. 2001;45(3):365 76. Fleming MG. Digital dermoscopy. Dermatol Clin 2001; 19:359 67.
[9] Dal Pozzo V, Benelli C, Roscetti E. The seven features for melanoma: a new diagnostic algorithm for the diagnosis of malignant melanoma. Eur J Dermatol 1999;9:303 8. [10] Kenet RO, Kang S, Kenet BJ, et al. Clinical diagnosis of pigmented lesions using digital epiluminescence microscopy. Arch Dermatol 1993;129:157 74. [11] Menzies SW, Ingvar C, McCarthy WH. A sensitivity and specificity analysis of the surface microscopy features of invasive melanoma. Melanoma Res 1996;6: 55 62. [12] Stolz W, Riemann A, Cognetta AB, et al. ABCD rule of dermatoscopy: a new practical method for early recognition of malignant melanoma. Eur J Dermatol 1994;4:521 7. [13] Soyer HP, Argenziano G, Chimenti S, et al. Dermoscopy of pigmented lesions: an atlas based on the Consensus Net Meeting on Dermoscopy 2000. Milan: EDRA Medical Publishing and New Media; 2001. [14] Argenyi ZB. Dermoscopy (epiluminescence microscopy) of pigmented skin lesions. Dermatol Clin 1997;15:79 95. [15] Ferrari A, Soyer HP, Peris K, et al. Central white scar like patch: a dermoscopic clue for the diagnosis of dermatofibroma. J Am Acad Dermatol 2000;43: 1123 5. [16] Massi D, De Giorgi V, Carli P, et al. Diagnostic significance of the blue hue in dermoscopy of melanocytic lesions. Am J Dermatopathol 2001;23: 463 9. [17] Massi D, De Giorgi V, Soyer HP. Histopathologic
[18]
[19]
[20]
[21] [22]
[23]
[24]
[25]
[26]
Dermatol Clin 20 (2002) 647 658
Prognostic factors
Gary S. Rogers, MD*, Stephanie M. Braun, MD
Departments of Dermatology and Surgery, Tufts University School of Medicine, 750 Washington Street, Boston, MA 02111, USA
For more than 50 years, prognostic factors for melanoma have been evaluated [1]. These have included clinical information, such as age, gender, and anatomic site of the primary melanoma, and myriad histologic factors, such as thickness, level of invasion, ulceration, and regression, to name a few. Over the past several years, the American Joint Committee on Cancer (AJCC) has worked to revise the melanoma staging system to incorporate those factors that seem to give the most significant prognostic information. To establish the revised staging system, the committee used a database of 30,450 melanoma patients including 17,600 patients on whom complete information about relevant prognostic factors was available [2,3]. The final version of the new staging system was adopted in early 2002. This article first discusses prognostic factors for melanoma in order of significance as outlined by the new AJCC staging system including Breslow thickness, ulceration, and nodal status. Thereafter, other important prognostic factors, such as anatomic site of primary melanoma, age, gender, inflammatory response, regression, and mitotic rate are reviewed. The article concludes with a brief review of new technology, which may aid in the future staging evaluation of melanoma patients.
New TNM classification and staging systems for melanoma The TNM classification and stage groupings of the new system are summarized in Tables 1 and 2 respectively. The former staging system is listed in Table 3 for comparison. Table 4 is a quick guide to the revised staging system that one can use as a pocket reference. One main structural difference is that the new system includes pathologic staging, in addition to clinical staging, which becomes particularly important in patients with nodal disease. This allows for incorporation of new diagnostic tools, such as sentinel lymph node biopsy, in predicting prognosis. T classification Breslow thickness The Breslow thickness rather than the Clark level of invasion is the most important predictor of outcome in localized (stage I and II) disease [2 30]. Thickness is measured in millimeters from the top of the granular layer of the epidermis to the deepest point of invasion of tumor cells in the dermis [26]. The one instance in which Clark level becomes important is in thin melanomas ( 1 mm), where deeper level of invasion (IV or V) correlates with a worse prognosis [31 39]. Clark level is not included in the new staging system except in thin melanomas. For simplicity, the thickness measurements have been changed to whole integers ( 1, 1.01 to 2, 2.01 to 4, >4 mm) versus the somewhat arbitrary and difficult to remember divisions in the old system ( 0.75, 0.76 to 1.5, 1.6 to 3, 3.01 to 4, > 4 mm). The former divisions were based on Breslows origi-
* Department of Medical and Surgical Dermatology, New England Medical Center, 750 Washington Street, Box 114, Boston, MA 02111 E-mail address: grogrers@lifespan.org (G.S. Rogers).
0733-8635/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 0 7 3 3 - 8 6 3 5 ( 0 2 ) 0 0 0 2 7 - X
648
G.S. Rogers, S.M. Braun / Dermatol Clin 20 (2002) 647658
Table 1 Revised AJCC TNM Classification for Melanoma (2002) Breslow Thickness Tis T1 In situ 1 mm Ulceration (+ or ) a. and Clark Level II or III b. + or Clark Level IV or V a. b. + a. b. + a. b. +
T2 T3 T4
1.01 2 mm 2.01 4 mm > 4 mm
excoriations [25,57]. Ulceration of a melanoma upstages the prognosis of that patient when compared with a patient with a nonulcerated melanoma of equivalent thickness [2]. In fact, the presence of ulceration is so important that patients with thick, ulcerated melanomas (>4 mm) lacking nodal metastases have a worse prognosis than some subgroups of patients with nodal disease [2]. The mechanism of ulceration has not yet been established. Some believe it to be related to a high mitotic rate, causing the tumor to outgrow its blood supply [58]. N classification Number of metastatic lymph nodes The most significant prognostic factor in the new N classification is the number of metastatic lymph nodes rather than the size of the metastasis within the node. Size, which was the basis for the old N classification, does not seem to be an independent prognostic indicator [21,59]. The number of nodes is divided into one, two to three, or greater than four metastatic nodes, each category with progressively worsening prognosis [3,12 15,19 21,59 64]. Tumor burden (micrometastases versus macrometastases) The classification also takes into account tumor burden, which is divided into micrometastasis and macrometastasis. Micrometastases are clinically occult metastases diagnosed histologically after a sentinel lymph node biopsy or elective lymphadenectomy. Macrometastases are clinically apparent nodes that are proved to be metastases histologically after lymphadenectomy or when there is gross extracapsular extension of the nodal metastasis. Patients with macrometastases have a lower survival rate than those with micrometastases [2,3,46,64,65]. New procedures, such as reverse transcriptase polymerase chain reaction (RT-PCR) and immunostains of lymph nodes, are being evaluated to improve the detection of micrometastasis [2,66 68]. The identification of involved nodes is particularly important because finding a positive node may influence treatment options, such as lymph node dissection or adjuvant therapy. Satellite lesions and in-transit metastases A microscopic satellite is defined as a tumor nest greater than or equal to 0.05 mm in size within the same histologic section as the primary tumor but separated from the primary lesion by normal tissue. An in-transit metastasis is a lesion more than 3 cm from the primary tumor but has not reached the draining regional lymph nodes [58]. In the old AJCC
Number of Metastatic Regional Nodes N0 N1 N2 0 1 23
Tumor Burden a. Micrometastasis G b. Macrometastasis x a. Micrometastasis G b. Macrometastasis x c. In-transit metastases or satellites without metastatic nodes
N3
!4 or matted nodes or in-transit metastases or satellites with nodes Serum Lactate Dehydrogenase Normal
Site of Metastasis(es) M0 M1a None Distant skin, subcutis or nonregional lymph nodes Lung All other viscera Any distant metastasis
M1b M1c
Normal Normal Elevated
nal papers discussing the prognostic importance of thickness in melanoma [26,40]. Later work has shown that these are not naturally occurring breakpoints; the new divisions were chosen for ease [3,4,21]. Ulceration Ulceration of the primary melanoma is the second strongest independent predictor of outcome in localized disease and is also significant in regional (ie, nodal) disease [3,27 29,35,41 56]. It has been incorporated into the T classification. Ulceration is defined histologically as the lack of intact epidermis over the lesion [28,29,47,51]. It can be distinguished readily from traumatic disruptions of the epidermis, such as
G.S. Rogers, S.M. Braun / Dermatol Clin 20 (2002) 647658 Table 2 Revised AJCC Clinical and Pathologic Staging of Melanoma (2002) Stage 0 IA B II A B C III III A B Clinical Tis T1a T1b or T2a T2b or T3a T3b or T4a T4b Any T N0 N0 N0 N0 N0 N0 N1 3 M0 M0 M0 M0 M0 M0 M0 Pathologic same as clinical T1 4a T1 4a T1 4b T1 4b T1 4a T1 4a T1 4a/b T1 4b T1 4b Any T Any T N1a N2a N1a N2a N1b N2b N2c N1b N2b N3 Any N M0 M0 M0 M0 M0 M0 M0 M0 M0 M0 Any M1
649
% Five-Year Survival Rates* Nearly 100 95.3 89 90.9 77.4 78.7 63 67.4 45.1 69.5 63.3 52.8 49.6 59 46.3 29 24 26.7 6.7 18.8
IV
Any T
Any N
M1a
* Survival rates are based on data from Balch CM et al. Final version of the American Joint Committee on Cancer staging system for cutaneous melanoma. J Clin Oncol 2001; 19:3635 48.
staging system, satellites were included in the T classification. Satellite lesions are now considered equivalent to in-transit metastases, portend a prognosis equivalent to node-positive disease, and are included in the N classification of the new staging system [2 4,12 20,69]. M classification Site of distant metastases Three distinct subsets (M1a, M1b, and M1c) have been created in the new M classification based on small, but statistically significant, differences in survival between the groups. The most significant prognostic factor in the new M classification is the site of distant metastasis. Metastases to nonvisceral sites including skin, subcutis, and lymph nodes (categorized as M1a) have a better prognosis than those to viscera (M1c) when evaluating 1- and 2-year survival [3,70]. There is some evidence that patients with metastases to the lung may have a better prognosis at 1 year than those with metastases to other viscera and are given the distinct category M1b [3,70 72]. This survival advantage, however, does not hold true at 2 years [3,70]. Lactate dehydrogenase Also taken into account in the new M classification is serum lactate dehydrogenase, which can be elevated in metastatic disease [2,3,73 77]. Elevated
lactate dehydrogenase places the patient in the M1c subgroup. Other causes of elevated lactate dehydrogenase unrelated to melanoma, such as hemolysis or lymphoma, need to be eliminated first. Stage groupings Stages I and II (localized disease) Stages I and II are characterized by a melanoma of any thickness with or without ulceration, and absence of nodal, satellite, or distant metastases. Stage I patients are at low risk, whereas stage II patients are at intermediate risk for metastases [2]. With both stages, the presence of ulceration of the primary melanoma upstages the patient [2]. For example, a melanoma of thickness 1.01 to 2 mm without ulceration is stage IB, whereas a melanoma of the same thickness with ulceration is stage IIA. The corresponding 5-year survival rates are 89% and 74%, respectively [2]. As previously mentioned, patients with thin melanomas ( 1 mm) with level IV or V depth of invasion are also upstaged to stage IB. The 5-year survival rate of a less than or equal to 1-mm nonulcerated melanoma with Clark level II-III (ie, IA) is 95% versus 91% for the equivalent melanoma with Clark level IV-V [2]. Stage III (nodal and regional disease) This stage is characterized by nodal disease or intransit and satellite metastases, a melanoma of any
650
Table 3 Former AJCC Staging System (1992) TNM Staging Breslow Thickness Tis T1 T2 T3a T3b T4a T4b Melanoma in-situ 0.75 mm 0.76 1.5 mm 1.51 3 mm 3.01 4 mm > 4 mm or Clark Level I II III IV IV V Satellite(s) within 2 cm of primary tumor Size of Regional Nodal Metastasis N0 N1 N2a N2b N2c no lymph node involvement 3 cm > 3 cm In-transit metastasis In-transit metastasis Other
Other
tion, patients with metastases in clinically detectable nodes (ie, macrometastases) are upstaged in comparison with patients with clinically occult nodes (ie, micrometastases). The 5-year survival for one microscopically involved node is 61% versus 46% for one macroscopically involved node [3]. As in stages I and II, the presence of ulceration upstages the patient independent of the number of metastatic nodes or the tumor burden [3]. In addition, the finding of an in-transit and satellite metastasis without metastatic nodes places the patient in stage IIIB, which has a 5-year survival in the range of 46% to 59%. If metastatic nodes are found in combination with an in-transit and satellite metastasis, then the patient is stage IIIC, which has a 5-year rate of 26% [2]. Stage IV (distant metastases) This stage is characterized by the presence of distant metastases. Although the prognosis is dismal overall for all stage IV patients with median survival about 6 months [27], those with metastases to nonvisceral sites seem to do better than those with metastases to viscera. One-year survival rates are 59% and 41%, respectively, and 5-year survival rates are 19% and 10%, respectively. Patients with lung metastases seem to have a prognosis intermediate to the two other categories at the 1-year mark (57% survival) but by 2 years their prognosis is similar to those with metastases to other viscera (23.1% versus 23.6% survival, respectively) [2]. It also seems that the number of metastases may influence survival [80 82]. There have been case reports of significantly increased survival after resection of a solitary metastasis to the lung, brain, or bowel [83]. Other factors The prognostic factors discussed in the preceding sections have been shown most significantly to impact survival in melanoma. There are other historically evaluated factors, however, which may influence survival to a lesser extent. There are factors, such as molecular markers, which are in the early stages of investigation and may also affect prognosis. Both of these groups of factors are discussed next. Level of dermal invasion The Clark level of dermal invasion seems to be less significant than the Breslow thickness in predicting prognosis. Clark defined the following five anatomic levels: I, no dermal invasion; II, tumor cells in the
> 3 cm
Site of Distant Metastasis M0 M1a None Skin, subcutis, or nonregional lymph nodes Viscera
M1b
Stage Grouping IA IB IIA IIB III IV T1 T2 T3 T4 Any T Any T N0 N0 N0 N0 N1 or N2 Any N M0 M0 M0 M0 M0 M1
thickness with or without ulceration, and the absence of distant metastases. It is a heterogeneous group with significant differences in survival among the three subgroups of this stage. The 5-year survival of IIIA, IIIB, and IIIC are 67%, 53%, and 26%, respectively [3,12,14,15,19,20,36,59 62,78,79]. More specifically, prognosis worsens with increasing number of nodes regardless of whether there is microscopic or macroscopic involvement. For example, 5-year survival is 61% with one microscopically involved node versus 35% with greater than or equal to four microscopically involved nodes [3]. In addi-
G.S. Rogers, S.M. Braun / Dermatol Clin 20 (2002) 647658 Table 4 Summary of staging criteria Stage IA IB IIA IIB IIC IIIA IIIB Thickness 1 mm 1 mm or 1.01 2 mm 1.01 2 mm or 2.01 4 mm 2.01 4 mm or > 4 mm > 4 mm any any Ulceration + + + + + + + + +/ +/ +/ Regional Nodal Involvement 1 or 2 3 microscopic 1 or 2 3 microscopic 1 or 2 3 microscopic 1 macroscopic 2 3 macroscopic 4 or matted any nodes any In-transit or Satellite Metastases + + +/
651
Distant Metastasis +
IIIC
any
IV
any
papillary dermis; III, tumor cells filling the papillary dermis; IV, tumor cells in the reticular dermis; and V, invasion into the subcutis [6,84]. In his model, prognosis worsened with deeper depth of invasion. Most later studies, however, have not shown level of invasion to be an independent prognostic factor except in thin melanomas ( 1 mm), where those with level IV or V have decreased survival [3,31 39]. Also, from a technical perspective, there is more intraobserver variability associated with determining Clark level versus Breslow thickness [85]. Anatomic site The anatomic site of the primary melanoma has been shown repeatedly to be an independent prognostic factor [27,56,58,86 97]. Patients with lesions on the extremities, except hands and feet, have a better prognosis than those with lesions on the trunk, head, and neck [3,27,31,64,98,158]. More specifically, melanomas located on the scalp, hands, and feet are associated with a particularly poor outcome [10,48,99,114,159]. Gender There are conflicting data on whether women have a better prognosis than men in melanoma. Some studies have not shown gender to be an independent variable but a reflection of differences in melanoma presentation, such as site, thickness, and ulceration, between women and men [58,86,88 91,96,97,100].
Women often present with thin, nonulcerated melanomas on the extremities, whereas men are more likely to have thick, ulcerated melanomas on the trunk [58]. Other studies, however, have shown female gender to be independent of site, thickness, and ulceration, and that men have a poorer prognosis than women [27,31,56,87,92,94,95,101,158]. Several authors have noted an association between gender and age in terms of outcome in melanoma. For example, one study found that women younger than 50 years old have a better survival rate than men and women older than 50 years [102]. Another study showed that women of any age and men less than 50 years old do better in terms of prognosis than men greater than 50 years old [103]. Other authors have found no differences in survival when comparing women of various ages [27]. Age The age of the patient is an independent prognostic factor in melanoma [3,85,102,104]. Increasing age is associated with decreasing survival rates regardless of the thickness or ulceration status of the lesion [3,41,46,49,64,105 107]. More specifically, patients over 60 tend to have a worse prognosis [41,46,49,55,105,107 110]. There also may be a link between age and gender, because some have observed lower survival rates in older men with melanoma [103,104]. In addition, older age is associated with progression from regional to visceral disease, whereas younger age is associated with progression
652
from local to regional disease [111]. A decline in immune defense mechanisms seen with increasing age may account for the lower survival rates seen in the elderly [3]. Mitotic rate High mitotic rate, as measured by the number of mitoses per millimeter squared in the dermal component of the melanoma, is associated with a poor prognosis [1,56,87,92,93,112 114]. Clark et al [87] found an 8-year survival of 95% in patients with tumors with no mitoses, a 79.4% survival rate in those with tumors with a mitotic rate less than 6/mm2, and a 38.2% survival rate in those with a mitotic rate of less than 6/mm2. Others claim that the prognostic index, which is the product of the mitotic rate and the thickness in millimeters, is more important than either two of the variables independently [100]. For melanomas 1.5 to 3.99 mm in thickness, Kopf et al [115] observed decreased survival in those with a prognostic index greater than 19 versus those with a prognostic index less than 19. Inflammatory response The difficulty in evaluating host lymphocyte response to melanoma as a prognostic indicator is the lack of a uniform definition from study to study. Some authors refer to lymphocytes at the base of the tumor [27], others refer to lymphocytes around and within the tumor, whereas still others refer to only those lymphocytes infiltrating the tumor [58]. Clark et al [87] used the term tumor-infiltrating lymphocytes to define only those lymphocytes within the tumor as the host response. In his study, a brisk host response was associated with a favorable prognosis in localized disease. This work has been confirmed by others [116]. Another author found the lack of an inflammatory response to be associated with a less favorable prognosis in thicker lesions (>3.65 mm) [117]. Despite many studies that have shown that a brisk lymphocyte response favorably affects prognosis [27,86,87,93,112,113,118,119], other studies have not shown this to be true [42,47,88,92,96,115, 120 123]. There is some suggestion that the presence of other inflammatory components, such as plasma cells or macrophages, may adversely affect prognosis [124,125], but this claim is yet to be substantiated. Regression As discussed previously for inflammatory response, the lack of a standard definition of regression
has also made this prognostic factor somewhat difficult to evaluate. The aspects that currently define regression are the absence of tumor cells and the presence of papillary dermal fibrosis, vascular proliferation, and scattered lymphocytes and melanophages [58]. Regression has been studied particularly in thin melanomas to try to predict which lesions will metastasize. Several authors have not found regression to have prognostic significance [56,122,123,126 130]. Other studies, however, have shown regression to be a poor prognostic indicator in thin melanomas (<0.76 mm historically) [31,87,131 135]. For example, in a series of 103 patients with thin melanomas, those with greater than 77% regression had an increased incidence of visceral metastases, whereas no metastases developed in patients with equivalent lesions without regression [132]. In a series of 121 patients with thin melanomas, 19% had regression; yet, this subgroup accounted for 71.4% of those who developed metastases [131]. Melanoma growth patterns Several authors have attempted to stratify survival rates according to the four histogenetic subtypes of melanoma: (1) lentigo maligna, (2) superficial spreading, (3) nodular, and (4) acral-lentiginous [27]. It has become clear that although historically important, this variable does not seem to be an independent prognostic factor but is rather a reflection of inherent thickness differences between the subtypes [114,136]. For example, nodular melanoma by definition is thicker than superficially spreading melanoma [137]. Lentigo maligna melanoma, often seen on the face of elderly individuals, had long been thought to have a more benign course. Thickness for thickness, however, survival for lentigo maligna melanoma is not superior [136]. Cell type The predominance of spindled over epithelioid cells has been associated with a better prognosis in several reports [93,138,139]. Other investigators have not substantiated this finding [87]. Vascular invasion Whether vascular invasion is an independent indicator of prognosis for melanoma, as it is for other malignant tumors, is yet to be determined. Several investigators have found it to predict poor outcome [96,100,140], whereas others have not [87,93]. Vascular invasion is associated with melanomas with
653
deeper thicknesses and levels of invasion [140]. Also, because it is closely related to microscopic satellites, it is a significant predictor of nodal involvement [141].
Melastanin mRNA expression Melastanin is a calcium channel protein produced only by melanocytes in the skin and the choroid of the eye [149]. Different patterns of melastatin expression are observed in melanomas versus benign melanocytic nevi [150]. The loss of melastanin mRNA expression in melanomas is associated with the risk of metastases [151]. It can be used to stratify patients with localized disease (stage I and II) into low- and high-risk for developing metastases. Duncan et al [152] demonstrated that melastanin status correlates with disease-free survival independent of tumor thickness and mitotic rate. In a series of 150 patients with primary melanoma, there was an 8-year diseasefree survival of 100% in stage I patients with diffuse expression of melastatin mRNA versus 77% in those with melastatin loss. Similarly, 8-year disease-free survival in patients with stage II disease was 90% in those with diffuse melastatin mRNA expression versus 51% in those with melastatin loss. Proliferation markers Proliferation markers, such as Ki-67, and proliferating cell nuclear antigen are being evaluated for potential prognostic significance in melanoma. Their usefulness, however, is yet to be determined [153 157].
New molecular and biochemical prognostic markers Several new techniques are under investigation to help those caring for melanoma patients try to predict which clinically disease-free patients may be at risk for developing recurrences. RT-PCR for melanocyte-specific proteins The RT-PCR is a means by which submicroscopic quantities of melanoma cells can be detected in lymph nodes, blood, and other tissues by targeting markers specific to melanocytes [68,142 145]. Researchers have used melanin synthesis enzymes, most commonly tyrosinase and MART-1, as markers because they are produced by melanocytes but few other cell types. Other markers that have been evaluated include tyrosine-related proteins 1 and 2, MAGE-3, and gp100 [146]. In regards to lymph nodes, this procedure is being tested as an adjuvant to routine hematoxylin and eosin histologic evaluation of lymph node specimens. Because the presence of nodal disease is such a strong predictor of ultimate outcome, more accurate, early detection is expected to change staging, treatment options, and survival. Shivers et al [68] evaluated the histologically negative lymph nodes from sentinel lymph node biopsies from stage I and II patients with RT-PCR for tyrosinase. He demonstrated that 13% of patients with histologically negative but RT-PCR tyrosinase positive sentinel lymph nodes recurred. Only one patient recurred who was negative on both evaluations. Other investigators have shown that testing for multiple markers by RT-PCR on the same specimen further increases the detection of submicroscopic melanoma within the lymph node [144]. Several studies have focused on using RT-PCR technology to detect melanoma cells circulating in the peripheral blood. Curry et al [147] found 60.4% positivity for MART-1 and 14.6% positivity for tyrosinase in 48 melanoma patients with disseminated disease. They conclude that these markers are of prognostic significance but caution their use in making therapeutic decisions because of the high falsenegative rate. Other researchers have demonstrated that the number of markers in blood by RT-PCR is an independent predictor of disease recurrence, and it correlates with stage of disease [148].
References
[1] Allen AC, Spitz S. Malignant melanoma: a clinicopathological analysis of the criteria for diagnosis and prognosis. Cancer 1953;6:1 45. [2] Balch CM, Buzaid AC, Soong S-J, et al. Final version of the American Joint Committee on Cancer staging system for cutaneous melanoma. J Clin Oncol 2001; 19:3635 48. [3] Balch CM, Soong S-J, Gershenwald JE, et al. Prognostic factors analysis of 17,600 melanoma patients: validation of the American Joint Committee on Cancer melanoma staging system. J Clin Oncol 2001;19: 3622 34. [4] Balch CM, Buzaid AC, Atkins MB, et al. A new American Joint Committee on Cancer staging system for cutaneous melanoma. Cancer 2000;88:1484 91. [5] Ruiter DJ, Testori A, Eggermont AM, et al. The AJCC staging proposal for cutaneous melanoma: comments by the EORTC Melanoma Group. Ann Oncol 2001;12:9 11. [6] Clark Jr. WH, From L, Bernardino EA, et al. The histogenesis and biological behavior of primary human malignant melanoma of the skin. Cancer Res 1969; 29:705 27.
654
G.S. Rogers, S.M. Braun / Dermatol Clin 20 (2002) 647658 thickness and level of invasion in cutaneous melanoma. Hum Pathol 1977;8:1 2. Breslow A. Tumor thickness in evaluating prognosis of cutaneous melanoma [letter]. Ann Surg 1978; 187:440. Prade M, Sancho-Garnier H, Cesarini JP, et al. Difficulties encountered in the application of Clark classification and the Breslow thickness measurement in cutaneous malignant melanoma. Int J Cancer 1980; 26:159 63. Lock-Anderson J, Hou-Jensen K, Hansen JP, et al. Observer variation in histological classification of cutaneous malignant melanoma. Scand J Plast Reconstr Hand Surg 1995;29:141 8. Breslon A. cross-sectional area and depth of invasion in the prognosis of cutaneous melanoma. Ann Surg 1970;172:902 8. Balch CM, Song S-J, Shaw HM, Urist MM, McCarthy WH. An analysis of prognostic factors in 8500 patients with cutaneous melanoma. In: Balch CM, Houghton AN, Milton GW, Sober AJ, Soong SJ, editors. Cutaneous melanoma. 2nd edition. Philadelphia: Lippincott; 1992. p. 165 87. Balch CM, Murad TM, Soong S-J, Ingalls AL, Halpern NB, Maddox WA. A multifactorial analysis of melanoma: prognostic histopathological features comparing Clarks and Breslows staging methods. Ann Surg 1978;188:732 42. Balch CM, Soong S-J, Murad TM, Ingalls AL, Maddox WA. A multifactorial analysis of melanoma. II. Prognostic factors in patients with stage I (localized) melanoma. Surgery 1979;86:343 51. Rogers GS, Kopf AW. Hazard-rate analysis in stage I melanoma. Arch Derm 1986;9(122):999 1002. Slingluff Jr. CL, Vollmer RT, Reintgen DS, Seigler H. Lethal thin malignant melanoma: identifying patients at risk. Ann Surg 1988;208:150 61. Kelly JW, Sagebiel RW, Clyman S, Blois MS. Thin level IV malignant melanoma: a subset in which level is the major prognostic indicator. Ann Surg 1985;202: 98 103. Buttner P, Garbe C, Bertz J, et al. Primary cutaneous melanoma: optimized cutoff points of tumor thickness and importance of Clarks level for prognostic classification. Cancer 1995;75:2499 506. Mansson-Brahme E, Carstensen J, Erhardt K, et al. Prognostic factors in thin cutaneous malignant melanoma. Cancer 1994;73:2324 32. Marghoob AA, Koenig K, Bittencourt FV, et al. Breslow thickness and Clark level in melanoma: support for including level in pathology reports and in AJCC staging. Cancer 2000;88:589 95. Morton DL, Davtyan DG, Wanek LA, et al. Multivariate analysis of the relationship between survival and the microstage of primary melanoma by Clark level and Breslow thickness. Cancer 1993;71: 3737 43. Finley JW, Gibbs JF, Rodriguez LM, et al. Pathologic and clinical features influencing outcome in thin cu-
[7] Cascinelli N, Zurrida S, Galimberti V, et al. Acral lentiginous melanoma: a histological type without prognostic significance. J Dermatol Surg Oncol 1994;20:817 22. [8] McGovern VJ, Shaw HM, Milton GW, et al. Is malignant melanoma arising in a Hutchinsons melanotic freckle a separate disease entity? Histopathology 1980;4:235 42. [9] Kuchelmeister C, Schaumburg-Lever G, Garbe C. Acral cutaneous melanoma in Caucasians: clinical features, histopathology, and prognosis in 112 patients. Br J Dermatol 2000;143:275 80. [10] Urist MM, Balch CM, Soong S-J, et al. Head and neck melanoma in 534 clinical stage I patients: a prognostic factors analysis and results of surgical treatment. Ann Surg 1984;200:769 75. [11] Slingluff Jr. CL, Vollmer R, Seigler HF. Acral melanoma: a review of 185 patients with identification of prognostic variables. J Surg Oncol 1990;45:91 8. [12] Balch CM, Soong S-J, Murad TM, et al. A multifactorial analysis of melanoma: III. Prognostic factors in melanoma patients with lymph node metastases (stage II). Ann Surg 1981;193:377 88. [13] Bevilacqua RG, Coit DG, Rogatko A, et al. Axillary dissection in melanoma: prognostic variables in nodepositive patients. Ann Surg 1990;212:125 31. [14] Calabro A, Singletary SE, Balch CM. Patterns of relapse in 1001 consecutive patients with melanoma nodal metastases. Arch Surg 1989;124:1051 5. [15] Coit DG, Rogatko A, Brennan MF. Prognostic factors in patients with melanoma metastatic to axillary or inguinal lymph nodes: a multivariate analysis. Ann Surg 1991;214:627 36. [16] Coit DG. Prognostic factors in patients with melanoma metastatic to regional lymph nodes. Surg Oncol Clin North Am 1992;1:281 95. [17] Day C, Sober AJ, Lew RA, et al. Malignant melanoma patients with positive nodes and relatively good prognoses: microstaging retains prognostic significance in clinical stage I melanoma patients with metastases to regional lymph nodes. Cancer 1981;47: 955 62. [18] Drepper H, Beiss B, Hofherr B, et al. The prognosis of patients with stage III melanoma. Cancer 1993;71: 1239 46. [19] Morton DL, Wanek L, Nizze JA, et al. Improved long-term survival after lymphadenectomy of melanoma metastatic to regional lymph nodes: analysis of prognostic factors in 1134 patients from John Wayne Cancer Clinic. Ann Surg 1991;214:491 9. [20] Slingluff Jr. CL, Vollmer R, Seigler HF. Stage II malignant melanoma: presentation of a prognostic model and assessment of specific active immunotherapy in 1,273 patients. J Surg Oncol 1988;39:139 47. [21] Buzaid AC, Ross MI, Balch CM, et al. Critical analysis of the current AJCC staging system for cutaneous melanoma and proposal of a new staging system. J Clin Oncol 1997;15:1039 51. [22] Breslow A. Problems in the measurement of tumor
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30] [31]
[32]
[33]
[34]
[35]
[36]
[37]
G.S. Rogers, S.M. Braun / Dermatol Clin 20 (2002) 647658 taneous melanoma: correlation with newly proposed staging system. Ann Surg 2000;66(6):527 31. Salman SM, Rogers GS. Prognostic factors in thin cutaneous malignant melanoma. J Dermatol Surg Oncol 1990;16:413 8. Shaw HM, McCarthy WH, McCarthy SW, et al. Thin malignant melanomas and recurrence potential. Arch Surg 1987;122:1147 50. Breslow A. Tumor thickness, level of invasion and node dissection in stage I cutaneous melanoma. Ann Surg 1975;182:572 5. Averbook BJ, Russo LJ, Mansour EG. A long-term analysis of 620 patients with malignant melanoma at a major referral center. Surgery 1998;124:746 56. Shaw HM, Balch CM, Soong S-J, et al. Prognostic histopathologic factors in malignant melanoma. Pathology 1985;17:271 4. Gershenwald JE, Thompson W, Mansfield PF, et al. Multi-institutional melanoma lymphatic mapping experience: the prognostic value of sentinel lymph node status in 612 stage I or II melanoma patients. J Clin Oncol 1999;17:976 83. Cascinelli N, Marubini E, Morabito A, et al. Prognostic factors for stage I melanoma of the skin: a review. Stat Med 1985;4:265 78. Wagner JD, Gordon MS, Chuang TY, et al. Predicting sentinel and residual lymph node basin disease after sentinel lymph node biopsy for melanoma. Cancer 2000;89:453 62. Balch CM, Soong S-J, Ross MI, et al. Long-term results of a multi-institutional randomized trial comparing prognostic factors and surgical results for intermediate thickness melanomas (1.0 to 4.0mm). Ann Surg Oncol 2000;7:87 97. Balch CM, Wilkerson JA, Murad TM, et al. The prognostic significance of ulceration of cutaneous melanoma. Cancer 1980;45:3012 7. Balch CM, Soong S-J, Milton GW, et al. A comparison of prognostic factors and surgical results in 1,786 patients with localized (stage I) melanoma treated in Alabama and New South Wales, Australia. Ann Surg 1982;196:677 84. Balch CM, Soong S-J, Bartolucci AA, et al. Efficacy of an elective regional lymph node dissection of 1-to 4-mm thick melanomas for patients 60 years of age and younger. Ann Surg 1996;224:255 63. Balch CM, Urist MM, Karakousis CP, et al. Efficacy of 2-cm surgical margins for intermediate-thickness melanomas (1 to 4 mm): results of a multi-institutional randomized surgical trial. Ann Surg 1993; 218(3):262 7. McGovern VJ, Shaw HM, Milton GW, et al. Ulceration and prognosis in cutaneous malignant melanoma. Histopathology 1982;6:399 407. Heaton KM, Sussman JJ, Gershenwald JE, et al. Surgical margins and prognostic factors in patients with thick (>4mm) primary melanoma. Ann Surg Oncol 1998;5:322 8. Kim SH, Garcia C, Rodriguez J, et al. Prognosis of
655
[38]
[54]
[39]
[55]
[40]
[56]
[41]
[42]
[57]
[43]
[58]
[59]
[44]
[45]
[60]
[61]
[46]
[62]
[47]
[63]
[48]
[64]
[49]
[65]
[50]
[66]
[51]
[67]
[52]
[68]
[53]
thick cutaneous melanoma. J Am Coll Surg 1999; 188:241 7. Mraz-Gernhard S, Sagebiel RW, Kashani-Sabet M, et al. Prediction of sentinel lymph node micrometastasis by histological features in primary cutaneous melanoma. Arch Dermatol 1998;134:983 7. Masbeck A, Westerdahl J, Ingvar C, et al. Cutaneous malignant melanoma in southern Sweden 1965, 1975, and 1985: prognostic factors and histologic correlations. Cancer 1997;79:275 83. Mackie RM, Aitchison T, Sirel JM, et al. Prognostic models for subgroups of melanoma patients from the Scottish Melanoma Group database 1979 86 and their subsequent validation. Br J Cancer 1995;71: 173 6. Corona R, Mele A, Amini M, et al. Intraobserver variability on the histopathologic diagnosis of cutaneous melanoma and other pigmented skin lesions. J Clin Oncol 1996;14:1218 23. Greenstein DM, Rogers GS. Advances in the diagnosis and treatment of melanoma. Dermatol Surg Oncol 1991;21(11):919 26. Buzaid AC, Tinoco LA, Jendiroba D, et al. Prognostic value of size of lymph node metastases in patients with cutaneous melanoma. J Clin Oncol 1995;13: 2361 8. Balch CM. Cutaneous melanoma: prognosis and treatment results worldwide. Semin Surg Oncol 1992; 8:400 14. Gershenwald JE, Prieto V, Colome-Grimmer MI, et al. The prognostic significance of microscopic tumor burden in 925 melanoma patients undergoing sentinel lymph node biopsy [abstract]. Proc Am Soc Clin Oncol 2000;19:15a. Drepper H, Biess B, Hofherr B, et al. The prognosis of patients with stage III melanoma. Cancer 1993;71: 1239 46. Cascinelli N, Vaglini M, Nana M, et al. Prognosis of skin melanoma with regional node metastases (stage II). Surg Oncol 1984;25:240 7. Messaris GE, et al. Prognostic variables for patients with stage III malignant melanoma. Eur J Surg 2000; 166:233 9. Cascinelli N, Belli F, Santinami M, et al. Sentinel lymph node biopsy in cutaneous melanoma: the WHO Melanoma Program experience. Ann Surg Oncol 2000;7:469 74. van der Velde-Zimmermann D, Schipper ME, de Weger RA, et al. Sentinel lymph node biopsies in melanoma patients: a protocol for accurate, efficient, and cost-effective analysis by preselection for immunohistochemistry on the basis of Tyr-PCR. Ann Surg Oncol 2000;7:51 4. van der Velde-Zimmermann D, Roijers JF, BouwensRombouts A, et al. Molecular test for the detection of tumor cells in blood and sentinel nodes of melanoma patients. Am J Pathol 1996;149:759 64. Shivers SC, Xiangning W, Weiguo L, et al. Molecular staging of malignant melanoma: correlation with clinical outcome. JAMA 1998;280:1410 5.
656
G.S. Rogers, S.M. Braun / Dermatol Clin 20 (2002) 647658 gery. Philadelphia (PA): WB Saunders Co.; 1995; p. 1576 601. Clark WH. The histogenesis and biologic behavior of primary malignant melanomas of the skin. Cancer Res 1969;29(3):705 9. Colloby PS, West KP, Fletcher A. Observer variation in the measurement of Breslow depth and Clarks level in thin cutaneous malignant melanoma. J Pathol 1991;163:245 50. Bachaud J-M, Shubinski R, Boussin G, et al. Stage I cutaneous malignant melanoma: risk factors of locoregional recurrence after wide local excision and clinical perspectives. Eur J Surg Oncol 1992;18: 442 8. Clark WH, Elder DE, Guerry D, et al. Model for predicting survival in stage I melanoma based on tumor progression. J Natl Cancer Inst 1989;81:1893 904. Day CL, Mihm MC, Sober AJ. Prognostic factors for melanoma patients with lesions 0.76 1.69 mm in thickness: an appraisal of thin level IV lesions. Ann Surg 1982;195:30 4. Day CL, Lew RA, Mihm MC, Sober AJ. A multivariate analysis of prognostic factors for melanoma patients with lesions greater than or equal to 3.65 mm in thickness: the importance of revealing alternative Cox Models. Ann Surg 1982;195:44 9. Day CL, Mihm MC, Sober AJ. Predictors in late deaths among patients with clinical stage I melanoma who have not had bony or visceral metases within the first 5 years after diagnoses. J Am Acad Dermatopathol 1983;8:864 8. Gamel JW, George SL, Staley WE, Seigler HF. Skin melanoma: cured traction and survival time as functions of thickness, site histologic, type, age and sex. Cancer 1993;72:1219 33. Hacene K, Le Doussal VL, Brunet M, et al. Prognostic index for clinical stage I cutaneous malignant melanoma. Cancer Res 1983;43:2291 6. Songegaard K, Schou G. Therapeutic and clinicalpathological factors in the survival of 1,469 patients with primary cutaneous malignant melanoma in clinical stage I: a multivariate regression analysis. Virchows Arch 1985;408:249 58. Soong S-J. A computerized mathematical and scoring systems for predicting outcome in patients with localized melanoma. In: Balch CM, Houghton AN, Milton GW, et al, editors. Cutaneous melanoma. Philadelphia: JB Lippincott; 1992. p. 200 12. Stidham K, Johnson JL, Seigler HF. Survival superiority of females with melanoma: a multivariate analysis of 6383 patients exploring the significance of gender in prognostic outcome. Arch Surg 1994;129: 316 24. Thorn M, Ponten F, Bergstrom R, et al. Clinical and histopathologic predictors of survival in patients with malignant melanoma: a population-based study in Sweden. J Natl Cancer Inst 1994;86: 761 9. Vossaert KA, Silverman MK, Kopf AW, et al. Influ-
[69] Harrist TJ, Rigel DS, Day Jr. CL, Sober AJ, Lew RA, Rhodes AR, et al. Microscopic satellites are more highly associated with regional lymph nodes metastases than is primary melanoma thickness. Cancer 1984;53:2183 7. [70] Balch CM, et al. A multifactorial analysis of melanoma: IV. Prognostic factors in 200 melanoma patients with distant metastases (stage III). J Clin Oncol 1982; 1:126 33. [71] Manola J, Atkins M, Ibrahim J, et al. Prognostic factors in metastatic melanoma: a pooled analysis of Eastern Cooperative Oncology Group trials. J Clin Oncol 2000;18:3782 93. [72] Barth A, Wanek LA, Morton DL. Prognostic factors in 1,521 melanoma patients with distant metastases. J Am Coll Surg 1995;181:193 201. [73] Deichmann M, Benner A, Bock M, et al. S100-beta, melanoma-inhibiting activity, and lactate dehydrogenase discriminate progressive from nonprogressive American Joint Committee on Cancer stage IV melanoma. J Clin Oncol 1999;17:1891 6. [74] Eton O, Legha SS, Moon TE, et al. Prognostic factors for survival of patients treated systemically for disseminated melanoma. J Clin Oncol 1998;16:1103 11. [75] Franzke A, Probst-Kepper M, Buer J, et al. Elevated pretreatment serum levels of soluble vascular cell adhesion molecule I and lactate dehydrogenase as predictors of survival in cutaneous metastatic malignant melanoma. Br J Cancer 1998;78:40 5. [76] Keilholz U, Conradt C, Legha SS, et al. Results of interleukin-2-based treatment in advanced melanoma: a case record-based analysis of 631 patients. J Clin Oncol 1998;16:2921 9. [77] Sirott M, Bajorin D, Wong G, et al. Prognostic factors in patients with metastatic malignant melanoma: a multivariate analysis. Cancer 1993;72:3091 8. [78] Kirkwood JM, Ibrahim JG, Sondak VK, et al. Highand low-dose interferon alfa-2b in high-risk melanoma: first analysis of intergroup trial E1690/S9111/ C9190. J Clin Oncol 2000;18:2444 58. [79] Kirkwood JM, Strawderman MH, Ernstoff MS, et al. Interferon alfa-2b adjuvant therapy of high-risk resected cutaneous melanoma: the ECOG Trial EST 1684. J Clin Oncol 1996;14:7 17. [80] Presant CA, Bartolucci AA and the Southeastern Cancer Study Group. Prognostic factors in malignant melanoma. The Southeastern Cancer Study Group Experience. Cancer 1982;49;2192 6. [81] Einhorn LH, Burgess MA, Vallejos C, et al. Prognostic correlations and response to treatment in advanced metastatic malignant melanoma. Cancer Res 1974;34: 1995 2004. [82] Amer MH, Al-Sarraf M, Vaitkevicius VK. Clinical presentation, natural history and prognostic factors in advanced malignant melanoma. Surg Gynecol Obstet 1979;149:687 92. [83] Rogers GS. Treatment of stage I melanoma: surgical and adjuvant therapy. In: Arndt KA, Robinson JK, LeBeoit RM, editors. Cutaneous Medicine and Sur-
[84]
[85]
[86]
[87]
[88]
[89]
[90]
[91]
[92]
[93]
[94]
[95]
[96]
[97]
G.S. Rogers, S.M. Braun / Dermatol Clin 20 (2002) 647658 ence of gender on survival in patients with stage I malignant melanoma. J Am Acad Dermatol 1992; 26:429 40. Reintgen DS, Vollmer R, Tso CY, Seigler HF. Prognosis for recurrent stage I malignant melanoma. Arch Surg 1987;122:1338 42. Shumante CR, Carlson GW, Giacco GG, et al. The prognostic implications of location for scalp melanoma. Am J Surg 1991;162:315 9. Schmoeckel C, Braun-Falco O. Prognostic index in malignant melanoma. Arch Dermatol 1978;114: 871 3. Reintgen DS, Paull DE, Seigler HF, et al. Sex related survival differences in instances of melanoma. Surgery 1984;159:367 72. Meyskens FL, Berdeaux DH, Parks B, et al. Cutaneous malignant melanoma (Arizona Cancer Center Experience) I: Natural history and prognostic factors influencing survival in patients with Stage I disease. Cancer 1988;62:1207 14. ODoherty CJ, Precott RJ, White H, et al. Sex differences in presentation of cutaneous malignant melanoma and in survival from stage I disease. Cancer 1986;58:788 92. Lees VC, Briggs JC. Effect of initial biopsy procedure on prognosis in stage I invasive cutaneous malignant melanoma: review of 1086 patients. Br J Surg 1991;78:1108 10. Sahin S, Rao B, Kopf AW, et al. Predicting ten-year survival of patients with primary cutanoeus melanoma: corroboration of a prognostic model. Cancer 1997;80:1426 31. Austin PF, Cruse CW, Lyman G, et al. Age as a prognostic factor in the malignant melanoma population. Ann Surg Oncol 1994;1:487 94. Cohen HJ, Cox E, Manton K, et al. Malignant melanoma in the elderly. J Clin Oncol 1987;5:100 6. Austin PF, Cruse CW, Lyman G, et al. Age as a prognostic factor in the malignant melanoma population. Ann Surg Oncol 1994;1:487 94. Loggie B, Ronan SG, Bean J, et al. Invasive cutaneous melanoma in elderly patients. Arch Dermatol 1991;127:1188 93. Shaw HM, McGovern VJ, Milton GW, et al. The female superiority in survival in clinical stage II cutaneous malignant melanoma. Cancer 1982;49:1941 4. Wanek LA, Elashoff RM, Goradia TM, et al. Application of multistage Markov modeling to malignant melanoma progression. Cancer 1994;73:336 43. Cochran AJ. Histology and prognosis in malignant melanoma. J Pathol 1969;97:459 68. Leon P, Daly JM, Synnestvedt M, et al. The prognostic implications of microscopic satellites in patients with clinical stage I melanoma. Arch Surg 1991; 126:1461 8. Vollmer RT. Malignant melanoma: a multivariate analysis of prognostic factors. Pathol Annu 1989; 24:383 407. Kopf AW, Gross DF, Rogers GS, et al. Prognostic
657
[116]
[98]
[99]
[117]
[100]
[101]
[118]
[102]
[119]
[120]
[103]
[121]
[104]
[122]
[105]
[123]
[106]
[124]
[107] [108]
[125]
[109]
[126]
[110]
[127]
[111]
[128]
[112] [113]
[129]
[114]
[130]
[115]
index for malignant melanoma. Cancer 1987;59: 1236 41. Sondak VK, Liu PY, Tuthill RJ, et al. Adjuvant immuno therapy of resected, intermediate-thickness, node-negative melanoma with an allogeneic tumor vaccine; overall results of a randomized trial of the Southwest Oncology Group. Journal of Clinical Oncology 2002;20(8):2058 66. Brocker E-V, Kolde G, Steinaussen D, et al. The pattern of the mononuclear infiltrate as a prognostic parameter in flat superficial spreading melanomas. J Cancer Res Clin Oncol 1984;107:48 52. Day CL, Sober AJ, Kopf AW, et al. A prognostic model for clinical stage I melanoma of the trunk: location near the midline is not an independent risk factor for recurrent disease. Am J Surg 1981;142: 247 51. Hansen MG, McCarten AB. Tumor thickness and lymphocytic infiltration in malignant melanoma of the head and neck. Am J Surg 1974;128:557 61. Day CL, Sober AJ, Kopf AW, et al. A prognostic model for clinical stage I melanoma of the lower extremity. Surgery 1981;89:599 603. Day CL, Sober AJ, Kopf AW, et al. A prognostic model for clinical stage I melanoma of the upper extremity: the importance of anatomic subsites in predicting recurrent disease. Ann Surg 1981;193:436 40. Sober AJ, Day CL, Fitzpatrick TB, et al. Early death from clinical stage I melanoma. J Invest Dermatol 1983;80:53s 5s. Sober AJ, Day CL, Fitzpatrick TB, et al. Factors associated with death from melanoma from 2 to 5 years following diagnosis in clinical stage I patients. J Invest Dermatol 1983;80:53s 5s. Mascaro JM, Molgo M, Castel T, Castro J. Plasma cells within the infiltrate of primary cutaneous malignant melanoma of the skin: a confirmation of its histoprognostic value. Am J Dermatopathol 1987;9: 497 9. Brocker EB, Zwadlo G, Suter L, Brune M, Sorg C. Infiltration of primary and metastatic melanomas with macrophages of the 25F9-positive phenotype. Cancer Immunol Immunother 1987;25:81 6. Brogelli L, Umberto MR, Moretti S, Urso C. The prognostic significance of histologic regression in cutaneous melanoma. Melanoma Res 1992;2:87 9. Cooper PH, Wanebo HJ, Hagar W. Regression in thin malignant melanoma. Arch Dermatol 1985;121: 1127 31. Kelly JW, Sagebeil RW, Blois MS. Regression in malignant melanoma: a histologic feature without independent prognostic significance. Cancer 1985;56: 2287 91. Trau H, Kopf AW, Rigel DS, et al. Regression in malignant melanoma. J Am Acad Dermatol 1983;8: 363 8. Wanebo HJ, Cooper PH, Hagar RW. Thin ( < 1mm) melanomas of the upper extremities are biologically favorable lesions not influenced by regression. Ann Surg 1985;201:499 504.
658
G.S. Rogers, S.M. Braun / Dermatol Clin 20 (2002) 647658 [145] Wang X, Heller R, Van Voorhis N, et al. Detection of submicroscopic lymph node metastases with polymerase chain reaction in patients with malignant melanoma. Ann Surg 1994;220:768 74. [146] Sarantou T, Chi DDJ, Garrison DA, et al. Melanomaassociated antigens as messenger RNA detection markers for melanoma. Cancer 1997;57:1371 6. [147] Curry BJ, Myers K, Hersey P. Utility of tests for circulating melanoma cells in identifying patients who develop recurrent melanoma. Recent Results Cancer Res 2001;158:211 30. [148] Hoon DS, Bostick P, Kuo C, et al. Molecular markers in blood as surrogate prognostic indicators of melanoma recurrence. Cancer Res 2000;60:2253 7. [149] Runnels LW, Yue L, Claphan DE. TRP-PLIK, a bifunctional protein with kinase and ion channel activities. Science 2001;291:1043 7. [150] Deeds J, Cronin FE, Duncan LM. Patterns of melastatin mRNA expression in melanocytic tumors. Hum Pathol 2000;31:1346 56. [151] Duncan LM, Deeds J, Hunter J, et al. Down-regulation of the novel gene melastatin correlates with potential for melanoma metastases. Cancer Res 1998; 58:1515 20. [152] Duncan LM, Deeds J, Cronin FE, et al. Melastatin expression and prognosis in cutaneous malignant melanoma. J Clin Oncol 2001;19:568 76. [153] Niezabitowski A, et al. Prognostic evaluation of cutaneous malignant melanoma: a clinicopathologic and immunohistochemical study. J Surg Oncol 1999;70: 150 60. [154] Vecchiato A, Rossi CR, Montesco MC, et al. Proliferating cell nuclear antigen (PCNA) and recurrence in patients with cutaneous melanoma. Melanoma Res 1994;4:207 11. [155] Smolle J, Hofmann-Wellenhof R, Kerl H. Prognostic significance of proliferation and motility in primary malignant melanoma of the skin. J Cutan Pathol 1992;19:110 5. [156] Soyer HP. Ki 67 immunostaining in melanocytic skin tumors: correlation with histologic parameters. J Cutan Pathol 1991;18:264 72. [157] Ramsay JA, From L, Iscoe NA, Kahn HJ. MIB-1 proliferative activity is a significant prognostic factor in primary thick cutaneous melanomas. J Invest Dermatol 1995;105:22 6. [158] Vilmer C, Bailey C, Le Doussal V, et al. Thin melanomas with unusual aggressive behavior: a report on nine cases. Melanoma Group of French Federation of Cancer Centers. J Am Acad Dermatol 1996;34: 439 44. [159] Thorn M, Adami HO, Ringborg U, et al. The association between anatomic site and survival in melanoma: an analysis of 12,353 cases from the Swedish Cancer Registry. Eur J Cancer Clin Oncol 1989;25:483 91.
[131] Gromet MA, Epstein WL, Blois MS. The regressing thin malignant melanoma: a distinctive lesion with metastatic potential. Cancer 1978;42:2282 92. [132] Ronan SG, Eng AM, Briele HA, Shioura NN, Das Gupta TK. Thin malignant melanoma with regression and metastases. Arch Dermatol 1987;123: 1326 30. [133] Sondergaard K, Hou-Jensen K. Partial regression in thin primary cutanoeus malignant melanomas clinical Stage I. Virchows Arch 1985;408:241 7. [134] Paladugu RR, Yonemoto RH. Biologic behavior of thin malignant melanomas with regressive changes. Arch Surg 1983;118:41 4. [135] Woods JE, Soule EH, Creagan ET. Metastasis and death in patients with thin malignant melanoma (less than 0.76mm). Ann Surg 1983;198:63 4. [136] Koh HK, Michalik E, Sober AJ, Lew RA, Day CL, Clark W, et al. Lentigo maligna melanoma has no better prognosis than other types of melanoma. J Clin Oncol 1984;2:994 1001. [137] Kopf AW, et al. Thickness of malignant melanoma: global analysis of related factors. J Dermatol Surg Oncol 1987;13:345. [138] Heenan PJ, English DR, Holman CD, Armstrong BK. Survival among patients with clinical stage I cutaneous malignant melanoma diagnosed in Western Australia in 1975/1976 and 1980/1981. Cancer 1991;68: 2079 87. [139] Day CL, et al. Classification of malignant melanoma according to histologic morphology of melanoma nodules. J Dermatol Surg Oncol 1982;8:874. [140] Fallowfield ME, Cook MG. Vascular invasion in malignant melanomas: an independent prognostic variable? Am J Surg Pathol 1989;13:217 20. [141] Langley R, Barnhill R, Mihm MC, et al. Neoplasms: cutaneous melanoma. In: Freedberg IM, Eisen AZ, Wolff K, et al, editors. Dermatology in general medicine. 5th edition. New York: McGraw-Hill; 1999. p. 1080 116. [142] Bieligk SC, Ghossein R, Bhattacharya S, et al. Detection of tyrosinase mRNA by reverse transcriptasepolymerase chain reaction in melanoma sentinel nodes. Ann Surg Oncol 1999;6:232 40. [143] Blaheta HJ, Schittek B, Breuninger H, et al. Detection of melanoma micrometastasis in sentinel lymph nodes by reverse transcriptase-polymerase chain reaction correlates with tumor thickness and is predictive of micrometastatic disease in the lymph node basin. Am J Surg Pathol 1999;23:822 8. [144] Goydos JS, Ravikumar TS, Germino FJ, et al. Minimally invasive staging of patients with melanoma: sentinel lymphadenectomy and detection of the melanoma-specific proteins MART-I and tyrosinase by reverse transcriptase polymerase chain reaction. J Am Coll Surg 1998;187:182 90.
Dermatol Clin 20 (2002) 659 676
The pathology of malignant melanoma

Robert J. Friedman, MD, MSc (Med)a,*, Edward R. Heilman, MDb
b a NYU School of Medicine, Department of Dermatology, 560 1st Avenue, New York, NY 10116, USA State University of New York Health Sciences Center at Brooklyn, Department of Dermatology and Pathology, Brooklyn, NY, USA
Mines a melanoblastoma, a real merciless bastard. As a rule, its eight months and youve had it. . . But the point is that even if Id come earlier a melanoblastoma is such a swine. You have only to touch it with a knife and it produces secondaries. You see, it wants to live too, in its way. . . Nobodys cured of melanoblasatoma. There just arent any instances of recovery. In my case, cutting of a leg wouldnt be enough, and where could they cut higher up? A. Solzhenitsyn Cancer Ward
[5] developed a system of classifying melanomas into histogenetic subtypes based on morphologic and biologic classification. These subtypes included (1) lentigo maligna (melanoma); (2) superficial spreading; (3) acral-lentiginous; and, (4) nodular.
Histogenetic subtype classification Lentigo maligna (melanoma) type This type of malignant melanoma was originally described as being commonly found on sun-damaged skin of the head and neck region. It generally evolves over many years and sometimes even decades. It begins as a small, generally symmetric pigmented patch having irregular borders and subtle variation in pigment ranging from tan to brown and sometimes black. Usually, the small patch evolves over time to become a larger and larger patch, often retaining the aforementioned asymmetry in shape, irregularity in border, variegation in color, and increasing diameter. This variant of melanoma may remain completely flat clinically, and generally confined to the epidermis histologically, for many months or years, and sometimes for decades. In a variable amount of time, however, the melanoma may penetrate into the deeper levels of the skin and develop the potential for metastases commensurate with its depth of penetration into the subjacent dermis. Superficial spreading type
Vadims statement in Solzhenitsyns famous novel, which takes place in a Russian hospital in the 1930s, depicts the feelings of hopelessness surrounding a diagnosis of malignant melanoma made only 70 years ago. Abysmally low survival rates of the 1930s have been replaced with overall 5- and 10-year survival rates well over 80% today with, 100% survival in patients having adequately treated in situ malignant melanoma. The latter diagnosis now accounts for nearly 50% of all melanomas diagnosed in the twenty-first century.
Clinicopathologic features of malignant melanoma There continues to be a dearth of information related to the precise histogenetic mechanisms involved in the pathogenesis of melanoma. Certainly, over the past 35 years this subject has been the topic of much scientific study [1 4]. Clark et al [1,2] and Elder et al
* Corresponding author. E-mail address: rfriedmanmdceo@hotmail.com (R.J. Friedman).
This type of malignant melanoma was originally described as most frequently occurring on the trunk
660
R.J. Friedman, E.R. Heilman / Dermatol Clin 20 (2002) 659676
and proximal extremities. This variant of melanoma generally begins as a small pigmented macule, which usually is asymmetric, having irregular borders, variegation in color, with subtle nuances of tans and browns and an enlarging diameter. The flat (intraepidermal in situ) phase is generally shorter than the lentigo maligna type lasting from a few months to a few years before developing a dermal component and the potential for metastasis proportionate to the depth of dermal penetration. Nodular type This type of malignant melanoma was originally described as occurring anywhere on the cutaneous surface, but most commonly on the trunk and proximal extremities. Essentially, this is a malignant melanoma with a very short macular stage (in situ) and often presents as a rapidly growing pigmented nodule. In more recent studies, careful historic evaluation often reveals the relatively short macular stage, which precedes the development of an often rapidly growing nodular component. Acral-lentiginous type This type of malignant melanoma was originally described as evolving on the acral regions of the body (hands, feet, and nail matrix). It is seen in all races but is most frequently found in dark-skinned races (Asians, blacks, and Hispanics). The clinical appear-
ance is similar to that of melanomas of the lentigo maligna and superficial spreading types with a moderately long macular (in situ) stage (months to years) before the development of a dermal component and the potential for metastasis commensurate with the thickness of the lesion.
Unifying concept for melanoma In 1980, Ackerman [4] formulated the so-called unifying concept to describe better the pathogenesis of melanoma. This was based on the concept that the histologic findings found in the aforementioned histogenetic subtypes of Clark et al [1,2] had more similarities than differences. Ackerman [3,4] postulated that all melanomas begin with a proliferation of single melanocytes within the epidermis (surface epithelium). Following a series of intraepithelial events, the melanocytes may enter the subjacent dermis (and potentially the deeper subcutis) and proliferate therein. It is from within the dermis and beyond (subcutis) that the cells of the melanoma may enter lymphatic and vascular channels and metastasize to distant sites. In 1982, Ackerman [6] expounded on the inaccuracy of the concepts of radial and vertical growth phase, stating that these terms were somewhat ambiguous. Clearly, the so-called radial growth phase, which implies that the melanoma is expanding in a more-or-less horizontal fashion, does not account for the fact that during the in situ phase of melanoma
Fig. 1. Low-power photomicrograph (original magnification 40) of early melanoma in situ from atypical pigmented lesion of trunk. Note the proliferation of single atypical melanocytes along the dermoepidermal junction and at higher levels of the epidermis. Early confluence of nests along the dermoepidermal junction. There is a patchy lymphohistiocytic dermal inflammatory cell infiltrate.
661
melanocytes enlarges the diameter of the neoplasm within both the epidermis and dermal and subcutis.
Observations on the histologic features of melanoma: update for the twenty-first century During the 15-year period 1983 to 1998, the authors group [7] performed an in-depth clinicopathologic study of over 1000 step-sectioned melanomas. What follows is a summary and description of some of the clinical and histologic features from that study.
Fig. 2. Higher-power photomicrograph of lesion in Fig. 1 (original magnification 400).
Histogenesis the melanocytes are expanding both radially and vertically. Radially, in the sense that the melanocytes are expanding (proliferating in a horizontal fashion within the epidermis [epithelium]), and vertically in the sense that the melanocytes are often proliferating upward into the higher levels of the epidermis (buckshot scatter of melanocytes into the epidermis) and downward within the epithelial structures of adnexa (Figs. 1 and 2). The term vertical growth phase is also somewhat ambiguous in that it is really not purely vertical. In fact, the cells of the melanoma are proliferating in both a vertical fashion (growing downward within the dermis and subcutis) and expanding radially as the proliferation of neoplastic The authors observations [7] confirmed many of those of Ackerman et al [3,4,6]. Almost all melanomas, without regard to their site of origin, had their beginnings with a proliferation of single melanocytes within the epidermis (epithelium) (Figs. 3 5). The authors attempted to classify the pathobiology of melanoma into at least two distinct stages: the intraepidermal (intraepithelial) phase and the intradermal phase. The second phase, which may well have multiple subphases, dictates the ultimate biologic outcome for the patient, namely the ability, or lack thereof, of the neoplasm to develop the competence for metastasis within a given host.
Fig. 3. Low-power photomicrograph of a more fully developed melanoma in situ from the trunk. Note the proliferation of atypical melanocytes arranged both singly and in nests throughout the epidermis. There is a patchy lymphohistiocytic dermal inflammatory cell infiltrate (original magnification 60).
662
those have the features of a large atypical melanocytic nevus (so-called dysplastic Clarks nevus) (Friedman RJ, et al, unpublished observations, 2002). Difficulties in the identification of precursor lesions There are a number of difficulties encountered by the dermatopathologist in evaluating the presence of a potential precursor melanocytic nevus within a melanoma. There is clearly a dilemma in differentiating the small cell variant of melanoma [1] from the cells of a melanocytic nevus. Possibly even more important is the ongoing controversy in developing and understanding specific, reproducible, and universally agreed on criteria to distinguish between the cells of a melanoma and those of some atypical (dysplastic) nevi. An error in interpretation of either of these alters the frequency of association between a so-called precursor lesion and the melanoma with which it is seemingly associated. It is important for the reader to keep these comments in mind until such time that a general consensus concerning the concept of precursor lesions is reached. Melanocytic nevi and melanoma: summary of current ideas The authors experience [8,9] along with that of others [5,10,11] suggest that there is an increased incidence of an identifiable melanocytic nevus in association with a melanoma in melanomas of the trunk and proximal extremities (superficial spreading
Fig. 4. Higher-power photomicrograph of Fig. 3 showing proliferation of single and nested melanocytes throughout all levels of the epidermis, even into the cornified layer. Note the prominent pleomorphism and hyperchromatism of the nuclei (original magnification 400).
Precursor lesions The authors original studies first published in 1991 [7] found that most (83%) malignant melanomas had their origins de novo without an antecedent or associated melanocytic nevus. About 17% of melanomas in that series were found to have an antecedent or associated melanocytic nevus. More recently, as yet unpublished observations reveal that in carefully examined step-sectioned and serially step-sectioned melanomas, the incidence of an associated melanocytic nevus is much higher, approaching more than 25% of melanomas examined. Many of
Fig. 5. High-power photomicrograph of typical pagetoid-type of melanoma in situ from the arm (original magnification 200). Note the prominent proliferation of single and nested atypical epithelioid melanocytes in a pagetoid array throughout the epidermis. The nests of melanocytes vary in size and shape and tend toward confluence. The atypical melanocytes are characterized as large with abundant pale-staining cytoplasms. The nuclei are both pleomorphic and hyperchromatic.
663
types) in contrast to all other sites (and types) of melanoma. It has been stated by some [5,10,12] that the melanocytic nevus identified in such material may be the precursor lesion involved in the progression of some melanocytic nevi into melanoma. This concept is quite intriguing but also very speculative. The study of so-called tumor progression had been a topic of debate in the world of anatomic pathology. The prototypic system studied extensively is that of the colonic polyp and colon cancer. The query as to whether such an analogy exists in the melanocytic system remains as yet not fully answered. Recent studies (Friedman RJ, et al, unpublished observations, 2002) [7] suggest that whereas overall some 17% to 25% or more of melanomas studied have an associated melanocytic nevus, most of those ( > 90%) were found in melanomas of the trunk and proximal extremities. Further, most (73% and 91%, respectively) had many of the histologic features seen in socalled large atypical (dysplastic Clarks nevi) nevi. Although not conclusive, these observations suggest that some melanocytic nevi may have a precursor relationship with some melanomas. Confirmatory observations are clearly needed to help solidify these concepts. Tumor progression model Various authors have devised a so-called model of tumor progression, which may be applicable to the melanocytic system [13,14]. This model essentially is a stepwise pathway in which there is aberrance in a melanocytic nevus in terms of its normal life cycle
(simple lentigo, to junctional nevus, to compound nevus, to dermal nevus with potential involution later in life). The aberrancy in the affected melanocytic nevus has been described as the continued proliferation of melanocytes within the epidermis (intralesional transformation) of an otherwise developing compound nevus such that there becomes a proliferation of intraepidermal melanocytes with a disorderly growth pattern and random cytologic atypia (melanocytic dysplasia). Clark et al believe that the histologic changes seen initially in melanoma in situ and later in invasive melanoma may occur in some of these atypical variants of melanocytic nevi [1]. Whether or not these concepts have any validity remains to be proved. From a biologic standpoint, a number of groups, including the authors, have found that melanomas arising in association with melanocytic nevi, and possibly within atypical (dysplastic) nevi (Fig. 6), when compared with those melanomas arising de novo, have a significantly lower rate of metastasis and death (better prognosis) [8,15]. Further, Sober et al [16] reported that melanoma arising de novo were much more likely to be associated with early death in patients with clinical stage I disease. Nevus-associated melanomas seem to have a different biologic behavior that those arising de novo [7,15].
The pathology of melanoma in the twenty-first century: a blending of ideas The histogenetic classification system of Clark et al [1,2] was based originally on histologic studies
Fig. 6. Another view of the neoplasm depicted in Fig. 5. Note the buckshot scatter of the atypical melanocytes throughout all levels of the epidermis (original magnification 200).
664
Table 1 Melanoma frequency by antatomic site Site Trunk or proximal extremity Head or neck Acral Other (mucosal) Number (%) 791 140 50 21 (79) (14) (5) ( < 2)
included completely removed (biopsy en toto) lesions; step-sectioning at 2-mm intervals; and adequate clinical follow-up information Approximately one third (309 of 1002) of melanomas received sequentially from clinicians in a large subspecialty dermatopathology laboratory (Dermpath, Inc., Scarsdale and Port Chester, NY) were found to be in situ. A breakdown by anatomic site of the 1002 melanomas is found in Table 1. Melanoma in situ Melanoma of the trunk and proximal extremities The most common sites for melanoma to occur are those on the trunk and proximal extremities [28, 29] (Figs.1 8). Further, for melanomas found in association with a melanocytic nevus, these anatomic sites also predominated. The Clark histogenetic subtype superficial spreading melanoma is the prototype for melanomas of the trunk and proximal extremities. In the authors comprehensive study published in 1991 [7], about 28% of melanomas found in these anatomic sites were in situ, accounting for some 71% of all in situ melanomas identified. Histologically, the earliest identifiable features of melanomas of the trunk and proximal extremities consisted of an increased number of single melanocytes, many of which had a normal cytologic appearance along and perhaps slightly above the dermoepidermal junction. A few cytologically atypical cells having somewhat larger hyperchromatic and pleomorphic nuclei were identified. Melanomas having such histologic features were generally small (< 4 mm in diameter). Somewhat larger melanomas (those identified more easily clinically and approximately 4 to 6 mm in diameter) are characterized histologically by having
of the so-called radial growth phase of melanomas. By definition, Clark et al [1,2] used the concepts that the radial growth phase consisted of both the intraepidermal component of the evolving melanoma along with the presence of small collections of melanocytes, and single cells of melanocytes within the papillary dermis (microinvasion). It seems, however, that in all practicality the pattern of growth within the epidermis and the reputed differences dictated the classification of melanomas by subtype. After reviewing the literature and the authors indepth clinicopathologic observations, the authors conclude that the previously reported differences in intraepidermal growth patterns are not quite so constant as that previously reported. Further, there clearly seems to be many more similarities than differences histologically between melanomas. From a clinical perspective, melanomas seem to have their clinical beginnings in a common fashion, which does not seem to be anatomic site dependent. Although evolution of a lesion may differ from site to site (eg, melanomas from sun-damaged skin of the head and neck, so-called lentigo maligna [melanoma], may have a longer biologically benign [in situ] stage when contrasted with melanomas from the trunk and proximal extremities), it is unclear whether this is site determined or determined by some other differences (eg, severe solar damage). For example, some melanomas on severely sun-damaged skin of the upper back and shoulders have a long biologically benign growth phase (in situ phase) similar to that of melanomas of the head and neck. Is it the marked solar elastosis that is present (which may inhibit neoangiogenesis and slow growth of the neoplasm), or the fact that the lesion is present on a specific anatomic site, which dictates clinicohistologic differences? As clinicians try to understand better the relationship between morphology and biology, simplistic ideas often go by the wayside. The following discussion is based on a review of the pertinent literature coupled with a comprehensive study of over 1000 step-sectioned melanomas from patients throughout the New York metropolitan area studied and followed during a 15-year period (1983 to 1998) [7]. The criteria for inclusion in the study
Fig. 7. Still higher-power (original magnification 400) photomicrograph of lesion in Fig. 5 showing marked atypicality of the melanocytes. A giant atypical melanocyte is seen.
665
Fig. 8. Photomicrograph of melanoma in situ on sun-damaged skin of the upper back (original magnification 200). Note the presence of atypical melanocytes both singly and in nests along the dermoepidermal junction and at higher levels of the epidermis. Dermal solar elastosis is present.
increased numbers of both single and nested melanocytes, many of which are noted to have cytologic atypia. Some or many of these melanocytes are noted to be present at higher levels of the epidermis. Involvement of at least the upper portions of the adnexal epithelium is common. The intraepidermal lesion was generally asymmetric in configuration with single melanocytes arranged in a nonequidistant fashion from one another at the periphery. It is the peripheral asymmetry of the intraepidermal melanocytic proliferation, coupled with the presence of melanocytes at varying levels of the epidermis that confers the asymmetry, border irregularity, and play in color commonly seen clinically in melanomas. As the melanocytes continue to proliferate within the epidermis and the diameter of the lesion enlarges, the initially ill-defined nests become better defined. In most instances, the nests tend to be shaped irregularly and have a tendency toward confluence. The nests themselves are symmetrically located within the expanding intraepidermal neoplasm. Their presence at higher levels of the epidermis interspersed with atypical single melanocytes often in pagetoid array is a classic finding in melanomas of the trunk and proximal extremities, particularly melanomas arising on skin that is not very sun damaged. The brown-black color seen in many melanomas is caused by the presence of melanin-laden melanocytes at all levels of the epidermis, including the cornified layer. The essence of dermoscopy also has its basis in these observations. There is some variability in the histology of the epidermis in melanomas from the trunk and proximal
extremities. In most instances, the epidermis is normal to somewhat hyperplastic. In some cases, the rete ridges may be elongated (lentiginous). The extent of solar elastosis is also quite variable from minimal or none to quite marked. The latter is the case of melanomas involving the posterior shoulders and upper back (see later) and may be associated with a histologic pattern more consonant with that of melanomas found on sun-damaged skin of the head and neck. The cytologic appearance of the atypical melanocytes seen in melanomas of the trunk and proximal extremities is somewhat variable, ranging from the classic pagetoid (large melanocyte with abundant eosinophilic-staining cytoplasms, oftentimes containing melanin pigment) to small cuboidal and spindleshaped types of melanocytes. The inflammatory cell infiltrate seen within the dermis in association with melanomas from the trunk and proximal extremities is also quite variable ranging from patch, to absent, to moderate, to dense. The cell type in most instances for in situ lesions generally consists of lymphocytes and histiocytes, including a few melanophages. In more advanced lesions involving the dermis and subcutis, plasma cells may be seen, generally portending a poorer prognosis. Histologic Regression. The presence of histologic regression (generally partial, but rarely total) is manifest by the presence of a dense, lichenoid inflammatory cell infiltrate of lymphocytes and histiocytes, often accompanied by melanosis (the presence of
666
Fig. 9. Photomicrograph (original magnification 60) of site of near totally regressed melanoma of the back. The epidermis had a flattened rete ridge pattern with only a rare melanocyte identified along the dermoepidermal junction. There is marked fibrosis and thickening of the papillary dermis with numerous melanophages and a patchy to moderately dense lymphohistiocytic inflammatory cell infiltrate.
numerous melanophages within the infiltrate) and later, focal or broad zones of fibrosis and thickening of the papillary dermis (Fig. 9). It is this latter finding that is manifest clinically as focal or broad zones of pink-white within the lesion. The implications of these histologic findings are related to the fact that the aforementioned histologic findings occur as a response to the presence of a few (or more) atypical melanocytes within the subjacent dermis, which were destroyed or partially destroyed by the bodys immune response to the neoplasm. Often only the intraepidermal portion of the neoplasm is left unscathed. Rarely, however, the immune process may destroy most if not all of the melanoma, including both the epidermal and dermal components, leaving behind only the battlefield of dermal fibrosis, a patchy lymphohistiocytic infiltrate, and scattered melanophages. Although the lesion may be completely destroyed by the bodys immune system, there was a time when the neoplastic melanocytes were present within the dermis with the inherent potential for metastasis. One should be wary of the risk for metastasis in patients having partially or completely regressed melanomas. Melanomas of the trunk and proximal extremities arising on severely sun-damaged skin. This histologic variant of melanoma of the trunk and proximal extremities (and rarely seen on the dorsal acral areas where there is marked solar damage in melanomas of acral sites) is characterized histologically as follows. The epidermis is generally atrophic (thinned), often with a flattened rete ridge pattern. Extensive dermal
solar elastosis is present. Early histologic changes within the epidermis include an increased number of single melanocytes with only a minimal amount of cytologic atypia. In time, nests of melanocytes, often tending toward confluence, are identified. There is very little upward migration of either nests or single melanocytes, but often there is extensive involvement of the epithelial structures of adnexa. Although there is little difference in the clinical presentation of early melanomas of this variety as compared with the remaining more commonly seen melanomas of the trunk and proximal extremities (pigmented, asymmetric maculae with irregular borders and ranging in color from tan to brown to somewhat darker brown), the larger lesions tend to be somewhat lighter (less dark brown and black) because of the relative absence of melanin pigment at higher levels of the epidermis. There seems to be little difference in the biology of these lesions in terms of prognosis; however, there may be some differences historically in terms of the duration of the in situ phase of growth. Melanoma of sun-damaged skin of the head and neck In the authors comprehensive study of over 1000 melanomas [7], melanomas of the head and neck accounted for about 14% of patients studied (Figs. 10 18). The smallest lesions clinically thought to be melanoma were approximately 4 mm in diameter. Histologically, these lesions were characterized by an increased number of melanocytes, some of which were larger than normal and having subtle pleomorphic and hyperchromatic nuclei. All were present on
667
Fig. 10. Photomicrograph of melanoma in situ from sun-damaged skin of face. The melanocytic proliferation is confined essentially to the dermoepidermal junction and only slightly above it. Single atypical melanocytes predominate over nests, which tend to vary in their sizes and shapes. Prominent dermal solar elastosis is present along with a patchy lymphohistiocytic inflammatory cell infiltrate (original magnification 100).
markedly solar-damaged skin as evidenced by abundant solar elastosis of the superficial dermis. A subtle clue to the diagnosis of an early evolving melanoma in situ on sun-damaged skin of the head and neck is the uneven distribution of single melanocytes at and only slightly above the dermoepidermal junction. Often there is also an uneven distribution of melanin pigment within the cytoplasms of the melanocytes. In most instances, the epidermis was noted to be atrophic; however, in some cases it was either normal or even hypertrophic. As the neoplasm evolves and becomes clinically larger, nest of melanocytes become apparent. Such nests were similar to those seen in me-
lanomas at other sites. In time, the nests tend toward confluence. Epithelial adnexal involvement is a relatively early finding in melanomas from these sites. In about 70% of melanomas of the head and neck, the intraepidermal component of the neoplasm was confined to the lower one half of the epidermis. The remaining 30% of head and neck melanomas had histologic changes more consistent with melanomas of the trunk and proximal extremity. In such cases, there was considerably less solar elastosis noted. All cell types were found in melanomas of the head and neck. Most commonly, however, the spindle and cuboidal cell types were identified. About one third of
Fig. 11. Higher-power (original magnification 250) photomicrograph of Fig. 10 depicting increased number of atypical melanocytes (epithelioid and spindle shaped) arranged mostly as single units, but also in a few nests along the dermoepidermal junction and slightly above it.
Fig. 12. Still higher-power photomicrograph of Fig. 10 depicting increased numbers of atypical melanocytes within the lower one half of the epidermis (original magnification 400)
668
Fig. 13. Photomicrograph of melanoma in situ from cheek. There are an increased number of single and nested melanocytes along and slightly above the dermoepidermal junction and involving the epithelial structure of adnexa (left side). Marked solar elastosis is present within the dermis and there is a patchy to moderately dense lymphohistiocytic inflammatory cell infiltrate within the dermis (original magnification 100).
melanomas were noted to have epithelioid-pagetoid melanocytes. As the lesion continues to grow, often reaching many millimeters or even centimeters in diameter, the classic findings include asymmetric proliferation of both single atypical melanocytes and irregularly arranged nests and confluent nests of melanocytes mostly along the dermoepidermal junction and only slightly above it. Historically, the in situ stage of development of melanomas of the head and neck lasts longer than that of melanomas of the trunk and proximal extremities. Sometimes, the clinically enlarging pigmented macule may evolve over years or even decades. The more prominent the sun damage, the longer the intraepidermal evolution. Prominent
adnexal involvement is a common finding. A variably dense lymphohistiocytic inflammatory infiltrate may be found in the subjacent dermis. Histologic features of regression may be seen at times. There are two histologically distinct patterns seen in melanomas from the head and neck. On markedly sun-damaged skin, the atypical melanocytic proliferation tends to be confined to the lower epidermis and evolves over many months, years, or even decades as melanoma in situ. On less sun-damaged skin, the pattern is more consonant with melanomas found on the trunk and proximal extremities with buckshot scatter of more epithelioid-pagetoid melanocytes
Fig. 14. Higher-power photomicrograph of Fig. 13 depicting the increased number of atypical melanocytes as single units at and slightly above the dermoepidermal junction (original magnification 250).
Fig. 15. Higher-power photomicrograph of Fig. 13 showing involvement of the upper portion of the hair follicle with atypical melanocytes. Atypical melanocytes are also seen along and slightly above the dermoepidermal junction (original magnification 250).
669
Fig. 16. Higher-power photomicrograph depicting prominent involvement of entire hair follicle epithelium with single and nested atypical melanocytes (original magnification 250).
throughout the epidermis. In such cases, evolution of the in situ lesion into one with a dermal component seems to be shorter. Aside from a suggestion that melanomas with the latter pattern have clinically darker brown or black in their lesion, the two histologic subtypes of melanoma of the head and neck are otherwise indistinguishable clinically. Melanoma arising on the skin of the acral-mucosal regions. In the authors large series of patients [7], melanomas of the acral sites, including nail matrix, and melanomas of the mucosa accounted for only 5% of all melanomas studied (Figs. 19 21). Approximately 35% were found on blacks and Asians.
Fig. 18. Higher-power photomicrograph of superficially invasive (0.32 mm) melanoma from sun-damaged skin of the side of the neck. Note the more pagetoid pattern where single and nested melanocytes are present throughout the epidermis in a somewhat buckshot array. Within the superficial papillary dermis, there are a few nests of atypical melanocytes (original magnification 200).
The very early histologic features of early evolving melanoma of the palms and soles are subtle and at times difficult to distinguish from those seen in acral melanocytic nevi. Histologically, these changes include increased numbers of singly arranged melanocytes, many having prominent dendritic processes. Sometimes, these processes are present at all levels of the epidermis and are a subtle clue to the diagnosis of melanoma in situ at this site. At times, there may be only a subtle increased number of single melanocytes, but a prominent number of thick dendritic
Fig. 17. Photomicrograph of an early melanoma in situ of sun-damaged skin from the temple. In this photo, there are increased numbers of mostly single atypical melanocytes along the dermoepidermal junction. A few atypical melanocytes are present at slightly higher levels of the epidermis. Marked dermal solar elastosis is present in the superficial dermis where there is a patchy lymphohistiocytic inflammatory cell infiltrate (original magnification 40).
670
Fig. 19. Photomicrograph of melanoma of sole of foot showing typical acral epithelium and an increased number of atypical melanocytes arranged both singly and in nest at and above the dermoepidermal junction (original magnification 100).
processes found within the upper reaches of the epidermis. As the lesion evolves, the number of spindle or cuboidal melanocytes, some of which may be cytologically atypical and often having prominent dendritic processes, are identified. Increased numbers of nests, some of which tend toward confluence, are identified within the epidermis. Initially, these nests are confined to the dermoepidermal junction. In time, they are present also at higher levels of the epidermis. The epidermis is usually normal or hyperplastic. Often, the rete ridge pattern is accentuated. Involvement of the epithelial structures of adnexa is commonly marked. Solar elastosis is minimal or absent.
A second pattern of histologic involvement has also been identified. The latter pattern is characterized by a more pagetoid nature. There is often buckshot scatter of epithelioid-pagetoid, cuboidal-spindle atypical melanocytes arranged both singly and in nests throughout all levels of the epidermis and frequently involving the sweat duct epithelium.
Melanoma in situ: a summary Primary cutaneous malignant melanoma develops fundamentally as a single pathological process. . ..the various types of malignant melanoma have many more features in common than they have differ-
Fig. 20.Higher-power (original magnification 200) photomicrograph of lower portion of epidermis showing and increased number of spindle-dendritic melanocytes having pleomorphic and hyperchromatic nuclei at and above the dermoepidermal junction. Note the prominent dendritic processes.
Fig. 21. Still higher-power photomicrograph at interface between the stratum spinosum and stratum corneum showing increased number of atypical spindle-shaped melanocytes and their prominent dendritic processes (original magnification 400).
671
ences. . . [3]. It is clear that, after reviewing in detail over 1000 step-sectioned melanomas, there is little advantage in subclassifying melanomas by subtype because all melanomas have their origins similarly from neoplastic intraepidermal melanocytes. The basic pathogenesis of melanoma within the epidermis is similar regardless of anatomic site. Histologically, there is little evidence to suggest that any one anatomic site has one specific microscopic pattern. Further, on the clinical side, melanomas have recognizable origins in a similar way regardless of anatomic site. Melanomas of the head and neck, trunk and proximal extremities, and acral sites all have a common clinical presentation. The authors clinicopathologic review of over 1000 patients strongly supports the use of the concept of the ABCDs in diagnosing melanoma clinically (present in 96% of patients having melanoma in situ and 85% of patients having invasive melanomas) [27]. The histologic features that account for the ABCDs clinically are an asymmetric proliferation of melanocytes within the epidermis and dermis; growing asynchronously (asymmetric lesion having irregular borders [A and B]); with melanocytes at varying levels of the epidermis and dermis (color variability [C]); and relentlessly expanding at various rates in all dimensions (diameter enlarging [D]). As with any model, there are rare exceptions to this rule, especially seen in those melanomas that are more rapidly growing (eg, some spitzoid melanomas seen in children; some melanomas having a rapidly growing dermal component where the dermal component of the neoplasm overshadows the intraepidermal component). The authors concur with Ackerman et al [3] that melanomas should be considered to develop fundamentally as a single pathologic process with variations in morphology resulting from yet to be determined influences (eg, sun damage). The biology of growth of a melanoma involving the dermis Previously, the discussion was centered on the histology of melanoma on melanomas confined to the epidermis. Left untreated, cutaneous melanoma may evolve from in situ melanoma to melanoma that has permeated into the subjacent dermis or even subcutis. The authors have termed the first histologic stage of melanoma the biologically benign intraepidermal growth phase. They term the second histologic stage of melanoma the intradermal growth phase. The intradermal growth phase has two components: the early intradermal growth phase and the late intradermal growth phase. A number of investigators [10,12,15] believe that there is a stage in the evolu-
tion of melanoma following the intraepidermal (in situ) stage when neoplastic melanocytes are present within the dermis (early intradermal stage), yet the melanoma does not as yet have competence for metastasis. Clark et al [17] and Elder et al [12] describe a stage in the evolution of melanoma termed progressive autonomous growth without tumorogenecity, or the capacity for metastasis. Histologically, this stage of growth is characterized by atypical melanocytes involving the epidermis with possible involvement of the papillary dermis with single melanocytes or small nests of melanocytes (microinvasion). Herlyn et al [18,19] illustrated the fact that melanocytes of microinvasive melanomas were difficult to grow in vitro, failed to form permanent cell lines, grew to low density only, and were not tumorigenic in nude mice. Elder et al [14] showed that in a study comprised of 93 patients having in situ or microinvasive melanoma, the survival rate at 7 years was 100%. The second component of intradermal growth of a malignant melanoma is one the authors term late intradermal growth phase and others term vertical growth phase [10]. Because the neoplastic melanocytes are not only proliferating in a vertical direction, but are essentially growing in all directions, the authors prefer their terminology to that of Clark et al [10] and substitute that terminology in subsequent discussion. Herlyn et al [18,19] showed that the neoplastic melanocytes of late intradermal growth phase melanomas are fully transformed and immortal in tissue culture and grow to a high cell density. Further, they have been found to have high plating efficiency and are tumorigenic in nude mice. Finally, chromosomal analyses illustrate multiple random and nonrandom abnormalities. Such neoplasms tend to be relentless in their growth, often forming multiple nodules both clinically and histologically. Some of these nodules may be comprised of different cell types, some of which are quite atypical and may exhibit high rates of mitosis. Some are biologically capable of metastasis. Necrosis is a common feature seen in the rapidly growing neoplastic melanocytes. As the neoplasm expands within the dermis, it evokes neoangiogenesis and may enter such vessels and spread to local or distant sites. At times, the rapidly growing neoplasm may cause ulceration of the overlying epidermis and if left untreated may proliferate deeply into the underlying subcutis. Clinically, the infiltration of the dermis with atypical melanocytes alters the surface configuration of the lesion. The initially macular in situ melanoma may develop a superimposed papule, and later may become plaquelike or even nodular.
672
Histologically, the nature of the intradermal growth phase is similar from site to site. The cellular morphology may be epithelioid, spindle, cuboidal, or small. Any given site can have any different cell type; however, epithelioid cells seem to be more prominent on melanomas of the trunk and proximal extremities, spindle-shaped cells tend to be more common on head and neck melanomas, and spindle-dendritic cells tend to be more common on acral melanomas. There are a few rules common to most melanomas within the dermis-subcutis: (1) confluence of dermal nests of atypical melanocytes; (2) lack of maturation of nuclei with progressive descent into the dermis; (3) mitotic figures may be present, particularly at the base of the dermal component of the melanoma; and (4) atypical melanocytes are often necrotic. The nodular melanoma: does it exist? Most melanomas of the trunk and proximal extremities were greater than 5 mm in diameter by the time a dermal component of the melanoma was identified. Less than 1% of the melanomas identified in the authors series [7] were found to have a dermal component when they were less than 5 mm in diameter. One can presume that such lesions, although rare, account for the so-called nodular melanoma described by Clark et al [1,2]. In the authors series, the thickest melanoma measuring less than 5 mm in diameter was 0.72 mm in thickness. In all instances, however, these small melanomas having early involvement of the dermis had readily identifiable (on step-sectioning) intraepidermal (in situ) growth components. An additional 23 melanomas having diameters between 6 and 16 mm had thickness measurements of approximately 1 mm. Many of these lesions could be classified as the so-called nodular type of melanoma using the Clark method of classification. All nonulcerated lesions were found to have (on step-sectioning) unequivocal intraepidermal (in situ) components. Clark et al [1,2] stated nodular melanoma differs histogenetically and morphologically from the other three types of melanoma by having an absence of, or a very limited intraepidermal component. As noted previously, a very small percentage of melanomas studied could have been classified as nodular. In nearly all cases, however, a readily apparent intraepidermal component of the lesion was identified on step-sectioning. Heenan and Holman [20] in reviewing over 400 melanomas having the histologic features of nodular melanoma found in many of them a distinct intraepidermal component. They concluded, as do the authors, that some melanomas having their origin at any anatomic site may develop within a
previous melanoma in situ, a more rapidly growing nodular component. Whether this fact is caused by an inherent difference in the biologic nature of the neoplastic cells, a diminished host-response, or other as yet unknown factors remains a mystery. It is the authors opinion that nodular melanoma may not be a discrete histogenetic subtype, but rather a result of an interplay of factors on a melanoma from any site and with any microscopic morphology. Melanomas involving the dermis from the trunk and proximal extremities Generally, most melanomas from the trunk and proximal extremities having dermal involvement were greater than 5 mm in diameter. Most were over 10 mm in diameter at the time of excisional removal. Four distinct cell types were found in trunk and proximal extremity melanomas: (1) epithelioid, (2) spindleshaped, (3) small-cuboidal, and (4) giant. The epithelioid cell type was the most common cell type found at these sites. Classically, the epithelioid cell is a large cell having abundant pale-staining cytoplasm, often containing abundant melanin pigment. They tend to grow within the dermis as sheets or confluent irregularly shaped nests with poorly defined cell borders. The nuclei are generally large and vesicular and tend to vary somewhat in their sizes. A prominent eosinophilic-staining nucleolus or multiple nucleoli is common. Mitotic figures may be identified, particularly at the base of the lesion. The next most common cell type is the spindleshaped cell. This cell type is defined as being an elongated melanocyte having tapering ends. Nuclear atypia may vary from cell to cell. Melanin pigment may or may not be present. This type of melanocyte was more often seen on melanomas from markedly sun-damaged skin of the posterior shoulders and upper back. In such cases, rarely this variant of melanocyte was seen in association with marked desmoplasia (fibrosis) in a variant of melanoma called desmoplastic melanoma. This variant of melanoma is more commonly seen on sun-damaged skin of the head and neck, specifically the face or scalp. An infrequently found cell type ( < 8%) in trunk and extremity melanomas is the so-called small-cell type. When present within the dermal component of the neoplasm, it often mimics a melanocytic nevus. Careful observation usually shows a significant amount of nuclear atypia, although at times this may be absent. Other findings that may help to distinguish small melanoma cells from nevus cells include (1) size often greater than 10; (2) occasional necrotic cells; (3) plasma cells at base of neoplasm rarely; (4) nests of melanocytes and individual cells
673
do not become smaller (and may become larger) with progressive descent into the dermis; (5) cells may contain melanin pigment; and (6) absence of fibrillary collagen between individual cells. A rare cell type (giant cell) may be seen in less than 1% of melanomas from the trunk and proximal extremities. When present, these cells have a single huge, bizarre, and hyperchromatic nucleus. Melanomas involving the dermis arising on sun-damaged skin of the head and neck The dermal component of melanomas arising on sun-damaged skin of the head and neck is comprised of atypical melanocytes having no distinctive difference from the melanocytes found on other body sites. Spindle-shaped melanocytes predominated (>50%) in a recent study. Melanomas arising on the head and neck have two distinctive histologic patterns: basilar and pagetoid. The dermal components of both types were similar, with a greater percentage of spindle cells present in the basilar variant. Desmoplastic and neurotropic melanomas The desmoplastic melanoma is a rare variant of melanoma ( < 1%), which can be found on any anatomic site, but is more commonly found on severely sun-damaged skin of the head and neck. First described in 1971 by Conley et al [21], the desmoplastic melanoma is generally a deeply invasive variant of melanoma comprised of spindle-shaped cells and marked desmoplasia (fibroplasia). At times, the desmoplasia can be misdiagnosed as a fibromatosis or a hypertrophic scar. A search for the overlying intra-
epidermal component of the melanoma and use of immunoperoxidase staining are often necessary to make a correct diagnosis. Neurotropic melanoma is a variant of desmoplastic melanoma, which occurs predominantly on the head and neck. Clinically, the lesion often occurs as an innocuous pink papule sometimes within a larger pigmented macular lesion. Histologically, neurotropic melanomas have many of the features of desmoplastic melanoma but also involve and extend along the peripheral nerves well beyond the bulk of the neoplasm. Use of immunoperoxidase stains, such as S100, often is useful to delineate the extent of the neoplastic process. At times, the neurotropic melanoma extends along peripheral nerves through foramina (eg, on the face) and enters the brain. Uncommonly, neurotropic melanomas may metastasize. Melanomas involving the dermis in acral areas In the authors recent in-depth study of more than 1000 step-sectioned melanomas [7] there was no distinctive pattern or cell type specific for the dermal component of acral melanomas (Figs. 19 22). About 41% of acral melanomas had spindle cells as their predominant cell type. Histologic reporting of melanoma The dermatopathologists role is to present information to the clinician that is both accurate and clinically relevant. Accuracy in diagnosis is paramount followed by the reporting of relevant information about prognosis (eg, thickness, mitotic rate,
Fig. 22. Photomicrograph of metastatic melanoma to skin (satellite metastasis). Note the absence of epidermal involvement and the large nests of atypical melanocytes within the dermis (original magnification 100).
674
Fig. 23. Photomicrograph (original magnification 100) of epidermotrophic metastatic melanoma. Note the triangular-shaped proliferation of atypical melanocytes with the apex of the triangle being the focus of epidermotrophic metastatic melanoma. The remainder of the epidermis is devoid of melanoma and there is a broadly based lesion of metastatic melanoma in the dermis
ulceration, and so forth); site; excision margins; and any special features (eg, intraspecimen metastases). Measurement of thickness Breslow [22] originally described a method establishing what has become the most important prognostic parameter, lesion thickness. Using an ocular micrometer, thickness measurement is accomplished by measuring (in millimeters) the vertical distance from the top of the granular layer (of the epidermis) or base of the ulcer (if the lesion is ulcerated) to the deepest atypical melanocyte in the dermis or subcutis. Mitotic rate of melanoma Schmoekel et al [23] identified a so-called prognostic index to improve further the ability to determine prognosis. The mitotic index is defined as the maximal number of dermal mitotic figures per millimeters squared. The mitotic index multiplied by the Breslow thickness equals the prognostic index. Highrisk melanomas are defined as having a thickness of 3 or more mm and a prognostic index of 13 or more. Medium-risk melanomas are defined by having a prognostic index of 1.1 to 12.9. Low-risk melanomas have a thickness of 0.76 or less and a prognostic index of 1 or less. Ulceration Ulceration of the epidermis is an important prognostic feature and should be reported when present. Epidermal ulceration greater than 3 mm in breadth
has been reported to be associated with a poorer prognosis [24]. Microscopic satellites and vascular invasion Patients having intraspecimen satellitosis have a poorer prognosis and should be included in the pathology report (Figs. 22 24) [25]. The presence of atypical melanocytes within an endothelial-lined channel is also associated with a poorer prognosis and should be reported. Melanocytic nevi and melanoma The association of melanoma and melanocytic nevi has been reported as ranging between 18% and 83% [26]. Data from the authors previous study of
Fig. 24. High-power photomicrograph of intravascular metastasis. Note the atypical melanocytes within the endothelial-lined channel (original magnification 250).
675
1002 step-sectioned melanomas [7,8] indicated that approximately 17% of melanomas had histologic evidence of a coexisting melanocytic nevus. More recently (Friedman RJ, et al, 2002, unpublished observations), the authors have found that nearly 27% of serially step-sectioned melanomas have histologic evidence of a coexisting melanocytic nevus. Most of these had features of a so-called dysplastic Clarks atypical nevus. These findings suggest that there are as yet many unanswered questions regarding the meaning of the so-called dysplastic nevus. Surely, it is at least a marker for, and likely a formal precursor of, at least some malignant melanomas. Further refinements in the diagnosis of these atypical lesions to make them generally agreed on, clear, accurate, and reproducible are necessary to enable clinicians to understand better the significance of their relationship to melanoma. Tumor volume: a new and intriguing predictor of survival in melanoma Preliminary studies [7] show that computer calculation of tumor volume may be a more important predictor of survival than thickness.
Clinicians should keep their minds open to new concepts and try to separate what makes sense from that which does not.
What everybody echoes or in silence passes by as true today may turn out to be falsehood tomorrow, mere smoke of opinion. . . Thoreau, Walden
References
[1] Clark Jr. WH, From L, Bernardino EA, et al. The histogenesis and biologic behavior of primary malignant melanomas of the skin. Cancer Res 1969;29:705. [2] Clark Jr. WH, Goldman LI, Mastrangelo MJ. Human malignant melanoma. New York: Grune and Stratton; 1979. [3] Ackerman AB, Su WPD. The histology of cutaneous malignant melanoma. In: Kopf AW, et al, editors. Malignant melanoma. New York: Masson; 1979. p. 25. [4] Ackerman AB. Malignant melanoma: a unifying concept. Am J Dermatopathol 1980;2:309. [5] Elder DE, Jucovy PM, Tuthill RJ, et al. The classification of malignant melanoma. Am J Dermatopathol 1980;2:315. [6] Ackerman AB. Disagreements about classification of malignant melanoma. Am J Dermatopathol 1982;4:447. [7] Friedman RJ, Rigel DS, Kopf AW. Cancer of the skin. Philadelphia: WB Saunders; 1991. [8] Friedman RJ, Rigel DS, Heilman ER. The relationship between melanocytic nevi and malignant melanoma. Dermatol Clin 1988;6:249. [9] Kopf AW, Welkovich B, Frankel RE, et al. Thickness of malignant melanoma: global analysis of related factors. J Dermatol Surg Oncol 1987;13:345. [10] Clark Jr. WH. The development and biologic significance of neoplasia. The Lila Gruber Memorial Cancer Research Award. American Academy of Dermatology. Chicago, December, 1984. [11] Kraemer KH, Greene MH, Tarone R, et al. Dysplastic nevi and cutaneous melanoma risk. Lancet 1983; 2:1076. [12] Elder DE, Guery D, Epstein MN, et al. Invasive malignant melanomas lacking competence for metastasis. Am J Dermatopathol 1984;6:55. [13] Clark Jr. WH, Mastrangelo MJ, Ainsworth AM. Current concepts of the biology of human cutaneous malignant melanoma. Cancer Res 1977;242. [14] Elder DE, Guerry D, Clark Jr. WH. Early (radial growth phase) melanomas are biologically benign and share immunohistochemical markers with nevi. In: Proceedings of the International Pigment Cell Society. Tucson: 1986. [15] Friedman RJ, Rigel DS, Kopf AW, et al. Favorable prognosis for malignant melanomas associated with acquired melanocytic nevi. Arch Dermatol 1983;119:455. [16] Sober AJ, Day CL, Fitzpatrick TB, et al. Early death
Summary It is clear that much of what has been taught over the years concerning the pathology of melanoma may have little validity. Melanoma is viewed simply as a malignant neoplasm comprised initially of a proliferation of atypical melanocytes within the surface epithelium (epidermis). It has many features in common, regardless of anatomic site. It spreads within the epidermis first for months, possibly years or even for decades. At this stage (melanoma in situ) it is wholly curable if completely surgically excised. What determines how long a given melanoma remains in situ is not clear. It is probably a combination of factors, including host response to the neoplasm; physical barriers to growth and metastasis (perhaps solar damage); chemical or humoral growth factors or inhibitors (perhaps genetically determined); and other as yet undiscovered factors. Once a given neoplasm penetrates into the subjacent dermis, there are whole ranges of ill-defined events that act on its ability to continue to grow and develop the competence for metastasis (growth factors and inhibitors, neoangiogenesis factors and inhibitors, host immune responses, and so forth). Let us throw out all of our prejudices that may have developed or nurtured over the years. There is much to learn about the pathobiology of melanoma.
676
R.J. Friedman, E.R. Heilman / Dermatol Clin 20 (2002) 659676 from clinical stage I melanoma. J Invest Dermatol 1983;80(suppl 6):505. Clark Jr. WH, Elder DE, Guerry D, et al. A study of tumor progression: the precursor lesions of superficial spreading and nodular melanomas. Hum Pathol 1984; 15:1147. Herlyn M, Thurin J, Balaban G, et al. Characteristics of cultured human melanocytes isolated from different stages of tumor progression. Cancer Res 1985;45:5670. Herlyn M, Thurin J, Balaban G, et al. Cell biology of melanoma and non-melanoma melanocytic lesions. In: Elder DE, editor. Pathobiology of malignant melanoma. Basel: Karger; 1987. p. 166. Heenan PJ, Holman CDJ. Modular malignant melanoma: a distinct entity or a common end stage? Am J Dermatopathol 1982;4:477. Conley J, Lattes R, Orr W. Desmoplastic malignant melanoma (a rare variant of spindle cell melanoma). Cancer 1971;28:914. Breslow A. Thickness, cross-sectional area and depth of invasion in the prognosis of cutaneous melanoma. Ann Surg 1970;179:902. Schmoekel C, Nejad KK, Braun-Falco O. Malignant melanomas with high and low risk. In: Ackerman AB, editor. Pathology of malignant melanoma. New York: Masson; 1982. p. 321. Rigel DS, Friedman RJ. Malignant melanoma. In: Stone J, editor. Dermatologic allergy and immunology. St Louis: CV Mosby; 1984. Day Jr. CL, Harrist TJ, Gorstein F, et al. Malignant melanoma: prognostic significance of microscopic satellites in the reticular dermis and subcutaneous fat. Ann Surg 1981;194:108. Lopransi S, Mihm Jr. MD. Clinical and pathologic correlation of malignant melanoma. J Cutan Pathol 1979;6:180. Friedman RJ, Rigel DS. The clinical features of malignant melanoma. Dermatol Clin 1985;3:271. Kamino H, Ackerman AB. Malignant melanoma in situ: the evolution of malignant melanoma within the epidermis. In: Ackerman AB, editor. Pathology of malignant melanoma. New York: Mason Publishing; 1981. p. 59. Maize JC, Ackerman AB. Malignant melanoma. In: Maize JC, Ackerman AB, editors. Pigmented lesions of the skin. Philadelphia: Lea and Febiger; 1987. p. 165.
[17]
[24]
[18]
[25]
[19]
[26]
[20]
[27] [28]
[21]
[22]
[29]
[23]
Dermatol Clin 20 (2002) 677 680
Biopsy techniques Diagnosis of melanoma

Neil A. Swanson, MD*, Ken K. Lee, MD, Annalisa Gorman, MD, Han N. Lee, MD
Department of Dermatology, Oregon Health & Science University, 3181 SW Sam Jackson Park Road, Mailcode OP06, Portland, OR 97201,USA
A properly performed biopsy is the first step in the management of melanoma. It has become increasingly important to obtain biopsies of pigmented lesions correctly because there are significant ramifications beyond just making the diagnosis of melanoma. Factors, such as depth of invasion, ulceration, microsatellitosis, angiolymphatic invasion, and mitotic index, can impact management and prognosis. The decision to implement new techniques, such as sentinel lymph node biopsy and new adjuvant therapies, is often determined by the initial biopsy. It is critical that an adequate specimen be presented to the dermatopathologist so that a correct and complete diagnosis can be made. Proper biopsy technique impacts the diagnosis and treatment of melanoma, but it is also important in creating the best aesthetic results, because many biopsies are benign. This article reviews the decision-making process and discusses in detail the biopsy techniques and rationales for their use.
Decision-making process When examining a patient with one or several suspicious pigmented lesions, the question often arises, Do I need to perform a biopsy, and if so, which technique do I choose? There are specific characteristics of patients and individual lesions that portend higher risk to develop melanoma. Examples
* Corresponding author. E-mail address: andertam@ohsu.edu (N.A. Swanson).
include patients with any of the following: a personal or family history of melanoma; a fair complexion; the presence of multiple nevi or atypical (dysplastic) nevus syndrome; a history of numerous or severe sunburns; and an advanced aged. Risk factors for individual lesions include appearance de novo; a change in size, texture, or shape; the ABCD criteria; and the symptom of pruritus. Clinically asymptomatic nevi that begin to itch should alert both patient and physician to pay closer attention to that particular lesion. A clinician must also listen to the patient. If they sense that there is something changing or different in a particular lesion, it is often best to obtain a biopsy of that lesion. Experienced clinicians have many anecdotes of clinically benign lesions removed purely based on patient request that turn out to be melanoma. Lastly, a clinician must weigh the given risks of a particular lesion with the patient or family concern (in the case of children) for scarring, inherent in all biopsy procedures. There are several tools available that can help to decide whether or not to obtain a biopsy of a particular pigmented lesion. These include dermoscopy, precise photography, mole mapping by computer, and others in developmental stages. Dermatologists develop expertise to determine which group of patients and which particular lesions are concerns for the development of melanoma. If suspicion is moderate or high, dermatologists routinely remove the pigmented lesion and submit it for histologic analysis by a dermatopathologist. This is both reassuring to the patient and clinician, and frequently leads to a diagnosis of melanoma in its earlier, less advanced stage.
678
N.A. Swanson et al. / Dermatol Clin 20 (2002) 677680
Biopsy techniques Glossary: definition of terms An excisional biopsy refers to en toto removal of a suspicious lesion. This is performed with a margin (as defined later) of clinically normal tissue. It is the preferred method of removing a pigmented lesion for histologic interpretation. An incisional biopsy is used to sample only a part of a suspicious lesion for histologic evaluation. Because it does not remove the entire lesion, incisional biopsies may limit the dermatopathologists ability to detect a melanoma based on sampling error. Shave biopsy refers to a shallow removal of a lesion at a depth confined to the dermis. It can be performed by a scalpel, a dermablade, a razor blade, or scissors. Hemostasis is usually obtained with aluminum chloride, and the biopsy site is allowed to heal by secondary intention. A saucerization is a biopsy that occurs through viable dermis into subcutaneous fat. It is performed by angling a scalpel at approximately 45 degrees to the skin and removing a disk of tissue, including all or part of the suspicious lesion, well into the subcutaneous fat. Punch biopsy refers to the use of a sharp circular instrument to remove tissue well into the subcutaneous fat. Punches are available in sizes ranging from 1.5 mm to 1 cm. They are used for excisional or incisional biopsies. Punch biopsies, with rare exception, are closed with simple interrupted or vertical mattress stitches for wound edge eversion to obtain the best cosmetic result. The fusiform (ellipse) allows for full-thickness removal of the suspicious lesion and a margin of surrounding skin. Closure is obtained with both deep and cuticular stitches. The margin removed is defined as the area of normal-appearing tissue surrounding the lesion to be removed and has two components. The peripheral margin is the area of normal skin extending radially from the clinically suspicious lesion, whereas the deep margin is the depth to which skin and subcutaneous tissue are entered and removed during the biopsy. Excisional biopsy The excisional biopsy is the preferred method of removing a clinically suspicious pigmented lesion. It provides the dermatopathologist with the maximal opportunity to diagnose a melanoma in a given biopsy sample. If present, the maximum depth of invasion of the melanoma can be measured and other histologic criteria of importance. It also helps define
the extent of further surgery necessary based principally on the depth of the melanoma. When performing an excisional biopsy, the clinician should first examine the lymph nodes in the suspected draining basin from the lesion in question. If the lesion turns out to be a melanoma, inflammation from the biopsy procedure can result in a false-positive dermatopathic node. Clinicians often use a Woods lamp to help assess the clinical margin, sometimes extending beyond the obvious clinical appearance of the lesion. The authors take a 1- to 1.5-mm margin surrounding the lesion so defined into the subcutaneous fat [1,2]. When possible, the biopsy is aligned along relaxed skin tension lines and along the draining lymphatics from that site. The former allows for an easier and more cosmetically acceptable procedure if the lesion is a melanoma and requires reexcision. The latter allows for a more accurate sentinel lymph node biopsy if indicated. The excisional biopsy can be performed with either a punch or a fusiform (elliptical) excision (Fig. 1). By choosing a punch 1- to 1.5-mm greater
Fig. 1. (A) Fusiform excisional biopsy outlined with a 1- to 1.5-mm peripheral margin around a suspicious pigmented lesion. (B) Tissue from the same excisional biopsy illustrating the deep margin into subcutaneous fat.
679
in diameter than the lesion to undergo biopsy, a fullthickness, complete specimen can be removed easily and quickly, with a cosmetically acceptable scar. If the lesion is larger or in a cosmetically sensitive area, such as the head and neck, a fusiform (elliptical) excision with full-thickness closure provides adequate tissue for the pathologist and an excellent cosmetic result if the pigmented lesion is benign. Saucerization A saucerization biopsy can be performed to provide an excisional specimen for the dermatopathologist (Fig. 2). This technique is easy to perform, time effective, and often preferred by patients and physicians in areas of the body where it is difficult to create an elegant scar. These areas include the upper back, shoulders, upper arms, and anterior chest. The authors often use a saucerization technique on lower extremities and ears. The key is to make sure this is a saucerization (ie, a biopsy into fatty tissue). Because in areas such as those described scars often spread or become hypertrophic, a saucerization is often the biopsy method of choice. It leaves a smaller, round, cosmetically acceptable scar instead of a longer linear spread excisional scar, providing adequate tissue for histologic interpretation (see Fig. 2). Incisional biopsy An incisional biopsy can be used to diagnose melanoma in a worrisome pigmented lesion. Because it does not sample the entire lesion in question, however, incisional biopsies are not as accurate diagnostically or prognostically as their excisional coun-
Fig. 3. This figure illustrates two points. First, this is the type of large pigmented lesion in a cosmetically sensitive area where an incisional biopsy might be performed. The authors recommend two to three 3- to 4-mm punch biopsies from the darkest area or a long, thin ellipse from the same. If the latter, the authors ask that the pathologist section the specimen longitudinally instead of the typical bread loaf. Second, the dotted margin is the clinical margin pre-Woods lamp, and the solid margin after Woods lamp. This tool can be very useful to determine an accurate clinical margin.
terpart. It should be performed on very large lesions and in cosmetically sensitive regions (Fig. 3). If it is performed, the most darkly pigmented or raised area of the lesion should undergo biopsy. This can be performed using a punch biopsy; small fusiform excision (ellipse); or a saucerization down to the level of the subcutaneous fat. Although most clinicians believe that an incisional biopsy does not spread tumor or influence survival [3], there still exists a theoretical risk that cutting through the tumor may lead to local spread of the melanoma. Studies indicate that an incisional biopsy does not influence prognosis [4 7]. Others have shown that an incisional biopsy, especially of a deeper melanoma, may negatively influence local recurrence or survival [8,9]. When performed, the tissue specimen should be cut and processed along the longitudinal axis to increase the cross-section area available for the dermatopathologist histologically.
Special circumstances For the clinically obvious melanoma, often in patients with a prior personal or family history of melanoma, the authors choose an excision with appropriate margins (both peripheral and deep) as the initial biopsy technique. This is used almost exclusively on the trunk or extremities. The margins chosen are similar to that for a thin melanoma, which is 1 cm
Fig. 2. A saucerization excisional biopsy with 1.5-mm margins surrounding a suspicious pigmented lesion, with a depth to fat. A duoderm dressing is in place and the wound is left to granulate pending pathologic findings.
680
metic results, bearing in mind that the excisional technique is ideal because it removes the suspicious lesion en toto. Excisional biopsies should extend to the subcutaneous fat by means of a punch biopsy, a fusiform ellipse, or a saucerization. Incisional biopsies can be performed in certain circumstances, but should be done so with caution because sampling error may lead to missed diagnosis or inaccurate histologic criterion, such as depth.
References
Fig. 4. A clinically highly suspicious pigmented lesion can undergo biopsy with a 1-cm margin, as in this case. If the melanoma is confirmed and the depth is less than 1 mm, the patient has been treated adequately by the excisional biopsy. If deeper, the final margin during re-excision can be adjusted to reflect the biopsy margin. [1] Brown MD, Johnson TM, Swanson NA. Changing trends in melanoma treatment and the expanding role of the dermatologist. Dermatol Clin 1991;9:657 67. [2] Holmstrom H. Surgical management of primary melanoma. Semin Surg Oncol 1992;8:366 9. [3] Penneys NS. Excision of melanoma after initial biopsy. J Am Acad Dermatol 1985;13:995 8. [4] Eldh J. Excisional biopsy and delayed wide excision versus primary wide excision of malignant melanoma. Scand J Plast Reconstr Surg 1979;13:341 5. [5] Drzewiecki KT, Ladefoged C, Christensen HE. Biopsy and prognosis for cutaneous malignant melanomas in clinical stage I. Scand J Plast Reconstr Surg 1980;14: 141 4. [6] Lederman JS, Sober AJ. Does wide excision as the initial diagnostic procedure improve prognosis in patients with cutaneous melanoma? J Dermatol Surg Oncol 1986;12:697 9. [7] Epstein E, Bragg K, Linden G. Biopsy and prognosis of malignant melanoma. JAMA 1969;208:1369 71. [8] Lees VC, Briggs JC. Effect of initial biopsy procedure on prognosis in stage 1 invasive cutaneous malignant melanoma: review of 1086 patients. Br J Surg 1991;78: 1108 10. [9] Lederman JS, Sober AJ. Does biopsy type influence survival in clinical stage I cutaneous melanoma? J Am Acad Dermatol 1985;13:983 7.
peripherally and well into the subcutaneous fat (Fig. 4). If the diagnosis is melanoma and the maximum lesional depth is less than 1 mm, the authors have treated the patient at the same time the biopsy procedure has been performed.
Summary The biopsy of a suspicious pigmented lesion is critical to establishing a correct and complete diagnosis. It allows the dermatopathologist accurately to diagnose melanoma and to gauge maximum depth of invasion (and other histologic criterion). This, in turn, influences the extent of further necessary surgery or other adjuvant therapy. Furthermore, choosing the appropriate biopsy technique provides adequate cos-
Dermatol Clin 20 (2002) 681 699
Surgical approaches to malignant melanoma Practical guidelines

Richard L. Shapiro, MD, FACS
Division of Surgical Oncology, Department of Surgery, New York University School of Medicine, 530 First Avenue, Suite 7G, New York, NY 10016, USA
The incidence of cutaneous malignant melanoma has been rising steadily over the last century world wide. In the United States, the lifetime risk of developing melanoma is estimated to be 1 in 75 in comparison with 1 in 1500 in 1935 [1 3]. Melanoma is currently the sixth and seventh most common solid malignancy diagnosed in men and women, respectively [4]. In the year 2002, it is estimated that 53,600 Americans will be diagnosed with cutaneous melanoma and 7400 will die of their disease [5]. Although accounting for only 5% of all skin cancers diagnosed annually, over 75% of all skin cancer deaths are caused by melanoma. Efforts during the past decade to educate primary care physicians [6,7] and the general public about the characteristic early cutaneous manifestations of melanoma have resulted in most (82%) patients being diagnosed in the early stages, when the primary tumor is confined to the skin [8]. Early diagnosis correlates significantly with increased survival. Five-year survival rates exceeding 90% are achieved in patients with localized disease in comparison with rates of approximately 60% and 5% in those with regional lymph node and distant metastases, respectively [9].
Epidemiology and risk factors Melanoma is diagnosed slightly more often in men than in women and occurs within a broad range of ages beginning in the third decade of life. Melanoma
E-mail address: richard.shapiro@med.nyu.edu
most commonly arises on the skin of the back in men and on the lower extremities in women. In darkerskinned ethnic groups (African-American, Asian, and Hispanic), melanoma frequently arises in the volar and plantar skin (acral) or in the nail bed (subungual). Although over 95% of melanomas originate in the skin, these tumors also arise in other anatomic locations, such as the eye, and mucous membranes including the vagina and anus [10]. Approximately 3% to 10% of patients present with metastatic disease in the absence of a clinically demonstrated primary lesion [11,12]. Patients with metastases from an unknown or occult primary melanoma have the same prognosis and should be managed in the same manner as patients with known primary lesions [13,14]. The typical patient with melanoma has lightcolored eyes, reddish or blonde hair, and a fair complexion that tans poorly and burns easily during brief periods of intense sun exposure [15,16]. Blistering sunburns sustained as a child or teenager are a significant risk factor for the development of melanoma and seem to be more deleterious than prolonged exposure to the sun in the later years [17]. The development of atypical (dysplastic) nevi is another significant risk factor. Although patients with greater than 100 dysplastic nevi (atypical mole syndrome) have a 10% to 15% incidence of melanoma in their lifetime, the presence of even a single atypical mole also increases the risk [18 21]. The prophylactic excision of all dysplastic nevi is not justified, however, because in most patients melanoma arises from normal skin as opposed to from within a preexisting nevus [22 25]. Complete excisional biopsy, however, of any pre-existing mole that has changed in appearance, itches, or bleeds is recommended. Chil-
682
R.L. Shapiro / Dermatol Clin 20 (2002) 681699
dren with large congenital nevi are also at increased risk for melanoma, although this occurs uncommonly (10% to 15%), and occurs most often in patients with truncal lesions within the first decade of life [26 28]. Any patient who has been diagnosed previously with a melanoma has a 3% to 6% risk of developing a second primary cutaneous melanoma in the course of their lifetime [29,30]. In addition, there is a 3% to 10% incidence of melanoma in first-degree family members [31 33]. Lifelong total cutaneous surveillance and follow-up of all melanoma patients and their immediate families is strongly recommended.
Diagnosis and proper biopsy technique A thorough physical examination includes an inspection of the entire cutaneous surface (including the volar and plantar skin, webspaces, and mucous membranes) of a completely undressed patient under appropriate lighting conditions. The ABCD rule helps to identify those pigmented lesions most likely to be melanoma [34 36]. Lesions that are asymmetric, with irregular borders, color variation, and diameters exceeding 6 mm (the size of a pencil eraser) are considered to be suspicious. In addition, any pigmented lesion that becomes darker or lighter in color, increases in size or becomes raised, or itches or bleeds should immediately arouse suspicion. Complete surgical excision is the biopsy method of choice for all cutaneous lesions suspected of being melanoma [37]. This technique can be performed rapidly in the office setting using local anesthesia and enables an accurate assessment of tumor thickness should the lesion indeed prove to be melanoma. Punch and shave biopsies are sometimes performed but may be suboptimal because tumor thickness may be underestimated. Following a diagnosis of melanoma by such methods, complete excisional biopsy should then be performed to assess tumor thickness
more accurately so that definitive surgery can be planned appropriately. Incisional biopsy of unusually large or broad lesions is acceptable only if a diagnosis of melanoma is confirmed and tumor thickness is assessed accurately. For extremity lesions, surgical biopsy incisions should always be oriented vertically as not to interfere with or complicate subsequent definitive excision and reconstruction. Immunohistochemistry is used routinely to compliment standard histopathologic techniques in confirming the diagnosis of melanoma. The monoclonal antibody HMB-45 [38,39] and the polyvalent antibody recognizing the S-100 antigen [40,41] are used most extensively. Although a combination of the two may improve the histopathologic characterization of difficult lesions, neither has proved to be completely reliable. For example, the S-100 protein, although expressed in almost all melanomas, is also detected in other tissues of neural crest derivation. Although HMB-45 staining may be used to distinguish unusual melanomas from unusual benign nevi, it is sometimes not identified in metastatic lesions or in amelanotic and desmoplastic melanomas [42].
Assessment of tumor thickness and staging Tumor thickness is a measure of the vertical growth phase of melanoma and is the most powerful prognostic indicator of the potential for local recurrence, metastases, and death [43]. An accurate assessment of tumor thickness, preferably through complete excisional biopsy, is crucial for the planning of an appropriate metastatic survey, surgery, adjuvant treatment, and follow-up schedule. Melanoma thickness is assessed in millimeters by ocular microscopy as described by Breslow [44] or by increasing levels of dermal penetration in the manner described by Clark et al [45,46]. The microstaging techniques of both Clark and Breslow have enabled a more accurate
Table 1 Melanoma staging and prognosis Stage IA IB IIA IIB III IV Criteria Breslow 0.75 mm or Clark level II Breslow > 0.75 1.5 mm or Clark level III Breslow > 1.5 4 mm or Clark level IV Breslow > 4 mm or Clark level V Regional lymph node metastasis or in-transit metastasis Distant metastasis* TNM T1 N0 M0 T2 N0 M0 T3 N0 M0 T4 N0 M0 Any T, N1 or 2, M0 Any T, any N, M1 % Survival (5 y) 93 87 66 50 40 <10
From American Joint Committee on Cancer. 4th edition. 1992, with permission. * Includes metastasis to skin, subcutaneous tissues, or lymph nodes beyond the regional nodes (M1a) and visceral metastasis (M1b).
683
grouping of patients with primary melanomas based on reproducible measurements of tumor thickness and are predictors of survival. Melanoma is staged according to published American Joint Committee on Cancer guidelines using the TNM system [47] (Table 1). This system incorporates Breslow and Clark levels within the T classification. When Breslow and Clark levels are in discordance, the thicker assessment predominates. Recently, a revision in the staging system for patients with melanoma has been proposed, which further stratifies groups of patients by including important prognostic factors, such as ulceration of the primary lesion and delineating the extent of nodal involvement by recording microscopic nodal involvement of the sentinel node and the number of nodes affected [48,49]. Other parameters, such as lactate dehydrogenase (LDH) level and anatomic site of distant metastatic disease, are also included in this new staging system. Patients seeking an opinion after lesion excision at another institution should be encouraged to submit all
outside reports and biopsy slides for in-house review to confirm the diagnosis of melanoma and tumor thickness before planning definitive surgery.
Preoperative metastatic work-up Metastatic work-up (Table 2) is determined primarily by the thickness of the primary melanoma and disease stage at presentation. No metastatic work-up is required for patients with malignant melanoma in situ. Patients with thin melanomas (< 1 mm) undergo chest radiograph and routine blood chemistries (including LDH), although the risk of distant metastases is minimal (< 3%). Patients with intermediate-thickness lesions (1 to 4 mm) undergo chest radiograph; routine blood chemistries; CT scans with intravenous contrast of the chest, abdomen, and pelvis; and CT or MRI scans of the brain. Patients with thick lesions (> 4 mm) have a significant risk of metastases to the regional nodes (60%
Table 2 Surgical treatment of primary cutaneous malignant melanoma (clinically NEGATIVE regional lymph nodesa ) Melanoma thickness (mm) In situ < 1 1 3.9 Pre-operative work-upb CXR; LDH CXR; LDH CT (chest, abdomin, pelvis) CT/MRI (brain) FDG-PET e CXR; LDH CT (chest, abdomin, pelvis) CT/MRI (brain) FDG-PET e Excision margins 5 mm 1 cm 2 cm Rx Regional nodes
Sentinel lymphadenectomyd
! 4
! 2 cmc
Sentinel lymphadenectomyd
Abbreviations: AJCC, American Joint Committee on Cancer; CNS, central nervous system; CXR, chest radiograph; FDG-PET, fluorodeoxyglucose positron emission tomography; FNA, fine needle aspiration; LDH, lactate dehydrogenase. a Patients with biopsy proved (FNA) regional lymph node metastases (AJCC stage III) undergo formal (complete) lymph node dissection at the time of wide and deep excision of the primary lesion in the absence of significant distant metastases. Patients presenting with biopsy proved distant metastases (AJCC stage IV) undergo wide and deep excision of the primary lesion without sentinel lymphadenectomy. Patients presenting with palpable regional lymph node metastases and distant metastases are candidates for palliative formal lymph node dissection only in the setting of minimal stage IV disease. b CT scans are performed with intravenous contrast in patients with melanoma; MRI scanning is more sensitive in the detection of CNS metastases; MRI or FDG-PET scanning may be useful in cases in which CT scan findings are indeterminate. c Minimal excision margins of 2 cm advised for melanomas ! 4 mm in thickness. Wider (3 cm) margins are often performed for these thick lesions, although no prospectively randomized data support this practice. d Intraoperative lymphatic mapping and sentinel lymphadenectomy are performed in the absence of confirmed palpable lymph node metastases or evidence of distant disease. Formal (complete) lymph node dissection is performed immediately if intraoperative microscopic analysis (touch preparation, frozen section, rapid immunostain) of the sentinel lymph node(s) reveals evidence of metastatic melanoma. If sentinel node metastases are confirmed later on final pathology, formal (complete) lymphadenectomy is performed as a second procedure. e FDG-PET scanning may be indicated to assess further patients with indeterminate radiologic findings or those at high risk for metastatic disease.
684
to 75%) and distant sites (30% 50%) and should undergo CT or MRI scanning as described for intermediate-thickness patients. The degree of suspicion with which ambiguous radiologic findings are viewed is dependent on the risk for metastases as assessed by physical examination, melanoma thickness, and disease stage. Indeterminate findings suspicious for metastases on CT scan may be investigated further using more advanced imaging techniques, such as MRI or 2-[18F]-fluoro-2-deoxyD-glucose (FDG) positron emission tomography (PET) scanning. PET using FDG may be particularly useful in the evaluation of those patients with more advanced (stage III and IV) or recurrent melanoma [50 52]. Cytologic confirmation of metastatic disease is essential and may be accomplished with needle aspiration biopsy performed under radiologic (CT or ultrasound) guidance.
evidence that local recurrence reflects the biologic aggressiveness of the primary lesion, not the inadequacy of the primary excision. Excision sites are most often closed primarily, although split- or full-thickness skin grafting or flap closure may be required for the reconstruction of larger defects. In patients requiring skin grafting for closure of a wound located on the extremity, skin should not be harvested from that limb, which could potentially reintroduce melanoma cells into the wound. Similarly, the changing of gloves and surgical instruments should be performed routinely after melanoma excision to avoid wound contamination.
Treatment of the regional lymph nodes Clinically negative regional lymph nodes Most patients with newly diagnosed melanoma exhibit no evidence of regional lymph node metastases. The potential for regional lymph node metastases is assessed most accurately by tumor thickness. Malignant melanoma in situ by definition has no significant potential for lymph node metastases. For thin melanomas ( < 1 mm) the risk of regional lymph node metastases is minimal ( < 5%); treatment of the regional lymph nodes is also not indicated. Patients with intermediate-thickness lesions (1 to 4 mm) have a 20% to 25% incidence of microscopic regional disease and a 3% to 5% risk of distant metastases and should theoretically derive the greatest therapeutic benefit from elective lymph node dissection. Prospectively randomized studies, however, have consistently failed to demonstrate a significant survival advantage with elective lymph node dissection [69 77]. However, the excision of clinically negative, microscopically positive lymph nodes in patients with melanoma serves but to identify those patients at high risk for systemic disease who may be candidates for adjuvant treatment. Sentinel lymphadenectomy [78 81] provides an alternative to routine elective lymph node dissection in patients who are at risk for subclinical micrometastases but have no clinical evidence of regional nodal disease. Cutaneous lymphoscintigraphy allows for a more precise delineation of the primary lymphatic drainage [82] of cutaneous melanomas and identification of a lymph node (or nodes) in the regional lymph node basin, termed the sentinel node, most likely to contain micometastases [83,84]. Intraoperative lymphatic mapping using a combination of intradermally injected vital blue dye and radiolabeled technetium sulfur colloid facilitates the selective
Excision margins The propensity of melanoma to disseminate and recur locally is well documented and has historically influenced the surgical approach to this tumor [53 56]. The surgical treatment of melanoma in terms of excision margin width has been studied extensively in prospectively randomized trials, however, and has become increasingly more conservative over the past several decades (see Table 2) [57 62]. Malignant melanoma in situ is excised with 5-mm margins. Thin (< 1 mm) melanomas have an extremely low rate of local recurrence ( < 2%) and are excised with 1-cm margins [63 65]. For intermediate-thickness (1 to 4 mm) lesions, the safety of 2-cm excision margins has been confirmed in prospectively randomized studies [66]. For thick melanomas (> 4 mm) the guidelines are less certain because no prospectively randomized study has specifically addressed the issue of excision margins for these lesions. Whereas 2-cm margins may be appropriate for these lesions based on currently available data, in selected circumstances (very thick lesions or in the presence of multiple satellite metastases) many surgeons advocate wider excision margins of 3 cm. It is unlikely, however, that margins exceeding 2 cm significantly impact on the higher rates of local recurrence (12%) and poor survival (55% at 5 years) with which these lesions are associated [67]. The failure of more radical procedures using wider (3 to 5 cm) excision margins or, in the past, limb amputation [68] to diminish local recurrence rates and increase survival in patients with thick (>4 mm) melanomas is further
685
Fig. 1. Details of cutaneous lymphatic mapping and the sentinel lymphadenectomy procedure. (A) Preoperative cutaneous lymphoscintigraphy in a patient with a malignant melanoma of the right chest wall. Approximately 800 mCi of filtered radiolabeled technetium 99-m sulfur colloid is injected intradermally around the site of the melanoma (large bright spot). The sentinel node is identified in the right axillary region (smaller, lateral bright spot). (B) Isosulfan blue dye (1 mL of 1% aqueous solution) is injected intradermally around the site of the melanoma. Surgical margins of 2 cm have been selected for the resection of this 1.5-mm thick melanoma. An elliptical excision allows for primary closure. (C) The blue-stained sentinel node and an afferent lymphatic are revealed through a small right axillary incision. A hand-held gamma detector probe (in the yellow plastic sheath) aids in the localization of the sentinel node. Once blue-stained lymphatic tissue is visualized by the surgeon, the probe is used to confirm the gamma signal in the blue-stained sentinel node.
identification and excision of the sentinel lymph nodes (Fig. 1) [85 87]. Rapid histopathologic staging is performed intraoperatively by microscopic examination of the sentinel node. If no microscopic evidence of metastatic melanoma is noted on histopathologic examination of the sentinel node, no other lymph nodes need to be excised because the incidence of so-called skip metastases to any other node in that particular regional drainage basin is less than 5% [88,89]. Immediate therapeutic dissection of the regional nodes is performed if metastases are noted in the sentinel lymph node. A large multi-institutional randomized study (Multicenter Selective Lymphadenectomy Trial) has been designed carefully to confirm the accuracy of this technique and the hypothesis that regional lymph node metastases occur rarely in the absence of metastasis to the sentinel node [90,91]. In patients with melanomas greater than or equal to 1 mm thick and no clinical evidence of regional
lymph node metastases, lymphoscintigraphy consisting of dynamic images and spot views is performed before surgery to define lymphatic drainage accurately and demonstrate the location of the sentinel lymph nodes. Sentinel lymphadenectomy is performed most accurately at the time of wide and deep excision of the primary melanoma. Consequently, patients who have already undergone definitive wide and deep excision of the primary lesion may not be ideal candidates for sentinel lymphadenectomy because lymphatic flow and drainage patterns may have been altered by that prior surgery. Several hours before surgery, the site of the primary lesion is injected with approximately 800 mCi of technetium 99m filtered sulfur colloid in four divided doses (see Fig. 1A). After the induction of anesthesia, 2 mL of vital blue (1% isosulfan blue) is injected intradermally at the site of the melanoma to stain the afferent lymphatics and sentinel node blue (see
686
Fig. 1B). The site is massaged manually for 20 minutes before incision in the skin overlying the sentinel lymph node as identified preoperatively by lymphoscintigraphy and confirmed intraoperatively with a hand-held gamma detector. Intraoperatively, the sentinel lymph node is identified visually by the appearance of blue stain in the node or in the afferent lymphatics and by detecting an enhanced gamma signal, compared with background levels, with the hand-held probe (see Fig. 1C). After the sentinel lymph node has been removed from the field, no other evidence of blue dye should be observed, nor should there be any significant residual radioactivity (exceeding background) detected with the gamma probe. The sentinel node is taken immediately to surgical pathology for rapid microscopic examination, which may include touch preparation, frozen section analysis, or rapid immunostains. If no definitive evidence of metastatic melanoma is noted, wide and deep excision is performed as planned and the procedure is terminated. If micrometastases are confirmed in the sentinel lymph node, then formal (therapeutic) lymphadenectomy is performed at that time. Postoperatively, the sentinel lymph node is serially sectioned and examined in detail using hematoxylin and eosin staining and S-100 and HMB-45 immunostaining. The use and clinical relevance of staging the sentinel node further on a molecular level using such techniques as polymerase chain reaction is currently being evaluated [92 95]. Patients with confirmed micrometastases in the sentinel node then routinely undergo complete regional lymph node dissection as a second procedure. Patients presenting with thick primary melanomas (> 4 mm) and no clinical evidence of regional lymph node metastases or radiologic evidence of distant disease are still at high (50% to 75%) risk to have microscopic nodal metastases and are also appropriate candidates for intraoperative lymphatic mapping and sentinel lymphadenectomy. Clinically positive lymph node metastases Any palpable lymph node in a patient with melanoma should be considered indicative of metastasis until proved otherwise. Fine needle aspiration biopsy is an accurate, reliable method of confirming metastatic melanoma [96 98]. If fine needle aspiration is not available or results are indeterminate, excisional biopsy of the lymph node is performed. Patients presenting with or subsequently developing regional lymph node metastases are at high risk for distant metastases and should undergo advanced imaging including CT scanning with intravenous contrast of
the chest, abdomen, and pelvis and CT or MRI scanning of the brain. FDG-PET scanning is also recommended to evaluate patients further with advanced melanoma. In patients with cytologically or histologically proved regional nodal metastases, formal (complete) lymph node dissection is performed. The development of palpable lymph node metastases is correlated significantly with substantially diminished survival (10% to 50%), which is influenced strongly by the number of and the extent to which the lymph nodes are involved and the primary melanoma thickness [99,100]. Regional lymph node dissection should not be performed routinely in patients with documented distant metastases that are extensive or in those patients with large lymph node metastases fixed to adjacent structures. Significant palliation of inoperable bulky or bleeding regional nodal metastases may be achieved with radiation therapy in such situations, which are associated with an extremely poor prognosis.
Adjuvant immunotherapy The rationale for using immunotherapy to treat patients with melanoma is based in part on the observation that evidence of partial regression is seen histologically in up to 20% of primary melanomas. Patients at high risk for distant metastasis, such as those with thick primary melanomas greater than 4 mm (stage IIB) or regional lymph node metastases (stage III), are currently offered immunotherapy in the form of interferon alfa-2 or an investigational melanoma vaccine. The ability of intravenous highdose interferon alfa-2 therapy to extend disease-free and overall survival in patients with stage IIB and III melanoma has been evaluated by the Eastern Cooperative Oncology Group in prospectively randomized trials [101,102]. In the high-dosage regimen interferon is administered intravenously 5 days a week for 1 month at 20 million units per square meter body surface area and then 3 times a week for 11 months by self-administered subcutaneous injection at a dose of 10 million units per meter [2]. Despite a modest survival benefit, however, treatment-limiting side effects (severe flulike syndrome) occur commonly in most patients and, rarely, can be life threatening (cardiovascular and hepatic toxicity). The efficacy of lower-dose regimens of interferon is currently under investigation by the Eastern Cooperative Oncology Group in clinical trials [103]. Melanoma antigen vaccines may alternatively be offered to this cohort of patients. A variety of vaccine preparations are available through university-based immunotherapy
687
programs who report encouraging, although as yet unproved, clinical results [104,105]. A distinct advantage of vaccine treatment, however, is the relative absence of significant side effects.
Postoperative follow-up Lesion thickness and disease stage influence the follow-up schedule for any patient with melanoma. For patients undergoing wide excision of malignant melanoma in situ, routine dermatologic surveillance is sufficient. For patients with thin melanomas (< 1 mm), follow-up consists of an inspection of the total cutaneous surface and a complete physical examination with attention to the surgical excision site and regional lymph nodes, and is performed at 3- to 6-month intervals for 2 years and at 6- to 12-month intervals thereafter. Chest radiograph and routine blood chemistries are obtained on a yearly basis. Patients with intermediate-thickness lesions return at 3-month intervals for 2 years and at 6-month intervals thereafter. Chest radiograph and LDH levels are performed each year, as are CT scans for 2 to 5 years after surgery. FDG-PET scanning is another modality that may have a role in screening high-risk patients for metastatic disease. Lifetime surveillance is essential, however, because the risk of metastatic disease extends over the full lifetime of the patient with melanoma. Most patients with thick primary lesions (> 4 mm) and those with lymph node metastases ultimately develop systemic disease. In these high-risk patients, followup physical examination occurs at 3- to 6-month intervals for life with CT scanning performed on at least a yearly basis. The risk of developing a second primary melanoma necessitates that all patients, irrespective of lesion thickness, continue lifelong follow-up with their dermatologist for total cutaneous examination, as should all of their first-degree relatives.
randomized trials confirm that local recurrence, and the incidence of in-transit, regional lymph node, and distant metastases, rises significantly ( P < .001) with increasing primary melanoma thickness. Local recurrence most often appears clinically as a blue-tinged subcutaneous nodule arising in close proximity (within 2 to 5 cm) to an excision site of a primary melanoma (satellite metastasis) or en route to the regional lymph node basin (in-transit metastasis). Any subcutaneous nodule arising in the vicinity of a melanoma excision site should be considered to be disease recurrence or progression until proved otherwise. Diagnosis is accomplished rapidly and accurately by fine needle aspiration biopsy. Excisional biopsy of the nodule under local anesthesia may sometimes be required for diagnostic confirmation. A complete metastatic survey, including CT or MRI and FDG-PET scans, should then be performed because most of these patients will now also have evidence of systemic metastases. Surgical treatment of locally recurrent malignant melanoma Although no standardized surgical approach to all patients with locally recurrent melanoma has been established, treatment guidelines have been developed based on clinical trials in patients selected by the extent and specific anatomic site of disease recurrence. The realization that local recurrence is not simply the result of inadequate surgical excision but is in fact an outward manifestation of the biologic aggressiveness of the primary melanoma has led to a more rational approach to the treatment of these patients. Complete surgical resection with primary wound closure is the most straightforward means of treating single recurrent lesions. Patients with multiple subcutaneous metastases grouped within a single site can similarly be treated with wide local excision with skin grafting or flap closure as necessary for wound coverage. Although wide resection margins are not as well defined in the resection of locally recurrent disease as they are in the treatment of primary cutaneous melanoma, recurrent lesions should be resected with a margin of normal tissue to avoid tumor spillage and wound contamination. Fracturing of the tumor mass is often followed by further rapid local recurrence. The changing of surgical instruments and gloves immediately after surgical resection or debulking of extensive metastases is recommended because rapid recurrence in the surgical wound and surrounding soft tissues is not uncommon. Despite complete surgical resection of multiple cutaneous metastases, further local and regional
Treatment of local recurrence Local recurrence in a patient with malignant melanoma is an ominous clinical event and is almost always associated with the development of systemic metastases. The survival of these patients is extremely poor averaging less than 5% at 10 years. Primary melanoma thickness remains the most significant prognostic indicator of local recurrence and death, with other important predictive variables being the presence of ulceration and anatomic location of the primary lesion [106]. Large multi-institutional
688
recurrence may occur in up to 67% of patients and is strongly associated with subsequent disease progression [107]. As many as 70% to 82% of such patients ultimately succumb to distant metastases [108]. Although systemic chemotherapy in the adjuvant setting is rarely effective [109], systemic immunotherapy in the form of interferon alfa-2b (Intron A) [110] or a variety of experimental melanoma vaccines may have a role in the treatment of patients subsequent to complete surgical resection of locally recurrent melanoma. The effectiveness of such modes of immunotherapy in improving survival in these patients remains unproved. Role of isolated limb perfusion Isolated limb perfusion was first reported by Creech et al [111] in 1958. The intravascular perfusion with high concentrations of chemotherapeutic agents for extended periods of time of an extremity harboring extensive, recurrent melanoma by isolating the involved extremity from the systemic circulation with tourniquet compression was a major advance for a problem that was often regarded as only amenable to amputation. An extracorporeal circulation was established using a membrane bubble oxygenator combined with a roller pump system to minimize limb ischemia in the isolated extremity. Although a variety of antineoplastic regimens including cisplatin, dacarbazine, actinomycin D, nitrogen mustard, and thiopeta have been evaluated [112], the alkalating agent L-phenylalanine mustard (melphalan) has remained one of the most effective and well-tolerated drugs used in isolated limb perfusions for melanoma [113]. The precise role of hyperthermia (maintaining constant tissue temperature of 38C to 40C) in isolated limb perfusion (isolated hyperthermic limb perfusion) is unclear. The theoretical advantage of hyperthermia in limb perfusion was predicated initially by laboratory studies that suggested that malignant cells may be adversely affected by heat [114]. Despite earlier reports of a significant treatment advantage with the addition of hyperthermia [115], a large multi-institutional study revealed similar complete response rates in patients undergoing perfusion with melphalan under normothermic (54%) or hyperthermic (56%) conditions [116]. The role of hyperthermia in isolated limb perfusion with melphalan was also evaluated in a large retrospective comparative study in which 218 patients treated with mild hyperthermia (39C to 40C) were compared with 166 patients perfused under controlled normothermic conditions (37C to 38C) [117]. The addition of mild hyperthermia did not significantly influence recurrence rates or survival.
The high rates of locoregional recurrence in patients with thick primary lesions and of in-transit and regional metastases in patients developing locally recurrent cutaneous melanoma makes adjuvant isolated limb perfusion an attractive treatment option after surgical resection of all visible tumor. Early retrospective studies comparing excision with excision plus isolated limb perfusion in patients with surgically resected disease yielded conflicting results [118]. A significant improvement in survival in patients with primary melanomas undergoing wide and deep excision combined with perfusion was reported in 1975 with 10-year overall survival rates of 83% compared with 57% in historic controls [119]. Larger case-controlled studies [120], however, have failed to reveal an advantage to isolated limb perfusion for either lowering local recurrence rates or improving survival. Moreover, a more recently performed, large prospectively randomized phase III trial of prophylactic isolation perfusion in patients with high-risk primary lesions ( ! 1.5 mm in thickness) also failed to show an advantage in terms of overall survival despite a modest decrease in intransit recurrence [121]. In this international, multiinstitutional study, 412 patients were randomized to wide excision only and 420 patients to wide excision plus hyperthermic isolated perfusion with melphalan. At a median follow-up interval of 6.4 years, patients undergoing isolated hyperthermic perfusion in addition to wide excision developed significantly fewer in-transit (3.3% versus 6.6%) and regional lymph node (12.6% versus 16.7%) metastases compared with those treated with wide excision alone. There was no significant benefit from isolated hyperthermic perfusion in terms of time to distant metastasis or survival. Based on a lack of convincing data, isolated limb perfusion cannot at this time be recommended as adjuvant treatment of patients with thick primary melanomas or those with completely resected locally recurrence. Extensive, recurrent satellite or in-transit metastases confined to a single extremity are not often amenable to surgical resection and rapidly become a significant clinical management problem despite the rapid development of distant metastases in most of these patients. Isolated limb perfusion using a variety of antineoplastic agents with and without hyperthermia has also been critically evaluated as a therapeutic approach in patients with extensive locoregional recurrence. Lienard et al [122] in The Netherlands reported on the effect of high-dose tumor necrosis factor (TNF) in combination with melphalan and interferon-g in isolated limb perfusions in patients with locoregionally recurrent malignant melanoma
689
and soft tissue sarcoma. Response rates of 91% were reported in patients receiving all three drugs in comparison with 52% when melphalan was used alone. Complete responses were seen in 100% of patients undergoing perfusion for in-transit metastases. After median follow-up of 26 months, 23% of patients have recurred in the perfused limb, 29% at distant sites and 17% at both regional and distant sites. This combination of agents has also been studied prospectively by Fraker et al [123,124] at the National Institute of Health. Patients with locoregionally recurrent melanoma (in-transit metastases) underwent surgical resection of recurrent tumor with or without hyperemic isolated limb perfusion with escalating doses of TNF in conjunction with melphalan and interferon-g. Overall response rates of 80% were reported in patients receiving all three agents compared with 64% in patients treated with melphalan alone ( P = .043). Increasing the dose of TNF from 4 to 6 mg enhanced the response to the perfusion from 92% to 100%, respectively. Systemic toxicity seemed to be related to perfusate leak rather than TNF dose and was easily managed. Regional complications, such as painful myopathy and neuropathy, occurred at higher levels of TNF (6 mg) and were dose limiting. Further studies confirm that more severe limb toxicity is not associated with improved clinical outcomes [125]. Isolated limb perfusion is, however, a complicated surgical procedure requiring special equipment and technical expertise. Although rare, serious complications do occur and are related to the regional and systemic effects of high doses of chemotherapeutic agents and technical aspects of the perfusion procedure [126]. Most patients develop erythema of the skin and mild edema in the extremity within several hours after completion of the procedure and may last for several days. Although common in the immediate postoperative period, significant edema of the extremity occurs in only about 25% of patients [127]. Severe blistering and necrosis of the skin occurs in a minority of patients. Paresthesias persisting for several weeks are common but usually resolve. Longstanding paresthesias and myopathy, however, occur in approximately 5% of patients. Severe myonecrosis requiring major amputation occurs rarely in less than 1% of patients. The long-term morbidity after isolated limb perfusion with melphalan has been studied extensively [128] at two major perfusion centers in The Netherlands. After a minimum of 1 year after a perfusion procedure, some degree of morbidity was observed in 166 (44%) of 367 patients. Lymphedema was the most commonly reported complication occurring in
108 (28%). Regional lymph node dissection, performed in 45% of patients at the time of perfusion, was found to be related significantly to the subsequent development of lymphedema. Other commonly reported long-term effects included pain (8%); muscle atrophy or fibrosis (4%); and recurrent infection (3%). Systemic toxicity occurs from the escape of chemotherapeutic agents into the general circulation. In a retrospective review from The Netherlands by Sonneveld et al [129], systemic effects were observed in 27% of a total of 368 patients undergoing single isolated perfusion procedures with melphalan. These consisted mostly of nausea and vomiting (20%); fever and hair loss (5%); and bone marrow depression (2%). Systemic toxicity of the antineoplastic agents administered is minimized by total isolation of the extremity being perfused through tourniquet compression and by flushing out the limb adequately with heparinized crystalloid at the end of the procedure. The integrity of the perfusion circuit and extracorporeal circulation is monitored by the local administration of a tracing substance that can be detected easily and rapidly in the systemic circulation. In the past, this was accomplished by injecting fluorescein into the perfusion circuit and using a Woods lamp to detect florescence in areas outside of the area of tourniquet compression, such as the sclera or urine. Although small leaks of melphalan and other standard agents into the systemic circulation are generally well tolerated, more quantitative techniques for assessing leak rates are essential when using more biologically active and potentially lethal cytokine agents, such as TNF into the perfusion circuit. Barker et al [130] have addressed this problem through the introduction of iodine 131 labeled albumin or technetium-labeled autologous red blood cells into the perfusion circuit and using a precordial gamma detector to monitor continuously increases in counts to detect quantitatively significant perfusate leaks. Similar techniques using technetium 99m labeled autologous red blood cells as a radioactive tracer have also been developed to monitor leaks continuously and quantitatively during isolated limb perfusion [131]. Intratumoral therapy Direct intralesional therapy using a variety of rin, immunomodulators, such as bacille Calmette-Gue or chemotherapeutic agents, such as cisplatin, has been shown to be effective in patients with multiple recurrent lesions. Intralesional bacille Calmette-Gue rin has been associated with reported response rates
690
of 50% to 65% in injected lesions and up to 21% in uninjected lesions, which are in close proximity to the treated lesions [132,133]. Other agents including methanol extracted residue [134], dinitrocholorobenzene [135], and interferon-a [136] have shown similar responses. Interestingly, intralesional interferon treatment also induced antitumor effects at distant sites among patients with concomitant skin and distant metastases. These unexpected systemic responses were observed in the lung and soft tissues. Complete response was observed in 3 of 51 patients. The duration of complete response was 6, 14, and 20 months. Partial response was noted in six other patients and lasted between 2 and 6 months. More recently an approach has been developed using intratumoral injection of cisplatin-epinephrine gel. The rationale is to deliver high concentrations of active antineoplastic drug (cisplatin) for extended periods at the tumor site, inducing a local effect and avoiding systemic toxicity. Platin concentrations in tumors are 10 to 1000 times higher than in human plasma after intravenous infusions of cisplatin. High drug levels are maintained for prolonged periods of time (> 72 hours), with slow diffusion of the drug from the injection site [137,138]. In 25 patients with a total of 107 clinically troublesome cutaneous, regional lymph node, and soft tissue metastases who had failed one or more previous therapies, intratumoral cisplatin-epinephrine gel treatment resulted in a 44% objective response rate [139]. Systemic toxicity was minimal and managed easily.
Special clinical situations Subungal melanoma Subungal melanoma is a rare clinical entity representing up to 3% of cases of melanoma in whites but a higher proportion of melanomas (15% to 35%) in dark-skinned ethnic groups [140 142]. Over 75% of subungual melanomas involve either the great toe or the thumb. Early signs of this lesion include a darkening of the nail bed. Dark pigmentation of the posterior nail fold, Hutchinsons sign, is a classic sign of subungual melanoma. Early diagnosis of subungual melanoma is rare, with the poor prognosis of this lesion reflecting most significantly advanced stage at diagnosis. Many patients with subungual melanoma report a recent history of trauma to the digit and attribute the lesion to a poorly healing wound. The differential diagnosis of subungual melanoma includes benign pigmented lesions of the nail bed mechanism (melanonychia
striata) [143]; chronic bacterial fungal infection; and subungual hematoma. The major pitfall in the diagnosis of subungual melanoma is an inadequate biopsy. Although nodular and amelanotic lesions do occur, most subungual hematomas appear as a sharply demarcated blueblack to brown discoloration of the nail, which does not involve the adjacent cuticle. A diagnosis of subungual hematoma may be confirmed by releasing the clotted blood through a large bore puncture or partial removal of the nail plate. Close visual inspection of the nail over a short period of time reveals that the pigmentation advances distally with growth of the nail plate. Formal biopsy of the nail bed is performed under digital or regional anesthesia block in the office or outpatient setting. The nail plate is then elevated carefully from the nail bed and removed so that the proximal aspect of the lesion in question is visualized clearly. An elliptical incision in the nail bed down to the underlying periosteum is then performed allowing for complete excisional biopsy of the lesion and primary closure of the defect with fine absorbable sutures. Larger defects may be repaired with nail bed flaps or skin grafting. Generous incisional biopsy through the central portion of pigmentation is performed for larger lesions not amenable to simple excision. Melanoma in situ of the nail bed is treated with wide local excision. Negative surgical margins of at least 5 mm are optimal. The surgical defect may be repaired with a local flap of skin or may require skin grafting. Invasive subungual melanomas of the lower extremity are treated most easily with amputation of the toe. The appropriate surgical resection margin width of 1 or 2 cm for lesions with thickness less than 1 mm or greater than or equal to 1 mm, respectively, is achieved through complete amputation of the affected toe. Ray amputation is performed for lesions extending into the webspace. In most patients, the resulting surgical defect is closed easily, heals well, and allows for ambulation without a specialized prosthesis or orthotic device, even when amputation of the great toe is required. For upper extremity subungual invasive melanomas, surgical treatment is more individualized. Amputation is performed through the joint most proximal to the lesion, which represents a more conservative and functionally superior approach to the more radical amputations performed in the recent past [144]. Wound closure is achieved with a flap of volar tissue while ideally maintaining a margin of at least 1 cm of normal tissue. For subungual melanomas of the thumb, a reconstruction is performed by webspace deepening
691
using a Z-plasty, reducing the length of digit loss by approximately 50% [145]. Sentinel lymph node mapping and excision are appropriately performed for those patients with melanomas greater than or equal to 1 mm in Breslow thickness in the absence of clinically palpable regional nodes. This procedure is most accurate and best performed before amputation of the affected digit. Patients presenting with palpable nodal metastases in the absence of significant distant metastases undergo complete regional lymphadenectomy. Plantar melanoma Melanoma arising on the sole of the foot, characteristically in an acral lentiginous growth pattern, is a rare clinical entity in whites accounting for only 2% to 8% of melanoma cases in white patients [146]. In dark-skinned ethnic groups, however, melanoma arises on the plantar surface of the foot in 35% to 90% of patients diagnosed with melanoma of African-American, Asian, or Hispanic descent [147]. Although the metastatic potential of these lesions is correlated significantly with the thickness of the primary melanoma as it is for cutaneous melanomas arising elsewhere, these lesions are often diagnosed at later stages and therefore generally have a less favorable prognosis. The frequent delay in diagnosis of plantar lesions may be explained in part by their rarity and their unusual and infrequently examined anatomic location. In addition, the increased thickness of the epidermis of the plantar surface may soften and distort the characteristic clinical appearance of the melanoma. Even melanomas that seem to be flat may be revealed after adequate biopsy to be thick lesions. A major pitfall in the early diagnosis of this lesion, however, is the failure to obtain a satisfactory biopsy specimen for histologic confirmation of malignancy. The extreme thickness of the plantar epidermis limits the use of shave biopsy as a diagnostic modality. In addition, the haphazard pigment pattern of these lesions also makes accurate diagnosis and assessment of lesion thickness by other techniques, such as punch or even incisional biopsy, also less likely to be successful. The preferred method of biopsy for these difficult lesions is complete excisional biopsy. Definitive wound closure may be deferred until rapid histologic diagnosis and margin inspection are complete. Once the diagnosis of melanoma is confirmed, the lesion is excised and staged according to guidelines established for other cutaneous primary melanomas of comparable thickness. Sentinel lymph node mapping and excision are recommended for those patients with melanomas greater than or equal to 1 mm in Breslow
thickness in the absence of clinically palpable regional nodes. Patients presenting with palpable nodal metastases and no other evidence of distant metastases undergo superficial inguinal lymph node dissection. Dissection of the deep inguinal nodes is performed in patients with involvement of Cloquets node or extensive disease in the upper aspect of the femoral triangle. Lesions confirmed to be melanomas on shave, punch, or incisional biopsy that approach or exceed 1 mm in thickness may be treated definitively as outlined previously and do not require a preliminary excisional biopsy procedure. Wound closure of the plantar surface requires special consideration. The exact location of the melanoma on the plantar surface, stage of disease, age, associated medical conditions, and lifestyle of the patient must be considered in the determination of wound closure. Defects on non weight-bearing aspects of the plantar surface or those in patients with sedentary lifestyles, significant medical comorbidities, or advanced metastatic disease may be closed most easily primarily or more commonly with split-thickness or full-thickness skin grafts. Closure of defects on the weight-bearing surface of the plantar region in ambulatory patients is accomplished with a variety of flap reconstructive procedures. These include relatively straightforward cutaneous rotational or advancement flaps and more complex reconstructive procedures, such as musculocutaneous free flaps with microvascular anastomosis. These latter procedures are usually performed with a plastic reconstructive surgeon who ideally has been involved in the care of the patient once the diagnosis of melanoma has been confirmed. Melanoma on the face Melanoma occurs rather commonly on the face and most often takes the form of an in situ (Hutchinsons melanotic freckle or lentigo maligna [148]) or thin invasive lesion. Despite their diminished biologic aggressiveness, however, the cosmetic and functional considerations of performing tumor surgery on the face makes treating even these lesions especially challenging. Surgical biopsy should be performed to assess melanoma thickness fully to plan definitive surgical treatment of the primary lesion appropriately and determine the risk of regional lymph node metastases and the need for procedures, such as lymphatic mapping and sentinel lymphadenectomy. As in other anatomic locations, complete excisional biopsy is the procedure of choice for confirming the diagnosis of melanoma and to measure lesion thickness. Other
692
techniques, such as shave, punch, and incisional biopsy, clearly have a role in certain circumstances but are less optimal because melanocytic neoplasms on the face, melanoma in situ in particular, are often far more extensive microscopically in terms of their radial extension than they appear clinically. If complete excisional biopsy of such a large lesion is required to examine fully the lesion and the result would be a large surgical defect not amenable to a cosmetically acceptable primary wound closure, the wound should not be closed at that time and be covered with a moist, sterile dressing until the pathologic examination of the surgical specimen is complete. Care should be taken, however, with any biopsy technique chosen, to avoid injury to the branches of the facial nerve. The marginal mandibular division, because of its superficial location and diminutive size, is particularly at risk. The possibility of facial nerve injury should be discussed with the patient and documented appropriately before any biopsy procedure. All melanomas, ranging from in situ to invasive lesions, should be excised with clear margins. On the face, however, achieving an appropriately wide resection margin may be difficult for melanomas located in close proximity to structures, such as the eye, nose, and mouth. A good rule of practice is to obtain as close to as possible the desired surgical margin based on the thickness of the melanoma. This practice is a reasonable approach in such circumstances because prospective randomized trials designed to define the width of melanoma excision have demonstrated that thinner margins obtained for melanomas greater than 1 mm in thickness are associated with slightly higher local recurrence rates but have no significant impact in long-term survival [149]. Traditional resection margin widths for the excision of in situ and thin invasive melanomas have recently been challenged by investigators advocating the technique of Mohs micrographic surgery as an alternative to wider local excision. Preliminary reports on the use of this technique in the treatment of such melanomas are promising with acceptably low rates of local recurrence [150]. A plastic reconstructive surgeon should be included in the planning and performance of any resection of a melanoma on the face that will result in a significant surgical defect. A range of reconstructive techniques is now widely available to make the final cosmetic and functional surgical result more acceptable. All patients should have the advantage of such a multidisciplinary approach, including the dermatologist, surgical oncologist, and plastic reconstructive surgeon, to their tumor.
All patients with melanomas on the face approaching greater than or equal to 1 mm in thickness are at risk for occult micometastases in the regional lymph nodes. Although for many years it has been appreciated that to some extent the anatomic location of the primary melanoma on the head and neck enables a prediction of the most likely site of regional nodal metastasis [151,152], the advent and refinement of cutaneous lymphoscintigraphy has better delineated the often complex lymphatic drainage patterns unique to this region. Accordingly, elective lymph node dissection of presumed sites of micrometastatic disease in the head and neck region has for the most part been replaced by cutaneous lymphoscintigraphy, using a combination of intradermally injected vital blue dye and a radiolabeled tracer substance, such as technetium sulfur colloid, and sentinel lymphadenectomy [153]. Regional lymphadenectomy is performed selectively in those patients with histologically documented micometastases in the sentinel nodes. Superficial parotidectomy with dissection of all facial nerve branches is recommended for patients with micrometastases in periparotid nodes. Selective cervical lymph node dissection based on the precise location of a positive cervical sentinel node has replaced more traditional modified radical and formal radical neck dissections. Patients presenting with or developing palpable lymph node metastases in the absence of significant distant metastases should undergo formal regional lymphadenectomy as described previously. In a small but not insignificant number of patients with melanoma of the face preoperative cutaneous lymphoscintigraphy reveals a complex pattern of lymphatic drainage from the primary lesion to multiple sentinel nodes widely dispersed throughout the head and neck region. The diagnostic accuracy may decline and risk of facial or spinal accessory nerve injury rises as the complexity and number of individual sentinel nodes to be identified and excised increases. It seems reasonable to forgo sentinel lymphadenectomy in individualized circumstances when multiple sentinel node sites are revealed by preoperative cutaneous lymphoscintigraphy. The reasons for this decision and the risks and benefits of not identifying potential microscopic regional nodal metastases should be discussed fully with the patient and carefully documented. Surgical treatment of cutaneous lesions of uncertain diagnosis Certain lesions pose serious diagnostic challenges, although the histopathologic criteria of melanoma
693
have been well described. In these difficult clinical situations, a second (or third) dermatopathologic opinion is prudent. Whenever possible, additional material should be acquired and a new set of slides prepared from the original cell blocks. Histologic and immunohistochemical staining should also be repeated on the new set of slides obtained. If any of the original lesion of concern remains at the primary site, a complete excisional biopsy should be performed to secure an accurate diagnosis. The treatment of patients with lesions that have not been definitively confirmed to be melanoma or a benign lesion or those lesions generating divergent dermatopathologic diagnoses is difficult but some guidelines may be established. A detailed discussion should be initiated with the patient, family, and referring physicians, which addresses specifically the advantages and disadvantages of treating the lesion as a melanoma as opposed to a benign lesion. The management of lesions suspicious for melanoma in situ is usually handled easily with complete surgical re-excision maintaining 5-mm margins if possible. Complete wide excision, maintaining 1-cm margins, is also recommended for lesions that may possibly represent invasive melanomas less than or equal to 1 mm in Breslow thickness. In anatomic areas where 1-cm surgical margins cannot be obtained easily because of cosmetic or functional considerations, such as on the face near the mouth, nose, or eye, complete excision is still advisable with as wide (though less than 1 cm) a margin as is reasonably possible. Lesions thought to represent, but not definitively confirmed to be, invasive melanoma greater than or equal to 1 mm are more difficult to manage because of the more complex surgery required and potential metastatic capability of lesions of this thickness. Once again, a detailed discussion with the patient and all parties concerned carefully delineating the pros and cons of treating such a lesion is essential and should be well documented in the official patient care record for medicolegal reasons. For those lesions arising in such anatomic areas as the chest wall, back, abdominal wall, or thigh, surgical excision with 2-cm margins seems reasonable because primary closure with an acceptable cosmetic and functional outcome can almost always be achieved. Although complete excision of the lesion with clear margins should performed, in those anatomic areas where 2-cm excision margins leave a defect not amenable to simple primary wound closure and necessitate more complex reconstructions or result in significant functional or cosmetic deformity, the actual excision margin width should be individualized and planned
and discussed carefully with the patient preoperatively. Sentinel lymph node mapping and biopsy may also be offered to patients with lesions that may represent melanomas greater than or equal to 1 mm in thickness. The risks and benefits of undergoing or declining this procedure should also be discussed carefully with the patient and family. Any patient who undergoes excision of a lesion of uncertain malignant potential should be followed carefully postoperatively as if the lesion were definitively proved to be melanoma. This includes routine physical examination and periodic laboratory and radiologic assessment appropriate for a patient with a melanoma of that particular thickness. Melanoma in pregnancy Although approximately one third of the increasing numbers of women who are diagnosed with melanoma each year are of childbearing age, melanoma accounts for about 8% of malignancies diagnosed during pregnancy [154]. The overall incidence of melanoma in pregnancy is estimated to be 0.14 to 0.28 cases per 1000 births [155]. Although occurring extremely rarely, melanoma is one of the most common tumors known to metastasize to the placenta and fetus [156,157]. Despite the fact that melanocytic nevi commonly become larger and darker under the hormonal influence of pregnancy presumably due to increased levels of estrogen and melanocyte-stimulating hormone [158,159], there exists no conclusive evidence that pregnancy significantly affects the biologic aggressiveness of a melanoma in terms of increasing the incidence of metastasis or lowering overall survival [160 163]. Moreover, pregnancy occurring either before or after the diagnosis and treatment of melanoma similarly seems to have no significant effect on the clinical course of the disease [164,165]. Based on the data presently available, the termination of pregnancy of a patient recently diagnosed with melanoma as a therapeutic measure cannot be recommended. Because the overwhelming (>75%) majority of melanoma occurrences happen within 2 to 3 years after treatment of the primary lesion, many women are encouraged to avoid becoming pregnant for that period of time postoperatively. The frequent observation that melanocytic lesions may become more pronounced during pregnancy makes the diagnosis of melanoma even more difficult. As in any other patient, however, any cutaneous lesion suspicious for melanoma in a pregnant patient should undergo biopsy without delay. This is accomplished with shave, punch, incisional, or pre-
694
ferably complete excisional biopsy, which may be performed safely in the pregnant patient under local lidocaine anesthesia, without the addition of epinephrine. Although most of these biopsy procedures may be carried out safely and rapidly in the office setting, the excision of larger lesions may be performed more prudently in the ambulatory surgery unit with an anesthesiologist, knowledgeable in the care of the pregnant patient, in attendance with intraoperative fetal monitoring during the later stages of pregnancy if appropriate. Once the diagnosis of melanoma is confirmed histologically, an abbreviated metastatic survey may be ordered but surgical treatment is planned commensurate with the thickness of the primary lesion and clinical stage of disease. Routine laboratory tests including determination of LDH level should be ordered. Radiologic work-up may include radiographs of the chest, which may be performed safely during pregnancy with the appropriate shielding. In patients with melanomas greater than or equal to 1 mm or those with palpable regional nodal metastases, sonographic examination of the abdomen and liver may be performed to search for visceral metastases instead of CT scanning with intravenous contrast, which is not recommended during the early stages of pregnancy. Wide local excision with 5-mm margins is performed under local anesthesia in patients with melanoma in situ. Invasive melanomas less than 1 mm are similarly treated under local anesthesia in the ambulatory surgery suite, with appropriate maternal-fetal monitoring, with wide and deep excision maintaining 1-cm margins. Patients with melanomas greater than or equal to 1 mm thick undergo wide and deep excision maintaining 2-cm margins under anesthesia administered by an anesthesiologist with experience in the care of the obstetric patient. Formal therapeutic regional lymph node dissection is performed concurrently in the presence of palpable nodal metastases. The treatment of the regional nodes in patients with melanomas greater than or equal to 1 mm thick and no clinical evidence of metastatic disease is less clear. Although intraoperative lymphatic mapping using vital blue dye and radiolabeled technetium sulfur colloid and sentinel lymphadenectomy is currently the accepted approach to nonpregnant patients with melanomas greater than or equal to 1 mm thick and no clinical evidence of nodal metastases, the accuracy and safety of this procedure in pregnant patients has not yet been studied completely and confirmed. The surgical approach to the regional lymph nodes in this group of patients must be individualized with all possibilities discussed in detail with the patient and her family.
Because no significant survival advantage has been demonstrated in prospective randomized clinical trials for patients with melanomas undergoing elective regional lymph node dissection, a reasonable approach is wide and deep excision of the primary melanoma and no treatment of the regional nodes at that time. Regional lymph node dissection may be performed at a later date should palpable nodal metastases become evident or after the completion of pregnancy.
References
[1] Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics, 1999. Cancer J Clin 1999;49:8 31. [2] Heston JF, Cusato MM, Young JL. Forty-five years of cancer incidence in Connecticut: 1935 79. National Cancer Institute monograph 70. NIH publication no. 86 2652. Washington, DC: Government Printing Office; 1986. [3] Glass AG, Hoover RN. The emerging epidemic of melanoma and squamous cell skin cancer. JAMA 1989;262:2097 100. [4] Jemal A, Thomas A, Murray T, Thin M. Cancer statistics, 2002. CA Cancer J Clin 2002;52:23 47. [5] Greenlee RT, Hill-Harmon MB, Murray T, Thun M. Cancer statistics, 2001. CA Cancer J Clin 2001;51: 15 36. [6] Friedman RJ, Rigel DS, Kopf AW. Early detection of malignant melanoma: the role of physician examination and self examination of the skin. CA Cancer J Clin 1985;35:130 51. [7] Rhodes AR, Weinstock MA, Fitzpatrick TB, et al. Risk factors for cutaneous melanoma: a practical method of recognizing predisposed individuals. JAMA 1987; 258:3146 54. [8] Ries LAG, Eisner MP, Rosary CL, et al. SEER cancer statistics review, 1973 1997. Bethesda, MD: National Cancer Institute; 2000. [9] Balch CM, Soong SJ, Shaw HM, et al. An analysis of prognostic factors in 8500 patients with cutaneous melanoma. In: Balch CM, Houghton AN, Milton GW, et al, editors. Cutaneous melanoma. 2nd edition. Philadelphia: JB Lippincott; 1992. p. 165 87. [10] Ross MI, Stern SJ, Wanebo HJ. Mucosal melanomas. In: Balch CM, Houghton AN, Milton GW, et al. Cutaneous melanoma. 2nd edition. Philadelphia: JB Lippincott; 1992. p. 325 35. [11] Reintgen DS, McCarty KS, Woodard B, Cox E, Seigler HF. Metastatic malignant melanoma from an unknown primary. Ann Surg 1983;156:335 50. [12] Giuliano AE, Moseley HS, Morton DL. Clinical aspects of unknown primary melanoma. Ann Surg 1980;191:98 103. [13] Balch CM, Soong S-J, Murad TM, Ingalls AL, Maddox WA. A multifactorial analysis of melanoma: III. Prognostic factors in melanoma patients with lymph node metastases. Ann Surg 1981;193:377.
R.L. Shapiro / Dermatol Clin 20 (2002) 681699 [14] Milton GW, Shaw HM, McCarthy WH. Occult primary malignant melanoma: factors influencing survival. Br J Surg 1977;64:805. [15] Rhodes AR, Weinstock MA, Fitzpatrick TB, Mihm Jr. MC, Sober AJ. Risk factors for cutaneous melanoma: a practical method of recognizing predisposed individuals. JAMA 1987;258:3145 53. [16] Epidemeology BM. Current trends, risk factors, and environmental concerns. In: Balch CM, Houghton AN, Sober AJ, Soong SJ, editors. Cutaneous melanoma. 3rd edition. St. Louis: Quality Medical Publishing; 1998. p. 551 71. [17] Gilcrest BA, Eller MS, Geller AC, Yaar M. The pathogenesis of melanoma induced by ultraviolet radiation. N Engl J Med 1999;340:1341 8. [18] Rigel DS, Rivers JK, Kopf AW, Friedman RJ, Vinokur AF, Heilman ER. Dysplastic nevi. markers for increased risk for melanoma. Cancer 1989;63: 386 7. [19] Elder DE. Acquired melanocytic nevi and melanoma: the dysplastic nevus syndrome. In: Ackerman AB, editor. Pathology of malignant melanoma. New York: Karger; 1987. p. 185. [20] MacKie RM, Freudenberger T, Aitchison TC. Personal risk factor chart for cutaneous melanoma. Lancet 1989;2:1989. [21] Ackerman AB. What naevus is dysplastic, a syndrome, and the commonest precursor of malignant melanoma? A riddle and an answer. Histopathology 1988;13:241 56. [22] Rigel DS, Friedman RJ, Kopf AW, Rogers GS, Heilman ER. Precursors of malignant melanoma. Dermatol Clin 1985;3:361 5. [23] National Institutes of Health. Precursors of malignant melanoma consensus report. Bethesda, MD: 1983, Oct 24 6;4(9):1 14. [24] Ackerman AB, Su WPD. Histopathology of cutaneous malignant melanoma. In: Kopf AW, Bart RS, Rodriguez-Sains R, Ackerman AB, editors. Malignant melanoma. New York: Masson; 1979. p. 25. [25] Gruber SB, Barnhill RL, Stern KS, Roush GC. Nevomelanocytic proliferations in association with cutaneous melanoma: a multivariate analysis. J Am Acad Dermatol 1989;21:773 86. [26] Rhodes AR, Melski JW. Small congenital nevocellular nevi and the risk of cutaneous melanoma. J Pediatr 1982;100:219. [27] Illig L, Weidner F, Hundeiker M, Gartmann H, Biess B, Leyh F, et al. Congenital nevi less than or equal to 10 cm as precursors to melanoma: 52 cases, a review, and a new conception. Arch Dermatol 1985;121:1274 80. [28] Hendrickson MR, Ross JC. Neoplasms arising in congenital giant nevi: morphologic study of seven cases and a review of the literature. Am J Surg Pathol 1981; 5:109 15. [29] Scheibner A, Milton GW, McCarthy WH, Nordlund JJ, Pearson LJ. Multiple primary melanoma: a review of 90 cases. Aust J Dermatol 1982;13:607. [30] Veronesi U, Cascinelli N, Bufalino R. Evaluation of
695
[31]
[32]
[33]
[34] [35]
[36]
[37] [38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
multiple primaries in malignant cutaneous melanoma. Tumori 1976;62:127. Greene MH, Clark Jr. WH, Tucker MA, Elder DE, et al. Acquired precursors of cutaneous malignant melanoma: the familial dysplastic nevus syndrome. N Engl J Med 1985;312:91. Greene MH, Clark Jr. WH, Tucker MA, Kraemer KH, Elder DE, Fraser MC. High risk of malignant melanoma in melanoma-prone families with dysplastic nevi. Ann Intern Med 1985;102:458. Reimer RR, Clark WH, Greene MH, Ainsworth AM, Fraumeni Jr. JF. Precursor lesions in familial melanoma: a new genetic preneoplastic syndrome. JAMA 1978;239:744. Koh HK. Cutaneous melanoma. N Engl J Med 1991; 325:171 82. Mihm Jr. MC, Fitzpatrick TB, Lane Brown MM, Raker JW, Malt RA, et al. Early detection of primary cutaneous malignant melanoma: a color atlas. N Engl J Med 1973;289:989 96. Wick MM, Sober AJ, Fitzpatrick TB, et al. Clinical characteristics of early cutaneous melanoma. Cancer 1980;45:2684 6. Harris MN, Gumport SL. Biopsy technique for malignant melanoma. J Dermatol Surg 1975;1:24 9. Gown AM, Vogel AM, Hoak D, Gough F, McNutt MA. Monoclonal antibodies specific for melanocytic tumors distinguish subpopulations of melanocytes. Am J Pathol 1986;123:195. Wick MR, Swanson PE, Rocamora A. Recognition of malignant melanoma by monoclonal antibody HMB45: an immunohistochemical study of 200 paraffinembedded cutaneous tumors. J Cutan Pathol 1988; 15:201. Cochran AJ, Wen D-R. S-100 protein as a marker for melanocytic and other tumors. Pathology 1985; 17:340. Nakajima T, Watanabe S, Sato Y, Kameya T, Shimosato Y, Ishihara K. Immunohistochemical demonstration of S-100 protein in malignant melanoma and pigmented nevus, and its diagnostic application. Cancer 1982;36:912. Wick MR, Swanson PE, Rocamora A. Recognition of malignant melanoma by monoclonal antibody HMB45: an immunohistochemical study of 200 paraffinembedded cutaneous tumors. J Cutan Pathol 1988; 15:201. Balch CM, Soong S, Shaw HM, et al. An analysis of prognostic factors in 8500 patients with cutanoeous melanoma. In: Balch CM, Houghton AN, Milton GW, et al, editors. Cutaneous melanoma. 2nd edition. Philadelphia: JB Lippincott; 1992. p. 165 76. Thickness BA. Cross-sectional areas and depth of invasion in the prognosis of cutaneous melanoma. Ann Surg 1970;17:902 8. Clark Jr. WH, From L, Bernardino EA, Mihm MC. The histogenesis and biologic behavior of primary human malignant melanomas of the skin. Cancer Res 1969; 29:705.
696
R.L. Shapiro / Dermatol Clin 20 (2002) 681699 primary cutaneous malignant melanoma: comparison of excision with margins of 1 or 3cm. N Engl J Med 1988;318:1159 62. Balch CM, Urist MM, Karakousis CP, et al. Efficacy of 2cm margins for intermediate-thickness melanomas (1 4cm): results of a multi-institutional randomized surgical trial. Ann Surg 1993;218:262 9. Roses DF, Harris MN, Rigel D, Carrey Z, Friedman R, Kopf AW. Local and in-transit metastases following definitive excision for primary cutaneous malignant melanoma. Ann Surg 1983;198:65 9. Veronesi U, Cascinelli N, Adamus J, Balch CM, Bandiera D, Barchuk A, et al. Narrow excision (1 cm margin): a safe procedure for thin cutaneous melanoma. The results of an international randomized clinical trial. Arch Surg 1991;126:438 41. Veronesi U, Cascinelli N, Adamus J, et al. Thin stage primary cutaneous malignant melanoma: comparison of excision with margins of 1 or 3cm. N Engl J Med 1988;318:1159 62. Balch CM, Urist MM, Karakousis CP, et al. Efficacy of 2cm margins for intermediate-thickness melanomas (1 4cm): results of a multi-institutional randomized surgical trial. Ann Surg 1993;218:262 9. Heaton KM, Sussman JJ, Gershenwald JE, Lee JE, Reintgen DS, Mansfield PF, et al. Surgical margins and prognostic factors in patients with thick (>4 mm) primary melanoma. Ann Surg Oncol 1998; 5:322 8. Jaques DP, Coit DG, Brennan MF. Major amputation for advanced malignant melanoma. Surg Gynecol Obstet 1989;169:1 6. Cascinelli N, Morabito A, Santinami M, et al. Immediate or delayed dissection of regional lymph nodes in patients with melanoma of the trunk: a randomised trial. WHO Melanoma Programme. Lancet 1998; 351: 793 6. Balch CM, Soong SJ, Bartolucci AA, et al. Efficacy of elective regional lymph node dissection of 1 to 4 mm thick melanomas for patients 60 years of age and younger. Ann Surg 1996;224:255 66. Balch CM. The role of elective lymph node dissection in melanoma: rationale, results, controversies. J Clin Oncol 1988;6:163 9. McCarthy WH, Shaw HM, Milton GW. Efficacy of elective lymph node dissection in 2,347 patients with clinical stage I malignant melanoma. Surg Gynecol Obstet 1985;161:575. Milton GW, Shaw HM, McCarthy WH, et al. Prophylactic lymph node dissection in clinical stage I cutaneous malignant melanoma: results of treatment in 1319 patients. Br J Surg 1982;69:108. Sim FH, Taylor WF, Irvins JC, et al. A prospective randomized study of the efficacy of routine prophylactic elective lymphadenectomy in management of malignant melanoma. Cancer 1978;41:948. Veronesi U, Adamus J, Bandiera DC, et al. Inefficacy of immediate node dissection in stage I melanoma of the limbs. N Engl J Med 1977;297:627.
[46] Clark Jr. WH, Ainsworth AM, Bernardino EA, Yang CH, Mihm Jr. MC, Reed RJ. The developmental biology of primary human malignant melanomas. Semin Oncol 1975;2:83 90. [47] Fleming ID, Cooper JS, Henson DE, editors. AJCC Cancer Staging Manual. Philadelphia (PA): Lippincott Raven Publishers; 1997. p. 201 6. [48] Buzaid AC, Ross MI, Balch CM, et al. Critical analysis of the current American Joint Commission on Cancer staging system for cutaneous melanoma and proposal for a new staging system. J Clin Oncol 1997; 15:1039 51. [49] Gershenwald JE, Thompson W, Mansfield PF, et al. Multi-institutional melanoma lymphatic mapping experience: the prognostic value of sentinel lymph node status in 612 stage I and II melanoma patients. J Clin Oncol 1999;17:976 83. [50] Gambhir SS, Czernin J, Schwimmer J, Silverman DHS, Coleman RE, Phelps ME. A tabulated summary of the FDG PET literature. J Nucl Med 2001; 42(5):15 93S. [51] Mijnhout GS, Hoekstra OS, van Tulder MW, Teule GJJ, Deville W. Systematic review of the diagnostic accuracy of F-18-fluorodeoxyglucose positron emission tomography in melanoma patients. Cancer 2001; 91:1530 42. [52] Gambhir SS, Hoh CK, Essner R, et al. A decision analysis model for the role of whole-body FDG PET in the management of patients with recurrent melanoma. J Nucl Med 1998;39:94P 5P. [53] Pemberton O. Observations on the history, pathology, and treatment of cancerous diseases. Part I: melanosis. London: Churchill; 1858. [54] Pringle JH. A method of operation in cases of melanotic tumors of the skin. Edinb Med J 1908;23:496 9. [55] Handley WS. The pathology of melanocytic growths in relation to their operative treatment. Lancet 1907; I:927. [56] Handley WS. The Hunterian lectures on the pathology of melanocytic growths in relation to their treatment. Lancet 1907;1:927 35. [57] Breslow A, Macht SD. Optimal size of resection margin for thin cutaneous melanoma. Surg Gynecol Obstet 1977;145:691 2. [58] Veronesi U, Cascinelli N, Adamus J, Balch CM, Bandiera D, Barchuk A, et al. Narrow excision (1 cm margin): a safe procedure for thin cutaneous melanoma. The results of an international randomized clinical trial. Arch Surg 1991;126:438 41. [59] Harris MN, Shapiro RL, Roses DF. Malignant melanoma: primary surgical treatment (excision and lymph node resection) based on pathology and staging. Cancer 1995;75:715 25. [60] Urist MM, Balch CM, Soong S-J, Shaw HM, Milton GW, Maddox WA. The influence of surgical margins and prognostic factors predicting the risk of local recurrence in 3445 patients with primary cutaneous melanoma. Cancer 1985;55:1398 402. [61] Veronesi U, Cascinelli N, Adamus J, et al. Thin stage
[62]
[63]
[64]
[65]
[66]
[67]
[68]
[69]
[70]
[71]
[72]
[73]
[74]
[75]
R.L. Shapiro / Dermatol Clin 20 (2002) 681699 [76] Veronesi U, Adamus J, Bandiera DC, et al. Delayed regional node dissection in stage I melanoma of the skin of the lower extremities. Cancer 1982;49:2420. [77] Veronesi U, Adamus J, Bandiera DC, et al. Stage I melanoma of the limbs: immediate versus delayed node dissection. Tumori 1980;66:373. [78] Morton DL, Wen D-R, Wong JH, et al. Technical details of Intraoperative lymphatic mapping for early stage melanoma. Arch Surg 1992;127:392 9. [79] Morton DL, Wanek L, Nizze JA, et al. Improved long-term survival after lymphadenectomy of malignant melanoma metastatic to the regional nodes. Ann Surg 1991;214:491. [80] Morton DL, Wen D-R, Cochran AJ. Management of early stage melanoma by intraoperative lymphatic mapping and selective lymphadenectomy: an alternative to routine elective lympadenectomy or watch and wait. Surg Oncol Clin North Am 1992;1:247 59. [81] Ross ML, Reintgen DS, Balch CM. Selective lymphadenectomy: emerging role for lymphatic mapping and sentinel node biopsy in the management of early stage melanoma. Semin Surg Oncol 1993;9:219 23. [82] Sappey PC. Anatomie, physiologie des vaisseaux lymphatiques consideres chez lhomme et les vertebras. Paris: A.Delhave and E. Lecrosnier; 1885. [83] Lock-Andersen J, Rossing N, Drzewiecki KT. Preoperative cutaneous lymphoscintigraphy in malignant melanoma. Cancer 1989;63:77 82. [84] Reintgen DS, Cruse CW, Wells K, et al. The orderly progression of melanoma nodal metastases. Ann Surg 1994;220:759 67. [85] Alex JC, Krag DN. Gamma-probe guided localization of lymph nodes. Surg Oncol 1993;2:137 43. [86] Albertini JJ, Cruse CW, Rapaport D, Wells K, Ross M, et al. Intraoperative radiolymphoscintigraphy improves sentinel lymph node identification for patients with melanoma. Ann Surg 1996;223:217 24. [87] Alex JC, Weaver DL, Fairbank JT, et al. Gammaprobe guided lymph node localization in malignant melanoma. Surgical Oncol 1993;2:303 308. [88] Morton DL, Thompson JF, Essner R, et al. Validation of the accuracy of intraoperative lymphatic mapping and sentinel lymphadenectomy for early-stage melanoma: a multicenter trial. Multicenter Selective Lymphadenectomy Trial Group. Ann Surg 1999;230: 453 63. [89] Albertini JJ, Cruse CW, Rapaport D, Wells K, Ross M, et al. Intraoperative radiolymphoscintigraphy improves sentinel lymph node identification for patients with melanoma. Ann Surg 1996;223:217 24. [90] Multicenter SLT. National Cancer Institute Grant No. PO1 CA29605 12. [91] Morton DL, Thompson JF, Essner R, et al. Validation of the accuracy of intraoperative lymphatic mapping and sentinel lymphadenectomy for early-stage melanoma: a multicenter trial. Multicenter Selective Lymphadenectomy Trial Group. Ann Surg 1999;230: 453 63. [92] Wang X, Heller R, VanVoorhis N, et al. Detection of
697
[93]
[94]
[95]
[96]
[97]
[98]
[99]
[100]
[101]
[102]
[103]
[104]
[105]
submicroscopic lymph node metastases with polymerase chain reaction in patients with malignant melanoma. Ann Surg 1994;220:768 74. Lukowsky A, Bellmann B, Ringk A, et al. Detection of melanoma micrometastases in the sentinel lymph node and in nonsentinel lymph nodes by tyrosine kinase polymerase chain reaction. Invest Dermatol 1999;113:554 9. Blaheta HJ, Schittek B, Breuninger H, et al. Lymph node micrometastases of cutaneous melanoma: increased sensitivity of molecular diagnosis in comparison to immunohistochemistry. Int J Cancer 1998;79: 318 23. Weiguo L, Stall A, Shivers SC, Lin J, et al. Clinical relevance of molecular staging for melanoma: comparison of RT-PCR and immunohistochemistry staining in sentinel lymph nodes of patients with melanoma. Ann Surg 2000;231:795 803. Basler GC, Fader DJ, Yahanda A, Sondak VK, Johnson TM. The utility of fine needle aspiration in the diagnosis of melanoma metastatic lymph nodes. J Am Acad Dermatol 1997;36:403 8. Hamilton J, Fernandez G, Cangiarella JF, Symmans WF, Waisman J. Aspiration biopsy assessment of suspected lymph node metastases in patients with malignant melanoma. Acta Cytol 1997;41:1606. Perry MD, Seigler HF, Johnston WW. Diagnosis of metastatic melanoma by fine needle aspiration biopsy: a clinical and pathologic correlation of 298 cases. J Natl Cancer Inst 1986;77:1013 21. Balch CM, Soong SJ, Murad TM, et al. A multifactorial analysis of melanoma III: prognostic factors in melanoma patients with lymph node metastases (stage II). Ann Surg 1981;193:377 88. Callery C, Cochran AJ, Roe DJ, et al. Factors prognostic for survival in patients with malignant melanoma spread to the regional lymph nodes. Ann Surg 1982;196:69 75. Kirkwood JM, Strawderman MH, Ernstoff MS, et al. Interferon alfa-2b adjuvant therapy of high-risk resected cutaneous melanoma: the Eastern Cooperative Oncology Group trial EST 1684. J Clin Oncol 1996; 14:7 17. Kirkwood JM, Ibrahim JG, Sosman JA. High dose interferon alfa-2b significantly prolongs relapse-free and overall survival compared with GM2-KLH/QS21 vaccine in patients with resected stage IIB-III melanoma: results of intergroup trial E1694/S9512/ C509801. J Clin Oncol 2001;19:2370 80. Kirkwood JM, Ibrahim JG, Sondak VK, et al. Highand low-dose interferon alfa-2b in high-risk melanoma: first analysis of intergroup trial E1690/S9111/ C9190. J Clin Oncol 2000;18:2444 58. Bystryn JC, Shapiro RL, Harris MN, et al. The use of vaccines in the treatment of malignant melanoma. Clin Dermatol 1996;14:337 41. Bystryn JC, Shapiro RL, Oratz R, et al. Partially purified tumor antigen vaccines. In: Devita VT Jr., Hellman S, Rosenberg SA, editors. Biological therapy of
698
R.L. Shapiro / Dermatol Clin 20 (2002) 681699 cancer. 2nd edition. Philadelphia: JB Lippincott; 1995. p. 668 79. Balch CM, Soong SJ, Shaw HM, et al. An analysis of prognostic factors in 8500 patients with cutaneous melanoma. In: Balch CM, Houghton AN, Milton GW, et al, editors. Cutaneous melanoma. 2nd edition. Philadelphia: JB Lippincott; 1992. p. 165. Hafstrom L, Rudenstam CM, Blomquist E, et al. Regional hyperthermic perfusion with melphalan after surgery for recurrent melanoma of the extremities. J Clin Oncol 1991;9:2091 8. Wong JH, Cagle LA, Kopald KH, Swisher SG, Morton DL. Natural history and selective management of in-transit melanoma. J Surg Oncol 1990;44:146 50. Buzaid AC, Bedikian A, Houghton AN. Systemic chemotherapy and biotherapy. In: Balch CM, Houghton AN, Sober AJ, et al, editors. Cutaneous melanoma. 3rd edition. Philadelphia: JB Lippincott; 1998. p. 408. Kirkwood JM, Strawderman MH, Ernstoff MS, et al. Interferon alfa-2b adjuvant therapy of high-risk resected cutaneous melanoma: the Eastern Cooperative Oncology Group Trial EST 1684. J Clin Oncol 1996; 14:7 16. Creech O Jr, Krementz ET, Ryan RF, et al. Chemotherapy of cancer: regional perfusion utilizing an extracorporeal circuit. Ann Surg 1958;148:616. Thompson JF, Gianoutsos MP. Isolated limb perfusion for melanoma: effectiveness and toxicity of cisplatin compared with melphalan and other drugs. World J Surg 1992;16:227 36. Fraker DL. Hyperthermic regional perfusion for melanoma of the limbs. In: Balch CM, Houghton AN, Sober AJ, et al, editors. Cutaneous melanoma. 3rd edition. Philadelphia: JB Lippincott; 1998. p. 288. Cavaliere R, Ciogatto EC, Giovanelli BC, et al. Selective heat sensitivity of cancer cells. Cancer 1967; 20:1351. Stehlin JS. Hyperthermic perfusion with chemotherapy for cancers of the extremity. Surg Gynecol Obstet 1969;129:305. Bulman A, Jamieson CW. Isolated limb perfusion with melphalan in the treatment of malignant melanoma. Br J Surg 1980;67:660. Klaase JM, Kroon BB, Eggermont AM, van Geel AN, Schraffordt Koops H, Oldhoff J, et al. A retrospective comparative study evaluating the results of mild hyperthermic versus controlled normothermic perfusion for recurrent melanoma of the extremities. Eur J Cancer 1995;31A:58 63. Fraker DL. Hyperthermic regional perfusion for melanoma of the limbs. In: Balch CM, Houghton AN, Sober AJ, et al, editors. Cutaneous melanoma, 3rd edition. Philadelphia: JB Lippincott; 1998. p. 285. McBride CM, Sugarbaker EV, Hickey RC. Prophylactic isolation-perfusion as the primary therapy for invasive malignant melanoma of the limbs. Ann Surg 1975;182:316. Franklin HR, Koops HS, Oldhoff J, et al. To perfuse or not perfuse? A retrospective comparative study to evaluate the effect of adjuvant isolated regional perfusion in patients with stage I extremity melanoma with a thickness of 1.5mm or greater. J Clin Oncol 1988;6:701. Schraffordt Koops H, Vaglini M, Suciu S, et al. Prophylactic isolated limb perfusion for localized, high risk limb melanoma: results of a multicenter randomized phase III trial. J Clin Oncol 1998;16:2906 12. Lienard D, Ewalenko P, Delmotte JJ, et al. High dose recombinant tumor necrosis factor in combination with interferon gamma and melphalan in isolation perfusion of limbs for melanomas and sarcomas. J Clin Oncol 1992;1:5260. Fraker DR, Alexander HR, Barlett DL, et al. A prospective randomized trial of therapeutic isolated limb perfusion (ILP) comparing melphalan (M), versus melphalan, tumor necrosis factor (TNF) and interferon-g (IFN): an initial report. Proceedings of the American Society of Surgical Oncology 1996;49:6. Fraker DR, Alexander HR, Andrich M, Rosenberg S. Treatment of patients with melanoma of the extremity using hyperthermic isolated limb perfusion with melphalan, tumor necrosis factor, and interferon -gamma: results of a tumor necrosis factor dose escalation study. J Clin Oncol 1996;14:479 89. Vrouenraets BC, Hart GAM, Eggermount AMM, Klaase JM, van Geel BN, Nieweg OE, et al. Relation between limb toxicity and treatment outcomes after isolated limb perfusion for recurrent melanoma. J Am Coll Surg 1999;188:522 30. Taber SW, Polk Jr. HC. Mortality, major amputation rates, and leukopenia after isolated perfusion with phenylalanine mustard for the treatment of melanoma. Ann Surg Oncol 1997;4:440 5. van Geel AN, van Wijk J, Wieberdink J. Functional morbidity after hyperthermic isolation perfusion of the limb for melanoma. Cancer 1989;63:1092. Vrouenraets BC, Klaase JM, Kroon BB, van Geel BN, Eggermount AM, Franklin HR. Long-term morbidity after regional isolated perfusion with melphalan for melanoma of the limbs: the influence of acute regional toxic reactions. Arch Surg 1995;130: 43 7. Sonneveld EJ, Vrouenraets BC, van Geel BN, Eggermont AM, Klaase JM, Nieweg OE, et al. Systemic toxicity after isolated limb perfusion with melphalan for melanoma. Eur J Surg Oncol 1996;22:521 7. Barker WC, Andrich MP, Alexander HR, et al. Continuous intraoperative external monitoring of perfusate leak using iodine-131 human serum albumin during isolation perfusion of the liver and limbs. Eur J Nucl Med 1995;22:1242. Sandrock D, Horst F, Gatzemeier W, Ghorbani M, Rauschecker H, Munz DL, et al. Leakage measurement during selective limb perfusion using a gamma probe. Eur J Nucl Med 1996;23:534 8. Morton DL. Immunological studies with human neoplasms. J Reticulendothel Soc 1971;10:137.
[106]
[121]
[107]
[122]
[108]
[109]
[123]
[110]
[124]
[111]
[125]
[112]
[113]
[126]
[114]
[127]
[115]
[128]
[116]
[117]
[129]
[130]
[118]
[131]
[119]
[132]
[120]
R.L. Shapiro / Dermatol Clin 20 (2002) 681699 [133] Mastrangelo MJ, Sulit HJ, Prehn LM. Intralesional BCG in the treatment of metastatic malignant melanoma. Cancer 1976;37:684. [134] Krown SE, Hilal EY, Pinsky CM. Intralesional injection of the MER into cutaneous metastases of malignant melanoma. Cancer 1978;42:648. [135] Cohen MH, Felix E, Jessup J, et al. Treatment of metastatic melanoma by intralesional injection of BCG and organic chemicals. In: Crispen RG, editor. Neoplasm immunity mechanisms. Philadelphia: Franklin Institute Press; 1975, p. 121. [136] Von Wussow P, Block B, Hartman F, Deicher H. Intralesional interferon-alpha therapy in advanced malignant melanoma. Cancer 1988;61:1071 4. [137] Holding JD, Linup WE, Bowdler DA, et al. Disposition and tumor concentrations of platinum in hypoalbuminaemic patients after treatment with cisplatin for cancer in the head and neck. Br J Clin Pharmacol 1991;32:173 9. [138] Yu NY, Conley F, Luck E, Brown D. Response of murine tumors to matrix-associated cisplatin intratumoral implants. NCI Monogr 1988;6:137 40. [139] Oratz R, Hauschild A, Sebastian G, Reiss H, Castro D. Intratumoral cisplatin/epinephrine injectable gel provides palliative tumor control for patients with metastatic melanoma. Presented at the Chemotherapy Foundation Symposium on Innovative Cancer Therapy for Tomorrow [abstract 72]. November 11 13, 1998. [140] Briggs JC. Subungual malignant melanoma: a review article. Br J Plast Surg 1985;38:174. [141] Finley RK, Driscoll DL, Blumenson LE, et al. Subungual melanoma: an eighteen year review. Surgery 1994;116:96 100. [142] Feibleman CE, Maize JC. Racial differences in cutaneous melanoma incidence and distribution. In: Ackerman AB, editor. Pathology of malignant melanoma. New York: Masson Publishing; 1981. p. 47. [143] Glat PM, Spector JA, Roses DF, Shapiro RL, Harris MN, Beasley RW, et al. The Management of pigmented lesions of the nail bed. Ann Plast Surg 1996;37:183 9. [144] Ebskov LB. Major amputation for malignant melanoma: an epidemiological study. J Surg Oncol 1993;52: 89 91. [145] Glat PM, Spector JA, Roses DF, Shapiro RL, Harris MN, Beasley RW, et al. The management of pigmented lesions of the nail bed. Ann Plast Surg 1996;37:187. [146] Reintgen DS, McCarty Jr. KM, Cox E, et al. Malignant melanoma in black American and white American populations: a comparative review. JAMA 1982; 248:1856 61. [147] Krementz ET, Reed RJ, Coleman III WP, et al. Acral lentiginous melanoma: a clinicopathologic entity. Ann Surg 1982;195:632 41.
699
[148] Clark Jr. WH, Mihm Jr. MC. Lentigo maligna and lentigo-maligna melanoma. Am J Pathol 1969;55: 39 45. [149] Veronesi U, Cascinelli N, Adamus J, et al. Thin stage primary cutaneous malignant melanoma: comparison of excision with margins of 1 or 3cm. N Engl J Med 1988;318:1159 62. [150] Zitelli JA, Brown CD, Hanusa BH. Mohs micrographic surgery for the treatment of primary cutaneous melanoma. J Am Acad Dermatol 1997;37:236 45. [151] Roses DF, Harris MN, Grunberger I, et al. Selective surgical management of cutaneous melanoma of the head and neck. Ann Surg 1980;192:629 32. [152] Byers RM, Medina JE, Wolf PF. Regional node dissection in the head and neck. In: Balch CM, editor. Pigment cell: surgical approaches to cutaneous melanoma. Basel: Karger; 1985. p. 83 104. [153] Morton DL, Wen D-R, Foshag LJ. Intraoperative lymphatic mapping and selective cervical lymphadenectomy for early stage melanomas of the head and neck. J Clin Oncol 1993;11:1751 6. [154] Kjems E, Krag C. Melanoma and pregnancy: a review. Acta Oncol 1993;32:371. [155] Wong DJ, Strassner HT. Melanoma in pregnancy. Clin Obstet Gynecol 1990;33:782. [156] Baergen RN, Johnson D, Moore T, Benirschke K. Maternal melanoma metastatic to the placenta: a case report and review of the literature. Arch Pathol Lab Med 1997;121:508. [157] Dildy III GA, Moise Jr. KJ, Carpenter Jr. RJ, Klima T. Maternal malignancy metastatic to the products of conception: a review. Obstet Gynecol Surv 1989; 44:535. [158] Ances IG, Pomerantz SH. Serum concentrations of beta-melanocyte-stimulating hormone in pregnancy. Am J Obstet Gynecol 1974;119:1062. [159] Sanchez JL, Figuero LD, Rodriguez E. Behavior of melanocytic nevi during pregnancy. Am J Dermatopathol 1984;6:9. [160] Wong JH, Sterns EE, Kopland KH, Nizze JA, Morton DL. Prognostic significance of pregnancy in stage I melanoma. Arch Surg 1989;124:1227. [161] Houghton AN, Flannery J, Viola MV. Malignant melanoma of the skin occurring during pregnancy. Cancer 1981;48:407. [162] Reintgen DS, McCarthy Jr. KS, Vollmer R, Cox E, Seigler HF. Malignant melanoma and pregnancy. Cancer 1985;44:1340. [163] Slingluff CL, Reintgen DS, Vollmer RT, Seigler HP. Malignant melanoma arising during pregnancy: a study of 100 patients. Ann Surg 1990;211:552. [164] Lederman JS, Sober AJ. Effect of prior pregnancy on melanoma survival. Arch Dermatol 1985;121:716. [165] Reintgen DS, McCarthy Jr. KS, Vollmer R, Cox E, Seigler HF. Malignant melanoma and pregnancy. Cancer 1985;44:1340.
Dermatol Clin 20 (2002) 701 708
Mohs micrographic surgery for the treatment of melanoma

John A. Carucci, MD, PhD
Chief, Mohs Micrographic and Dermatologic Surgery, Department of Dermatology, Weill Medical College of Cornell University, 525 East 68th Street, New York, NY 10021, USA
The incidence of melanoma continues to rise with approximately 50,000 diagnosed annually in the United States [1]. It is estimated that 1 in 70 Americans will develop melanoma [2]. Although the trend is toward earlier diagnosis of thinner melanomas, the mortality rate from melanoma continues to be significant [3]. Standard surgical excision with clear margins remains the treatment of choice [3]; however, the use of Mohs micrographic surgery (MMS) for melanoma continues to gain popularity [4]. This article focuses on the basis of MMS as a treatment option for melanoma by examination of tumor biology, standard surgical excision margins, and the Mohs method; interpretation of melanoma on frozen sections; and the use of special stains to increase the specificity of detection of melanoma.
Biologic behavior of melanoma Melanoma is a life-threatening cancer of melanocytes with the potential to metastasize to regional lymph nodes and distant organs including liver, bone, and brain [5]. Prognosis is proportional to Breslow depth (Table 1) [6]. Melanoma in situ has an excellent prognosis and a high cure rate after excision. Melanomas less than 0.76 mm are considered thin and carry an overall better prognosis than melanomas greater than 0.76 mm in depth. Lentigo maligna (LM), an indolent subtype of melanoma in situ that occurs on sun-damaged areas of the head and neck, may progress to invasive LM melanoma [7]. Because of the potentially aggressive nature of melanoma,
surgical excision with appropriate clear margins is the treatment of choice. Appropriate margins of excision have decreased throughout the twentieth century and the trend toward narrow margins continues today [8]. One of the paramount biologic questions in melanoma management is whether excision of the primary cancer is sufficient for cure. If so, then narrow clear margins should be sufficient for cure. If not, then wide margins might be required as originally thought, to decrease the potential for undetectable microscopic foci of tumor cells to persist and metastasize. This is not necessarily supported by the fact that findings of in-transit metastatic disease on excision results in poorer prognosis with higher risk of widespread metastatic disease [9]. In one study, primary melanomas greater than 1.5 cm with microscopic metastases were associated with nodal metastases more often than melanomas of similar thickness without satellites (53% versus 12%) [9]. Satellites were defined as nests of melanoma greater than 0.05 mm in the reticular dermis, panniculus, or vessels beyond the principal tumor but separated by normal tissue.
Treatment of primary melanoma Standard excision and margins Excision margins for melanoma range from narrow (0.5 cm) for melanoma in situ to 2 cm for lesions greater than 2 mm in Breslow depth (Table 2) [8]. The concept of wide local excision is based on the idea that excision of the cancer alone is not sufficient for cure. Wide margins are taken to excise any otherwise undetected micrometastatic disease; how-
E-mail address: jac2015@med.cornell.edu
702 Table 1 Prognosis based on melanoma depth Depth of invasion (mm) < 0.76 0.76 1.49 1.50 2.49 2.50 3.99 > 4.0
J.A. Carucci / Dermatol Clin 20 (2002) 701708
5-year Survival (%) 96 87 75 66 47
ever, the presence of satellite metastases has been associated with an increase in regional disease even when completely excised [9]. Clarification of margin assessment of standard excision specimens is necessary. Margins are assessed from representative sections of the specimen in bread loaf fashion allowing only sampling of the specimen [10]. This degree of examination may occasionally result in a false-negative assessment of clear margins in cases of infiltrating, or aggressive growth cancers. Similar misdiagnosis may result when vertically cut frozen specimens are relied on for intraoperative margin control.
Mohs micrographic surgery Mohs micrographic surgery facilitates optimal margin control and conservation of normal tissue in the management of nonmelanoma skin cancer [11 13]. Dermatologic surgeons specially trained in the technique perform MMS in an office setting under local anesthesia. Briefly, a tangential specimen of tumor with a minimal margin of clinically normalappearing tissue is obtained, precisely mapped, and processed immediately by frozen section for microscopic examination (Fig. 1). Optimal margin control is obtained by examination of the entire perimeter of the specimen and contiguous deep margin. Meticulous mapping allows for directed extirpation of any remaining tumor. A key defining feature of
Table 2 Standard excision margins based on melanoma depth Depth of invasion (mm) Melanoma-in-situ 1 1.01 4a >4 Standard excision margins (cm) 0.5 1 2 2 vs 2 +
a Consider sentinel lymph node dissection before definitive excision in cases where Breslow depth exceeds 1 mm or when melanoma invades to Clarks level IV.
Fig. 1. Mohs micrographic surgery. (A, B) A beveled specimen is obtained with a margin of normal-appearing tissue. (C, D) The tissue is divided, inked, and mapped before processing of frozen sections. Precise mapping allows for directed extirpation of remaining tumor. (Courtesy of Dr. Stuart Zweibel, Mt. Kisco, New York.)
703
MMS is that the surgeon excises, maps, and reviews the specimen personally, minimizing the chance of error in tissue interpretation and orientation. MMS has gained acceptance as the treatment of choice for recurrent skin cancers and primary skin cancers located on anatomic sites requiring maximal tissue conservation for preservation of function and cosmesis [14 17]. The use of the technique depends on the ability of the Mohs surgeon to detect cancer on the frozen sections.
Detection of melanoma on frozen sections The use of MMS for melanoma depends on the surgeons ability to detect melanoma when interpreting frozen sections. Zitelli et al [18] studied 221 specimens from 59 patients and found they were able to detect melanoma on frozen sections with 100% sensitivity and 90% specificity. Frozen sections were compared with paraffin sections from each sample. In another study, Zitelli et al [19] reported their ability to achieve narrow margins and low recurrence rates in 95% of 535 consecutive patients with melanoma treated by MMS. They state that their overall low recurrence rate (0.5%) and ability to achieve narrow margin excision (6 mm) in 83% of patients support their ability to detect atypical melanocytes on frozen sections. The authors caution that both experience and an excellent frozen section laboratory are necessary to visualize melanomas accurately. Others caution against using frozen section in the diagnosis of melanoma because of decreased ability to detect melanocytes with adequate sensitivity. Nield et al [20] report significant differences in tumor thickness values for melanoma obtained by frozen and paraffin sections. In their study, thickness was either overestimated or underestimated on frozen sections. In another study, Shafir et al [21] found that in 29 consecutive cutaneous melanomas, thickness measured on frozen section was 0.1 to 0.4 mm greater than that measured on paraffin-embedded sections. They further caution that frozen section diagnosis of regressing melanoma is difficult and is accomplished best with paraffin-embedded specimens.
3 mm are usually obtained. Frozen sections are processed and slides are examined by the Mohs surgeon. Remaining tumor is then removed in a directed fashion with margins of 3 mm in subsequent stages. Mohs surgery offers the advantage of examination of virtually 100% of the peripheral margin rather than sampling achieved by examination of conventional sections. MMS offers optimal patient convenience because tumors can be excised and repairs performed in a single day when the freshtissue technique is used. In contrast, MMS was originally performed by a fixed-tissue technique (chemosurgery) whereby tumors were fixed overnight with zinc chloride paste (Fig. 2). Initial tumor excision and excision of subsequent stages took place over the course of days. In these cases, wounds were managed by granulation or delayed reconstruction. Mohs [24] first reported the use of chemosurgery (fixed-tissue technique) for melanoma in 1950. In a study of 20 patients, he reported a 35% 5-year survival rate. He later reported a 50% 5-year survival rate in a larger cohort in 1956 [25]. In 1977, Mohs [26] reported three local recurrences in 103 cases treated using the fixed-tissue technique. In this report, more than 80% of melanomas treated by Mohs were Clarks level IV and V. Zitelli et al [22] in 1989 reported on 200 cases of melanomas with an overall cure rate of 65%. Melanomas treated in this study also were advanced with almost one third level V. Snow et al [27] reviewed the treatment results of 179 cases from the Mohs melanoma registry. There were 61 cases treated by the fresh-tissue technique and 113 cases treated by a hybrid fixedtissue technique. Data were compared based on treatment technique versus tumor invasiveness and
MMS for melanoma Mohs micrographic surgery is gaining popularity as a treatment for melanoma [19,22,23]. For the first stage, the tumor with a margin of clinically uninvolved tissue is excised, inked, and mapped by the Mohs surgeon. For melanoma, initial margins of
Fig. 2. In the fixed-tissue technique, zinc chloride paste is applied overnight before Mohs excision. (Courtesy Dr. Stuart Zweibel, Mt. Kisco, New York.)
704
5-year survival data were concordant regardless of treatment type. A landmark study of MMS, fresh-tissue technique, for the treatment of primary cutaneous melanoma was performed by Zitelli et al [19]. In this study, 535 patients with 553 primary melanomas were treated by fresh-tissue MMS. Metastatic rates and 5-year survival rates were equivalent to or better than historic controls treated by standard wide-margin surgery. Most tumors (83%) were excised with margins of 6 mm. Local recurrences caused by inadequate excision were rare (0.05%). The authors state that the ability to achieve excision with narrow margins with a low rate of local recurrence validates the ability to detect melanoma on frozen sections with adequate sensitivity and specificity.
the sensitivity of detection of melanoma by modifying the Mohs technique.
Modified Mohs surgery Special histologic stains Special histologic stains have been used in an effort to increase sensitivity and specificity of detection of malignant melanocytes at the margins of LM. Commonly used special stains have included S-100, Melan-A, HMB-45, and Mel-5 (Table 3). S-100, an acidic Ca2 + and Zn2 + binding protein found in the cytoplasm and nucleoplasm, may be involved in the regulation of diffusion of monovalent cations across membranes [31]. S-100 is sensitive for melanoma and is particularly useful in detection of spindle cell or desmoplastic melanoma [32]. Although sensitive for the dermal component of melanoma, S-100 may not be strongly positive in the epidermis. In addition to melanoma, S-100 stains benign melanocytic lesions, dendritic cells, histiocytes, Schwann cell, muscle, chondrocytes, and eccrine and apocrine cells [31,32].
Melanoma margins revisited In a related study by Zitelli et al [18], the margins of excision required to clear 553 melanomas cleared by MMS were determined. The authors found that margins of 6 mm were required in 83% of melanomas, 9-mm margins were required to remove 95% of melanomas, and 1.2-cm margins were required to remove 97% of melanomas. The authors concluded that predetermined margins of excision should include 1 cm of normal skin for melanomas on the trunk and proximal extremities less than 2 cm in diameter, or a 1.5-cm margin for lesions larger than 2 cm in diameter. They recommended margins of 1.5 cm for melanomas of the head, neck, hands, and feet less than 3 cm in diameter and a margin of 2.5 cm for lesions larger than 3 cm in diameter. Robinson [28], using Mohs surgery with both frozen and paraffin sections, showed that margins less than 6 mm removed only 23% of 16 LM and that margins up to 1.3 cm were required to remove all 16 tumors. Cohen et al [29] excised 45 LMs and lentigo maligna melanomas (LMM) by Mohs surgery and showed that margins of 6 to 10 mm were required to clear all tumors. In another study by Zalla et al [30], the average margin required for clearance was 8.3 mm for in situ melanoma and 11.1 mm for invasive melanoma. In contrast to Zitelli et al [19], only 50% of patients were clear with margins of 6 mm. Although 2.6 cm were required to clear 95% of all tumors, 50% of invasive tumors were cleared with margins of 6 mm. Accurate assessment of subclinical extension in these studies was based on the ability of the Mohs surgeon to assess the presence of melanoma cells on frozen sections. Attempts have been made to increase
Table 3 Special histologic stains used to enhance melanoma detection Histologic stain Antigen S-100 Acidic Ca and Zn2 + binding protein
2+
Positive staining
Mel-5
HMB-45
Melan-A
Desmoplastic melanoma Dendritic cells Muscle cells Histiocytes Schwann cells Chondrocytes gp75 (melanocyte Melanoma glycoprotein) Benign nevi Pigmented Bowens disease Pigmented actinic keratosis Lentigo Lichen planus like keratosis 30 35 kd Melanoma melanosomeClear cell tumor associated Angiolipoma glycoprotein Lymphangiomyomatosis MART-1 (22-kd Melanoma cytoplasmic Clear cell tumor melanosome Angiolipoma associated Lymphangiomyomatosis glycoprotein)
705
The Melan-A antigen is a 22-kd, cytoplasmic melanosome-associated melanocyte differentiation antigen and is present in 80% to 100% of melanomas [33,34]. Homogeneous staining was reported in up to 72% of tumors with Melan-A [33]. Zalla et al [30] found that Melan-A was the most consistently crisp stain when compared with S-100, Mel-5, and HMB-45 in the treatment of melanoma. HMB-45 is a mouse monoclonal antibody that recognizes a 30- to 35-kd, melanosome-associated sialated glycoprotein [35,36] that is localized to stage II and III melanosomes in neoplastic melanocytes. Although HMB-45 has been reported to be up to 97% sensitive for melanoma on paraffin sections, it is often negative on spindle cell, desmoplastic, or neurotropic melanomas [32]. Mel-5 recognizes gp75, an abundant glycoprotein in melanocytes [37]. Although staining for melanoma has been reported to be brisk, Mel-5 stains nonmelanocytic lesions including lichen planus like keratoses, pigmented actinic keratoses, and pigmented Bowens disease. Gross et al [38] report successful treatment of two patients with LM using a rapid (75-minute) Mel-5 staining technique. Both patients remained tumor free at 15 and 18 months. In that study, the authors used a negative control from contralateral skin and a positive control from the center of the LM. Zalla et al [30] report successful treatment of melanoma using Melan-A, HMB-45, Mel-5, and S-100. In that study, 68 patients were treated including 46 with melanoma in situ and 22 with invasive melanoma. In their study, Melan-A was reported to be the most consistently crisp and easily interpretable immunostain. Menaker et al [39] recently described a rapid HMB-45 technique for in situ and invasive melanoma. In their study of 20 patients, HMB-45 was 100% sensitive and 95% specific for detecting melanoma when compared with horizontally sectioned paraffin-embedded specimens.
Special situations Lentigo maligna Lentigo maligna is an indolent form of melanoma in situ that occurs predominantly on sun-exposed areas of the head and neck in elderly individuals [7]. There has been controversy as to whether LM represents a precancerous lesion rather than a skin cancer; however, the threat of progression to invasive LMM exists [42]. Because studies have shown that prognosis of LMM is dependent on depth of invasion as with other subtypes of invasive melanoma, detection and treatment of LM before progression to invasive disease are necessary [43]. Surgical excision with clear margins is the treatment of choice for LM [7]. Both conventional and MMS excision have been used successfully in the treatment of LM. Cohen et al [40] reported high cure rates for LM using MMS with rush paraffin sections. Gross et al [38] reported successful treatment of LM using MEL-5 immunostain to increase sensitivity. Stonecipher et al [44] used rush paraffin sections stained with HMB-45 for treatment of two cases of LM and three cases of LMM. Melanoma of the hands and feet Melanoma involving the nail unit tends to have poorer prognosis with 5-year survival rates for stage I and II disease ranging between 38% and 61% [45 47]. This is likely caused by delays in diagnosis and adequate treatment. Tseng et al [48] reported 48 patients with acral lentiginous melanoma (ALM) in which 71% had a depth of 1.5 mm or greater. Metastases were present in 56%. The authors recommended excision to bone followed by grafting for lesions less than 1 mm in depth and amputation of the distal phalanx and sentinel lymph node biopsy for lesions greater than 1 mm in depth. Fourteen patients with nail apparatus melanoma were treated with MMS by Brodland [49]. After a mean follow-up period of 7.7 years 6 patients were alive and disease free, 3 were alive and disease free after local recurrence, 2 had died disease free of other causes, 2 died from widespread disease, and 1 was disease free after lymph node dissection and interferon-alfa for regional disease. The author proposed that risk of local recurrence with little effect on survival was acceptable when compared with disability secondary to amputation. Banfield et al [50] report treatment of nail apparatus melanoma by MMS using paraffin sections stained with hematoxylin and eosin, Schmorls, and ammoniacal silver nitrate and
Rush paraffin sections Permanent sections have been used to aid recognition of malignant melanocytes in LM and LMM. Cohen [40] reports successful treatment in 45 patients with LM and LMM over a mean follow-up period of 29.2 months. In a follow-up study, Cohen et al [41] reported 1 recurrence in 45 patients over a follow-up period of 50 months for an overall cure rate of 97%. As with other types of staged procedures, extirpation and repair take place over several days.
706
immunostaining with HMB-45 and S-100. Their patient was tumor free at 1 year. Melanoma involving the foot can be a therapeutic challenge. Griego and Zitelli [51] report successful treatment of recurrent acral melanoma using HMB-45 immunostaining. In their report, the melanoma recurred 6 years after standard excision with 9 of 12 specimens positive for tumor. The recurrent tumor failed MMS and the fourth recurrence was treated with the modified technique using HMB-45. In that case the authors found that HMB-45 was positive in areas known to be positive by frozen and paraffin sections and in nearby areas where permanent sections were equivocal or negative. Their patient remained tumor free at 22 months. Melanoma of the oral cavity Melanoma of the oral cavity, although rare (0.2% of all melanomas), can be devastating with 5-year survival rates ranging from 20% to 35% [52]. Oriba et al [53] report treatment of advanced melanoma using the fixed-issue technique. Their patient was a 77-year-old man with a 6-cm nodular melanoma of the lip and buccal mucosa that extended to a depth of 4 mm and Clarks level V. In addition regional metastases were discovered at presentation. After a neck dissection was performed, fixed-tissue Mohs surgery was performed to debulk the primary tumor. The primary tumor was cleared, the wound healed secondarily, and the patient was subsequently treated with local radiation. The patient eventually died of systemic metastatic disease; however, the tumor did not recur locally. The authors proposed that tissuesparing chemosurgery allowed for increased quality of life as compared with radical surgery that would have otherwise been performed. Melanoma of the nose Melanoma on the nose was reported to have an incidence of 0.4% with desmoplastic melanoma being the predominant subtype [54]. Excision to cartilage followed by flap or graft repair has been recommended [3]. Mohs et al [55] reported a series of 10 patients with melanoma of the nose treated by MMS, fixedtissue technique. There were three patients who developed widespread metastatic disease over 5 to 7 years, although no local recurrences were recorded. Two of these three cases were nodular melanoma, Clarks level V, and the other was an amelanotic, desmoplastic LMM, Clarks level V. The authors reported a 62.5% 5-year survival rate for eight deter-
minate cases. Follow-up ranged from 5 to 10 years for successfully treated patients. Melanoma of the ear Melanoma of the ear occurs most frequently on the helix [3]. Cole et al [56] report successful treatment by excision of skin with cartilage sparing followed by graft repair. Others advocate wedge excision for melanomas thicker than 1 mm [3]. Mohs [57] reported a series of 17 patients with auricular melanoma treated by MMS, fixed-tissue technique. Four patients developed distant metastatic disease. Twelve (75%) of 16 patients were disease free at 5 years. Of the 12 successfully treated, 5 patients had melanomas invasive to Clarks level IV or V. Periorbital melanoma Periorbital melanoma accounts for up to 7% of head and neck melanomas [58]. Because of the thinness of the skin and high likelihood of subdermal invasion, wedge excision of lesions on the lash margin has been suggested with narrower margins reserved for lesions involving eyelid skin [58]. Mohs [59] used both the fixed- and fresh-tissue technique in the treatment of periorbital melanoma. The fixed-tissue technique was used when there was no involvement of the lid margins or bulbar conjunctiva. A modified fresh-tissue technique was used when these structures were invaded. Mohs reported on five cases of periorbital melanoma, three of which were invasive to Clarks level III. There were no recurrences in four of five cases over a follow-up period ranging from 4 to 36 years. In one case, there were both local and regional metastases over 2 years.
Summary Surgical excision with clear margins remains the treatment of choice for melanoma. Under appropriate conditions MMS may offer a means of achieving clear margins while sparing normal tissue. This requires a Mohs surgeon expert in interpretation of melanocytic lesions and excellent technical support. As the popularity of special histologic stains continues to grow these may offer and additional means to increase sensitivity and specificity in the detection of melanoma.
707
References
[1] Rigel DS, Friedman RJ, Kopf AW. The incidence of malignant melanoma in the United States: issues as we approach the 21st century. J Am Acad Dermatol 1996; 34:839 47. [2] Davidowitz S, Belafsky PC, Amedee RG. The epidemiology of malignant melanoma in Louisiana and beyond. J La State Med Soc 1999;151:493 9. [3] Naryan D, Ariyan S. Surgical management of primary melanoma. Clin Plast Surg 2000;27:409 19. [4] Kaspar TA, Wagner Jr. RF. Mohs micrographic surgery for thin stage I malignant melanoma: rationale for a modern management strategy. Cutis 1992;50: 350 1. [5] Rigel D, Carucci JA. Malignant melanoma: prevention, early detection, and treatment in the 21st century. CA Cancer J Clin 2000;50:215 36. [6] Breslow A. Thickness, cross sectional area, and depth of invasion in the prognosis of primary melanoma. Ann Surg 1970;172:902 8. [7] Cohen LM. Lentigo maligna and lentigo maligna melanoma. J Am Acad Dermatol 1995;33:923 36. [8] Cascinelli N. Margin of resection in the management of primary melanoma. Semin Surg Oncol 1998;14: 272 5. [9] Harrist T, Rigel DS, Day CL, et al. Microscopic satellites are more highly associated with regional lymph node metastases than is primary melanoma thickness. Cancer 1986;53:2183 7. [10] Abide J, Nahai F, Bennett RG. The meaning of surgical margins. Plast Reconstr Surg 1984;73:492 6. [11] Lawrence CM. Mohs micrographic surgery for basal cell carcinoma. Clin Exp Dermatol 1999;24:130 3. [12] Bernstein PE. Mohs 98: single-procedure Mohs surgery with immediate reconstruction. Otolaryngol Head Neck Surg 1999;120:184 9. [13] Shriner DL, McCoy DK, Goldberg DJ, Wagner Jr. RF. Mohs micrographic surgery. J Am Acad Dermatol 1998;39:79 97. [14] Arlette JP, Carruthers A, Threlfall WJ, Warshawski LM. Basal cell carcinoma of the periocular region. J Cutan Med Surg 1998;2:205 8. [15] Kumar B, Roden D, Vinciullo C, Elliott T. A review of 24 cases of Mohs surgery and ophthalmic plastic reconstruction. Aust N Z J Ophthalmol 1997;25: 289 93. [16] Nelson BR, Railan D, Cohen S. Mohs micrographic surgery for nonmelanoma skin cancers. Clin Plast Surg 1997;24:705 18. [17] Skouge JW. Mohs micrographic surgery for the treatment of difficult skin cancers. Md Med J 1997;46: 231 7. [18] Zitelli JA, Brown CD, Hanusa BH. Surgical margins for excision of primary cutaneous melanoma. J Am Acad Dermatol 1997;37:422 9. [19] Zitelli JA, Brown C, Hanusa BH. Mohs micrographic surgery for the treatment of primary cutaneous melanoma. J Am Acad Dermatol 1997;37:236 45.
[20] Nield DV, Saad MN, Khoo CT, Lott M, Ali MH. Tumour thickness in malignant melanoma: the limitations of frozen section. Br J Plast Surg 1988;41:403 7. [21] Shafir R, Hiss J, Tsur H, Bubbis JJ. Pitfalls in the frozen section diagnosis of malignant melanoma. Cancer 1983;51:1168 70. [22] Zitelli JA, Mohs FE, Larson P, Snow S. Mohs micrographic surgery for melanoma. Dermatol Clin 1989;7: 833 43. [23] Olhoffer IH, Bolognia JL. Whats new in the treatment of cutaneous melanoma? Semin Cutan Med Surg 1998;17:96 107. [24] Mohs FE. Chemosurgical treatment of melanoma: a microscopically controlled method of excision. Arch Dermatol 1950;62:269 79. [25] Mohs F. Chemosurgery in cancer, gangrene and infections: featuring a new method for the microscopically controlled excision of cancer. Springfield, IL: Thomas; 1956. p. 168 78. [26] Mohs FE. Chemosurgery for melanoma. Arch Dermatol 1977;113:285 91. [27] Snow SN, Mohs FE, Oriba HA, Dudley CM, Leverson G, Hetzer M. Cutaneous malignant melanoma treated by Mohs surgery: review of the treatment results of 179 cases from the Mohs Melanoma Registry. Dermatol Surg 1997;23:1055 60. [28] Robinson JK. Margin control for lentigo maligna. J Am Acad Dermatol 1994;31:79 85. [29] Cohen LM, McCall MW, Hodge SJ, Freedman JD, Callen JP, Zax RH. Successful treatment of lentigo maligna and lentigo maligna melanoma with Mohs micrographic surgery aided by rush permanent sections. Cancer 1994;73:2964 70. [30] Zalla MJ, Lim KK, Dicaudo DJ, Gagnot MM. Mohs micrographic excision of melanoma using immunostains. Dermatol Surg 2000;26:771 84. [31] Donato R. S-100 proteins. Cell Calcium 1986;7: 123 45. [32] Blessing K, Sanders DSA, Grant JJH. Comparison of immunohistochemical staining of the novel antibody Melan-A with S-100 protein and HMB-45 in malignant melanoma and melanoma variants. Histopathology 1998;32:139 46. [33] Jungbluth A, Busam KJ, Gerald WL, et al. A103: an anti-Melan-A monoclonal antibody for the detection of malignant melanoma in paraffin embedded tissue. Am J Surg Pathol 1998;22:595 602. [34] Busam K, Chen YT, Old LJ, et al. Expression of Melan-A (MART-1) in benign melanocytic nevi and primary cutaneous melanoma. Am J Surg Pathol 1998; 22:976 82. [35] Bacchi C, Bonetti F, Pea M, et al. HMB-45: a review. Appl Immunohistochem 1996;4:73 85. [36] Chiamenti A, Vella F, Bonetti F, et al. Anti-melanoma antibody HMB-45 recognizes a 30 35 kDa melanosome associated sialated glycoprotein. Melanoma Res 1996;6:291 8. [37] Bhawan J. MEL-5: a novel antibody for differential diagnosis of epidermal pigmented lesions of the skin
708
J.A. Carucci / Dermatol Clin 20 (2002) 701708 in paraffin embedded sections. Melanoma Res 1997;7: 43 8. Gross EA, Andersen WK, Rogers GS. Mohs micrographic excision of lentigo maligna using Mel-5 for margin control. Arch Dermatol 1999;135:15 7. Menaker G, Chiang JK, Tabila B, Moy RL. Rapid HMB-45 staining in Mohs micrographic surgery for melanoma in situ. J Am Acad Dermatol 2001;44: 833 6. Cohen LM. Management of lentigo maligna and lentigo maligna melanoma with tangential paraffin-embedded sections. J Am Acad Dermatol 1994;31:691 2. Cohen LM, McCall MW, Zax RH. Mohs micrographic surgery for lentigo maligna and lentigo maligna melanoma: a follow-up study. Dermatol Surg 1998;24: 673 7. Albert LS, Fewkes J, Sober AJ. Metastatic lentigo maligna melanoma. J Dermatol Surg Oncol 1990;16:56 8. Ringborg U, Afzelius LE, Lagerlof B, et al. Cutaneous malignant melanoma of the head and neck: analysis of treatment results and prognostic factors in 581 patients: a report from the Swedish Melanoma Study Group. Cancer 1993;71:751 8. Stonecipher MR, Leshin B, Patrick J, White WL. Management of lentigo maligna and lentigo maligna melanoma with paraffin-embedded tangential sections: utility of immunoperoxidase staining and supplemental vertical sections. J Am Acad Dermatol 1993;29: 589 94. OLeary JA, Berend KR, Johnson JL, Levin LS, Seisler HF. Subungual Melanoma. A review of 93 cases with identification of prognostic variables. Clin Orthop 2000;378:206 12. Muchmore J, Krementz ET, Carter RD, Sutherland CM, Godfrey RS. Regional perfusion for the treatment of subungual melanoma. Am Surg 1990:114 8. Pack G, Oropeza R. Subungual melanoma. Surg Gynecol Obstet 1967;124(3):571 82. [48] Tseng J, Tanabe KK, Gaad MA, et al. Surgical management of primary melanoma of the hands and feet. Ann Surg 1997:544 50. [49] Brodland D. The treatment of nail apparatus melanoma with Mohs micrographic surgery. Dermatol Surg 2001; 27:269 73. [50] Banfield C, Dawber RPR, Walker NPJ, Stables GI, Zeina B, Schomberg K. Mohs micrographic surgery for the treatment of in situ nail apparatus melanoma: a case report. J Am Acad Dermatol 1999;40:98 9. [51] Griego RD, Zitelli JA. Mohs micrographic surgery using HMB-45 for a recurrent acral melanoma. Dermatol Surg 1998;24:1003 6. [52] Me P. Malignant melanoma of the oral cavity. In: Clark WJ, Golman LI, Mastrangelo MJ, editors. Malignant melanoma. New York: Grune & Stratton; 1979. p. 125 7. [53] Oriba H, Stanley R, Snow SN, Mohs FE. Oral malignant melanoma treated with Mohs micrographic surgery by fixed tissue technique. Arch Otolaryngol Head Neck Surg 1998;124:199 201. [54] Padapoulos T, Rasiah K, Thompson JF, et al. Melanoma of the nose. Br J Surg 1997;84:986. [55] Mohs F, Snow SN, Larson PO. Mohs micrographic surgery fixed-tissue technique for melanoma of the nose. J Dermatol Surg Oncol 1990;16:111 1120. [56] Cole D, Mackay GJ, Walker BF, et al. Melanoma of the external ear. J Surg Oncol 1992;50:110 4. [57] Mohs F. Fixed tissue micrographic surgery for melanoma of the ear. Arch Otolaryngol Head Neck Surg 1988;114:624 31. [58] Zoltie N, ONeill TJ. Malignant melanoma of eyelid skin. Plast Reconstr Surg 1989;83:994. [59] Mohs F. Microscopically controlled surgery for periorbital melanoma: fixed tissue and fresh tissue techniques. J Dermatol Surg Oncol 1985;11:284 91.
[38]
[39]
[40]
[41]
[42] [43]
[44]
[45]
[46]
[47]
Dermatol Clin 20 (2002) 709 712
Chemotherapy approaches to melanoma

Anna C. Pavlick, MD
Division of Medical Oncology, Department of Medicine, New York University School of Medicine, Bellevue C & D, Room 556, 462 First Avenue, New York, NY 10016, USA
The activity of chemotherapeutic drugs on melanoma is limited. It is for this reason that adjuvant therapy for high-risk melanoma has focused on immunotherapy, such as interferon, interleukin, or vaccines. There are, however, several clinical trials exploring the use of chemotherapy or biochemotherapy in the setting of high-risk (stage III) patients. Chemotherapy in the treatment of melanoma is usually reserved for metastatic disease. Patients with metastatic melanoma have a very poor prognosis. Based on the new American Joint Committee on Cancer staging system, stage IV disease was further defined into three subgroups with very different survivals rates. Stage IV-A melanoma patients have metastatic disease to the skin and have a 5-year overall survival of 20%, whereas stage IV- B and C (lung and other visceral metastasis, respectively) have a 5-year overall survival of 9% [1]. To improve on this dismal prognosis, novel agents are being used in clinical trials. Table 1 provides a summary of chemotherapy trials.
Single agents Dacarbazine, an alkylating agent, is the most active single-agent chemotherapy for metastatic melanoma. It must be metabolized in the liver to the active agent, mitozolomide. Response rates average approximately 20%. Response duration is approximately 6 months. Only 2% of all patients treated with dacarbazine achieve a sustained complete remission [2].
The most common side effects of dacarbazine include nausea and vomiting, but this is usually well controlled with the aggressive use of 5-HT3 antiemetic agents. Neutropenia and thrombocytopenia are relatively rare. It is usually administered at a dose of 800 to 1000 mg/m2 once every 3 weeks. Cisplatin, a platinum compound, is reported to produce responses in 10% to 20% of patients [2]. It is commonly viewed as the standard second-line therapy for metastatic melanoma. The side effects from cisplatin include sensory-neural hearing loss or tinnitus, peripheral neuropathy, renal failure, and severe nausea and vomiting. The nitrosoureas, carmustine and lomustine, also have activity in melanomas. Responses are approximately 10% to 20% [3]. Side effects include myelosuppression, secondary malignancies, pulmonary fibrosis, and nausea and vomiting. The taxanes, both paclitaxel and docetaxel, are newer agents being investigated in the field of melanoma. As single agents, response rates have been reported to be 20% [4,5]. They are usually well tolerated. Their side effect profile includes a hypersensitivity reaction, complete alopecia, peripheral neuropathy, arthralgias, and myalgias. Myelosuppression is uncommon. Vinca alkaloids, namely vinblastine, as single agents have a 13% response rate [2]. Very few patients receive this agent alone. It is very well tolerated with no myelosuppression, but little activity.
Combination chemotherapy Combination chemotherapy has been extensively explored and may induce a faster response than single-agent dacarbazine, but with greater toxicity
E-mail address: anna.pavlick@med.nyu.edu
710
A.C. Pavlick / Dermatol Clin 20 (2002) 709712
Table 1 Chemotherapy trials for metastatic melanoma Trial Agents Single agents Combination chemotherapy Chemohormonal therapy Reference Comis [15] Buzaid et al [16] DelPrete et al [6] Berd and Mastrangelo [17] Kirkwood et al [18] Atkins et al [11] Dorval et al [19] Phase of Trial III III II II III II III Regimen DTIC CVD vs DTIC CBDT (20 mg/d) CBDT (160 mg/d 7 days prechemo) DTIC-IFN vs DTIC CDDP-DTIC with IL-2 CDDP-IFN vs CDDP-IFN, IL-2 Number of Patients 1936 91 20 147 45 38 101 Responses CR + PR (%) 20 24 vs 11 55 28 19 vs 20 42 25 vs 16
Biochemotherapy
Abbreviations: C, cisplatin; CDDP, cisplatin; D, dacarbazine; DTIC, dacarbazine; IFN, interferon-alfa; IL-Z, interleukin-Z; T, tamoxifen; V, vinblastine.
and no improvement in overall survival [2]. The two most commonly used combination regimens are a combination of cisplatin, vinblastine, and dacarbazine (CVD), and the Dartmouth regimen. CVD is administered intravenously every 3 weeks. The Dartmouth regimen combines cisplatin, dacarbazine, carmustine, and tamoxifen. Tamoxifen was included in the Dartmouth regimen, because there were data suggesting that antiestrogens may have an additive antitumor effect. Despite the impressive single-institution response rate of 40% [6], multi-institutional studies continue to reflect a response rate of only 20% [7,8].
Biochemotherapy Biochemotherapy is a combination of several chemotherapeutic agents that are administered with nonspecific immunomodulators. Recently, the combination of cytokines, namely interferon-alfa and interleukin-2, has been investigated with chemotherapy. The rationale for combining these agents was the independent activity seen in several clinical trials. Chemotherapy with interferon-alfa alone failed to demonstrate a significant improvement in overall response and survival with increased toxicity [9]. It is not recommended that interferon-alfa with chemotherapy be used outside a controlled clinical trial. Chemotherapy with interleukin-2 was also assessed and does not seem to improve responses rates or survival. Interleukin-2, however, may interact synergistically with cisplatin and multiagent regimens with this combination are currently being explored [10]. Chemotherapy in combination with interferonalfa and interleukin-2 is considered in the classic
biochemotherapy regimen [11]. The MD Anderson Cancer Center evaluated biochemotherapy given in an alternating, sequential, or concurrent manner. The alternating regimen consisted of interferon and interleukin alternating with CVD every 6 weeks. The response rate in this single-institution trial was only 33% [12]. A trial using sequential biologic agents with CVD demonstrated a 69% response rate compared with the 40% response rate of CVD alone [13]. Concurrent biochemotherapy combines all agents over 5 days in a 21-day cycle. This is more convenient and less toxic than sequential treatment, but efficacy is not yet known [14]. Combination biochemotherapy is clearly more toxic than single-agent therapy and large multi-institutional studies have not demonstrated a clear benefit.
Isolated limb perfusion Intra-arterial chemotherapy has been used for patients with in-transit metastases. This permits dose intensification to the tumor, which is not possible systemically. The most commonly used drug is melphalan. Hyperthermic limb perfusion for in-transit metastases can result in 60% to 80% overall response rates; however, less than half of these responses are complete responses, and, of this group, less than 15% are durable responses [20]. Several investigators are now exploring the use of tumor necrosis factor with chemotherapy to try to improve the durability of responses. Others are combining isolated limb perfusion with aggressive surgical resection. The value of these strategies is still being evaluated in controlled clinical trials.
A.C. Pavlick / Dermatol Clin 20 (2002) 709712
711
Novel agents in clinical trials There are several new advances in the treatment of metastatic melanoma. Some are in drug therapy, immunotherapy, and in the development of enzyme pathway inhibitors or anti-sense proteins that inhibit the transcription of mRNA into DNA. Temozolomide, an oral imidazotetrazine derivative, is a cytotoxic alkylating agent that was developed as an alternative to intravenous dacarbazine. Temozolomide was compared with dacarbazine in a randomized phase III trial in the treatment of metastatic melanoma, which demonstrated equivalent response rates [21]. Unlike dacarbazine, temozolomide penetrates the blood-brain barrier and may decrease the incidence or prevent the development of brain metastases. It may also act as a radiosensitizer with radiotherapy in the setting of brain metastases. This concept is currently being investigated. The side effects of temozolomide include mild myelosuppression and nausea. Temozolomide was combined with thalidomide in a phase I-II clinical trial. The results, although single institution and phase I-II trial, are comparable with those of combination chemotherapy, and it seems to be active also against metastatic disease to the brain [22]. Accrual in this trial is continuing. Another exciting development is the novel BCL-2 antisense protein, G-3139 (Augmerosin). Up-regulation of BCL-2 inhibits apoptosis conferring immortality to cancer cells. By inhibiting the transcription of messenger RNA into DNA, the cells lose their immortality and become exquisitely susceptible to the damaging effects of chemotherapy. In a combination phase I-II trial, 17 patients with advanced malignant melanoma were administered G-3139 followed by dacarbazine chemotherapy. Treatment was well tolerated with adverse events that included transient fever, flushing, and rash. Liver function abnormalities seen at the highest G3139 dose level were not related to the drug. Antitumor responses were reported in 6 of 14 evaluable patients, including one pathologically documented complete response (CR), two partial responses, and three minor responses (MRs). The median survival among all patients was over 7 months, and some responses extended for over 1 year [23]. A randomized phase III trial of G-3139 with dacarbazine versus dacarbazine alone is currently underway. The trial will be completed by the end of 2002. There are several other novel agents that are being explored in phase II clinical trials for metastatic melanoma. PS-341 is a proteasome inhibitor that demonstrated activity in metastatic melanoma in the phase I trials. It is currently being investigated as a
single agent and in combination therapy. The theory behind combining PS-341 with chemotherapy is to prime the cells, resulting in increased susceptibly to the cytotoxic effects of chemotherapy. Epothilone (BMS 277550) is a novel antitubulin chemotherapy agent that has shown activity in this disease. Phase I studies were completed in 2001, which not only defined the maximally tolerated dose and side effects of this drug, but also demonstrated activity in melanoma, breast, ovarian, and lung cancer. The side effects of epothilone include hypersensitivity reaction at the time of infusion, myelosuppression, neuropathy, myalgias, arthralgias, nausea, and vomiting. Epothilone is now being investigated in phase II trials. There are several other innovative treatment options also under investigation, such as novel vaccines, fusion proteins, and dendrite cell vaccines. All of these may allow for better control and cure of this deadly tumor.
Summary Melanoma, if detected early, is a curable malignancy. Unfortunately, however, many melanomas are detected at the stage that puts patients at a high risk of recurrence, if they are not already metastatic. New techniques are being developed to try to detect subclinical disease and allow for early intervention with adjuvant treatment. Because the incidence of melanoma is on the rise, there is a greater demand to improve the current therapy. Revision of the American Joint Committee on Cancer staging system has enabled patients to be stratified better into risk groups, which in turn dictates treatment recommendations. As evidenced by the data provided, the current chemotherapeutic options for metastatic melanoma are far from optimal. Clearly, in this malignancy, enrollment into a clinical trial is an excellent option for patients. The development of novel approaches to treating this disease provide hope for better management and possible cure of a notoriously aggressive disease with a dismal prognosis.
References
[1] Balch C, Buzaid A, Soong S-J, et al. Final version of the American Joint Committee on Cancer staging system for cutaneous melanoma. J Clin Oncol 2001; 19:3635 48. [2] Anderson C, Buzaid A, Legha S. Systemic treatment
712
A.C. Pavlick / Dermatol Clin 20 (2002) 709712 for advanced cutaneous melanoma. Oncology 1995; 9:1149 54. Jacquillat C, Khayat D, Banzet P, et al. Final report of the French multicenter phase II study of the nitrosourea, Fotemustine, in 153 evaluable patients with disseminated malignant melanoma including patients with cerebral metastasis. Cancer 1990;66:1872 3. Einzig A, Hochster H, Wiernik P, et al. A phase II study of taxol in patients with malignant melanoma. Invest New Drugs 1991;9:59 63. Bedikian A, Weiss G, Legha S, et al. Phase II trial of docetaxel in patients with advanced cutaneous melanoma previously untreated with chemotherapy. J Clin Oncol 1995;13:2895 8. DelPrete S, Maurer L, ODonnell J, et al. Combination chemotherapy with cisplatin, carmustine, DTIC and tamoxifen in metastatic melanoma. Cancer Treat Rep 1984;68:1403 7. Buzaid A, Murren J, Durerage H. High-dose cisplatin with DTIC and tamoxifen in the treatment of metastatic melanoma. Cancer 1991;68:1238 41. Flaherty L, Liu P, Daniels D, et al. The addition of tamoxifen to DTIC and cisplatin in a SWOG phase II trial in metastatic malignant melanoma Stage IV. Proc Am Soc Clin Oncol 1993;12:394. Morton R, Creagen E, Schaid D, et al. Phase II trial of recombinant leukocyte A interferon (IFN-alpha-2a) plus 1,3-bis (2-chloroethyl)-1-nitrosurea (BCNU) and the combination cimetidine with BCNU in patients with disseminated malignant melanoma. Am J Clin Oncol 1991;14:152 6. Flaherty L, Liu P, Fletcher W, et al. DTIC and outpatient Il-2 in the treatment of metastatic malignant melanoma: a phase II SWOG trial. J Natl Cancer Inst 1992;84:893 8. Atkins M, OBoyle K, Sosman J, et al. Multi-institutional phase II trial of intensive chemoimmunotherapy for metastatic melanoma. J Clin Oncol 1994;12: 1553 60. Legha S, Renz S, Plager C, et al. Biochemotherapy using Il-2 and IFN in combination with cisplatin, vinblastine and DTIC in advanced melanoma. Proceedings of the American Society of Clinical Oncology 1991;10:293. Legha S, Renz S, Bedekian A, et al. Treatment of metastatic melanoma with combined chemotherapy containing cisplatin, vinblastine and DTIC (CVD) and biochemotherapy using Il-2 and IFN. Ann Oncol 1996;7:827 33. Legha S, Buzaid A, Bedekain A, et al. Efficacy of concurrent biochemotherapy using Il-2 and IFN and chemotherapy with cisplatin, vinblastine and DTIC in metastatic melanoma. Proceedings of the American Association of Clinical Research 1994;35:87. Comis RL. DTIC (NSC-45388) in malignant melanoma: a perspective. Cancer Treat Res 1976;64:1123 7. Buzaid A, Legha S, Winn R, et al. Cisplatin, vinblastine and dacarbazine versus dacarbazine alone in metastatic melanoma: preliminary results of a phase III Cancer Community Oncology Program (CCOP) trial. Proc Am Soc Clin Oncol 1993;12:389. Berd D, Mastrangelo M. Combination chemotherapy of metastatic melanoma. J Clin Oncol 1995;13:796 7. Kirkwood J, Ernstroff M, Giuliano A, et al. Interferon alpha-2a and dacarbazine in melanoma. J Natl Cancer Inst 1990;82:1062 70. Dorval T, Negrier S, Chevreau C, et al. Results of a French multicentric randomized trial of chemoimmunotherapy (cisplatin, IL-2, with or without IFN) in metastatic malignant melanoma. Proc Am Soc Clin Oncol 1994;13:395. Franker DL. Hyperthermic regional perfusion for melanoma of the limbs. In: Balch CM, Houghton AN, Sober AJ, Soong S, editors. Cutaneous melanoma. St. Louis, MI: Quality Medical Publishing Inc; 1998. p. 281 300. Middleton MR, Grob JJ, Aaronson N, et al. Randomized phase III study of temozolomide versus dacarbazine in the treatment of patients with advanced melanoma. J Clin Oncol 2000;18:158 66. Hwu W, Panageas K, Krown S, et al. Phase I study of temozolomide and thalidomide in the treatment of metastatic melanoma. Proc Am Soc Clin Oncol 2001; 20:1424. Jansen B, et al. Systemic treatment with bcl-2 antisense down-regulates tumor content of bcl-2 protein and induces major antitumor responses in patients with melanoma when combined with dacarbazine [abstract]. In: Proceedings of the Annual Meeting of the American Association of Cancer Research 2000;41:23.
[3]
[14]
[4]
[15] [16]
[5]
[6]
[17] [18]
[7]
[8]
[19]
[9]
[20]
[10]
[21]
[11]
[22]
[12]
[23]
[13]
Dermatol Clin 20 (2002) 713 716
Radiation therapy of malignant melanoma

Jay S. Cooper, MD
Radiation Oncology, New York University Medical Center, 560 First Avenue, New York, NY 10016, USA
For many years, malignant melanomas were considered unresponsive to radiation therapy. Although modern data clearly show this not to be true, irradiated melanomas often regress more slowly than do other types of tumors, leading to an incorrect assessment of the effects of radiation if measured too soon.
Palliative therapy of macroscopic disease The benefit of radiation therapy for malignant melanoma is most easily observed in the palliative management of metastatic lesions, both because relatively little shrinkage of the tumor is needed to provide symptomatic relief (and is relatively easy to detect) and because such treatment is the most common use of radiation therapy for malignant melanoma. Approximately one in four treated patients experience excellent relief (complete response); approximately one in three experience partial relief; and the rest do not benefit clinically (little or no relief). Placed in the perspective of other uses of radiation therapy, the subjective response of irradiated malignant melanomas that have metastasized, for example, to the brain or bone is virtually identical to the response observed following treatment of other types of disseminated neoplasms that are generally regarded as radiosensitive [1 3]. Precisely why some tumors respond well, whereas others do not, has intrigued many scientists and has led to considerable debate about the existence of a biologic factor that potentially could be exploited to increase the likelihood of response. The degree of damage caused by radiation is a complex function of the way that radiation is deliv-
E-mail address: jay.cooper@med.nyu.edu
ered; it is not merely a reflection of the total dose, but is influenced by the amount in each fraction and the time interval between fractions. Moreover, different tissues, both normal and malignant, have their varying inherent responses to differences in patterns of dose delivery. Approximately 25 years ago, investigators observed that some animal and human melanomas grown in vitro could withstand (better than other types of tumors), and subsequently repair (if given sufficient time), the amount of damage produced by one typical (conventional) single fraction of radiation therapy [4,5]. Consequently, some hypothesized that repackaged radiation therapy, delivering a particularly high-dose-per-fraction, might overcome the inherent repair capacity of melanomas and lead to their control. Three independent clinical series [6 8] seemed to confirm this hypothesis. These series reported a direct correlation of response with the size of the dose delivered in each fraction of treatment and, in contrast to the effect typically observed with other types of tumors, no correlation with the total dose delivered. Other investigators [9 11], however, repported that conventionally irradiated malignant melanomas responded equally well or that highdose-per-fraction techniques were superior to conventional techniques for the production of short-term tumor responses, but inferior to conventional techniques for the production of long-term local control. Between 1983 and 1988, the Radiation Therapy Oncology Group designed and conducted a prospective randomized trial [12] comparing 800 cGy delivered four times at weekly intervals (high-doseper-fraction therapy) with 250 cGy daily for 20 fractions over 26 to 28 days (conventional-doseper-fraction therapy) for malignant melanoma. A mathematic model [13] predicts that these two very different-appearing radiation therapy regimens produce approximately equivalent damage in normal
714
J.S. Cooper / Dermatol Clin 20 (2002) 713716
tissues. The results showed no difference in response rates between tumors treated with high-dose-per-fraction techniques and those treated with conventional techniques. There was a 24.2% complete response rate (plus a 35.5% partial response rate) for lesions treated with high-dose-per-fraction radiation versus a 23.4% complete response (plus a 34.4% partial response rate) following conventional radiation. Although this trial failed to confirm the hypothesis that justified it, this trial provides irrefutable evidence that malignant melanomas do respond to radiation therapy and clearly demonstrates that complete response can be expected approximately 25% of the time.
elderly patients treated for lentigo malignas that were characterized as larger lesions located in the head and neck area. Such therapy, however, has not gained wide acceptance and questions remain about the potentially curative role of radiation therapy in the setting of small, visibly detectable disease.
Potentially curative treatment of microscopic disease Moving down one order of magnitude, radiation therapy has somewhat more clearly been shown to have beneficial effects when applied to microscopicsize disease. It has long been accepted that surgical resection of all gross evidence of one or more intracerebral metastases typically leaves behind small amounts of tumor that are subclinical (ie, too small to be detected by current clinical means). For such patients, radiation therapy clearly has been shown to kill at least some of the remaining tumor cells. Skibber et al [21] reported 34 patients who had clinically solitary, intracerebral, metastatic malignant melanomas that were macroscopically completely excised. Twelve were treated by surgery alone and 22 were treated by surgery and elective postoperative irradiation. Intracerebral recurrence decreased from 75% (9 of 12) in patients who did not receive radiation therapy to 23% (5 of 22) in patients who did. This, in turn, influenced survival; median survival was tripled in the group receiving radiation therapy from 6 months to 18 months ( P = .002). Similarly, Hagen et al [22] reported the outcome following resection alone of clinically single intracerebral metastatic malignant melanomas (16 patients) or in combination with postoperative radiation therapy (19 patients). The median time to recurrence of central nervous system disease was extended from 6 months following surgery alone to 27 months by the addition of radiation therapy, and 85% of patients treated by surgery alone died because of their central nervous system disease versus only 24% of patients who also received radiation therapy. Most recently, elective irradiation has been considered in other situations wherein a high-risk for recurrence remains (eg, following resection of aggressive primary tumors, resection of locally recurrent disease, or resection of metastases in regional lymph nodes) despite expertly done surgery. For example, OBrien et al [23] reported a 24% local recurrence rate following surgery of cutaneous melanomas 4-mm or more thick and a 34% recurrence rate following a therapeutic neck dissection [24]. Byers [25] reported a 46% regional failure rate following modified cervical
Potentially curative radiation therapy of macroscopic disease Having established that malignant melanomas are not radioresistant, it is appropriate to ask if there are any that radiation therapy can eradicate. At least in theory, the thinnest tumors (which likely contain the fewest number of malignant cells), the noninvasive lentigo malignas, should be the best candidates to demonstrate the potential of curative-intent radiation therapy. Several physicians have reported 95% to 100% control of lentigo malignas by using extremely large doses of relatively nonpenetrating radiation [14 16]. Such cure may be obtained by indiscriminate desquamation, however, rather than by selective killing of the tumor cells. More convincing evidence comes from Harwood and Cummings [17] who treated both lentigo malignas and lentigo maligna melanomas with x-ray beams ranging from 100 to 280 kV (approximately the voltage used for diagnostic-quality radiographs and slightly higher) and doses ranging from 3250 cGy in five fractions over 1 week (for lesions less than 2 cm in diameter) to 4500 cGy in 10 fractions over 2 weeks (for lesions 2 to 5 cm in diameter). With follow-up ranging from 1.5 to 13 years, 15 of 17 lentigo malignas and 21 of 23 lentigo maligna melanomas were controlled by a single course of radiation therapy. Others have reported similar results. Christie and Tiver [18] treated seven lentigo malignas with 100-kV therapy and relatively conventional doses (4400 to 5750 cGy in 11 to 23 fractions over 15 to 34 days) and observed local control in all patients with follow-up ranging from 8 to 37 months. Panizzon [19] treated 104 lentigo malignas with superficial radiation therapy and reported a 100% cure rate after 7.5 years of follow-up; similarly, 18 lentigo maligna melanomas exhibited an 89% cure rate. Tsang et al [20] observed a 5-year actuarial tumor control rate of 86% in 36
J.S. Cooper / Dermatol Clin 20 (2002) 713716
715
neck dissection. If only a solitary node less than 3 cm was involved the local recurrence rate was 22%. If there were either multiple involved nodes or extracapsular extension present, however, the local-regional recurrence rate was 50%. Lee et al [26] reported that resection of histologically involved lymph nodes resulted in a 30% nodal basin recurrence rate. Furthermore, some factors were associated with particularly high risks of recurrence: extracapsular extension of disease (63%); multiple nodes invaded by tumor (46% for 4 to 10 nodes and 63% for more than 10 nodes); and cervical lymph node involvement (43%). It is now becoming clear that the risk of recurrence in such situations can be reduced substantially by radiation therapy. The largest series of patients treated electively because of high-risk malignant melanomas comes from the MD Anderson Cancer Center in Houston, Texas [27,28]. Their definition of high risk included malignant melanomas at least 1.5 mm thick (79 patients); clinically detectable regional nodal disease (32 patients); and recurrent previously resected disease (63 patients). Treatment was designed to deliver 6 Gy per fraction for a total of five fractions over 2.5 weeks. Although this high-dose-per-fraction technique was designed to exploit the inherent radiobiologic behavior of some melanomas (as discussed previously), it has the additional convenience of relatively few treatments that some patients who are ambivalent about taking an elective therapy find appealing. After a median follow-up of 35 months only 6 patients experienced isolated local-regional recurrence, 9 patients experienced both local-regional and distant recurrence, and 58 patients developed solely distant metastatic disease. Overall, the 5-year actuarial local-regional control rate was 88%, which far exceeded the historic experience at the same institution. As additional evidence of this principle, OBrien et al [29] reported a nonrandomized comparison of patients who had high-risk malignant melanomas that were treated with or without elective radiation therapy in Sidney, Australia. Forty-five high-risk patients (65% having at least two involved lymph nodes and 48% having extracapsular extension of disease) received six fractions of 5.5 Gy each. Local-regional failure occurred in only 6.5% and all failures were evident within 1 year of treatment. Unfortunately, only 40% survived for at least 5 years. In comparison, 107 other patients (40% having two or more involved nodes and 19% having extracapsular extension [ie, a relatively more favorable group]) received no radiation therapy and experienced an 18.7% local-regional failure rate. This group had a 35% 5-year survival rate. Corry et al [30] reported the outcome of 42 patients who received more conventionally fractionated
elective radiation therapy for high-risk malignant melanomas in Melbourne, Australia, most often 50 to 60 Gy delivered in 2-Gy-per-day fractions. With 5 years of follow-up, the first site of treatment failure was nodal recurrence in 20% and an additional 2% experienced simultaneous nodal and distant recurrence of disease. Sadly, the overall 5-year survival rate was only 33% because 52% of these high-risk patients developed distant metastases, despite remaining free of local-regional disease. The author recently reported [31] his experience in using high-dose-per-fraction elective radiation therapy for high-risk malignant melanomas. Despite the aggressive nature of the tumors treated (multiple involved lymph nodes [21 patients]; close or microscopically involved surgical margins [9 patients]; extracapsular extension [6 patients]; previously resected, recurrent disease [3 patients]; or primary tumors more than 4 mm thick [4 patients]) the patients were provided with an 84% local-regional control rate at 5 years. Unfortunately, the patients (like those of Ang et al [27,28], OBrien et al [23,24,29], and Corry et al [30]) failed to have a commensurate improvement in survival, only 39% living 5 years, suggesting that distant metastases are shed from the high-risk sites before they are irradiated. Perhaps most remarkable is the similarity in outcome in these four worldwide reports. The localregional recurrence rates, the distant metastasis rates, and the 5-year survival rates fall within a 10% range in all reports, firmly suggesting that the reported observations are universally applicable. When the results of elective radiation therapy are grouped with the results of the randomized Radiation Therapy Oncology Group study, it becomes evident that (1) malignant melanomas are not impervious to radiation therapy, and that (2) radiation therapy can be beneficial, particularly when only small volumes of disease need to be treated. Clearly, more effective systemic agents are needed to improve overall survival rates. It is likely that when systemic therapy improves, however, the importance of local-regional control may increase and the value of elective radiation therapy will become more evident. Until then, radiation therapy has a relatively small, but potentially important, role to play in the management of aggressive malignant melanomas.
References
[1] Carella RJ, Gelber R, Hendrickson F, et al. Value of radiation therapy in the management of patient with cerebral metastases from malignant melanoma;
716
J.S. Cooper / Dermatol Clin 20 (2002) 713716 R.T.O.G. Brain Metastases Study I & II. Cancer 1980; 45:679 83. Hilaris BS, Raben M, Calabrese AS, et al. Value of radiation therapy for distant metastases from malignant melanoma. Cancer 1963;16:765 73. Katz HR. The results of different fractionation schemes in the palliative irradiation of metastatic melanoma. Int J Radiat Oncol Biol Phys 1981;7:907 11. Barranco SC, Romsdahl MM, Humphrey RM. The radiation response of human malignant melanoma cells grown in vitro. Cancer Res 1971;31:830 3. Dewey DL. The radiosensitivity of melanoma cells in culture. Br J Radiol 1971;44:816 7. Habermalz HJ, Fischer JJ. Radiation therapy of malignant melanoma. Cancer 1976;38:2258 62. Hornsey S. The relationship between total dose, number of fractions and fraction size in the response of malignant melanoma in patients. Br J Radiol 1978; 51:905 9. Overgaard J. Radiation treatment of malignant melanoma. Int J Radiat Oncol Biol Phys 1980;6:41 4. Lobo PA, Liebner EJ, Chao JJ, et al. Radiotherapy in the management of malignant melanoma. Int J Radiat Oncol Biol Phys 1981;7:21 6. Rounsaville MC, Cantril ST, Fontanesi J, et al. Radiotherapy in the management of cutaneous melanoma: effect of time, dose and fractionation. Front Radiat Ther Oncol 1998;22:62 78. Trott KR, von Lieden H, Kummermehr J, et al. The radiosentivity of malignant melanomas. Part 2: Clinical studies. Int J Radiat Oncol Biol Phys 1981;7: 15 20. Sause WT, Cooper JS, Rush S, et al. Fraction size in external beam radiation therapy in the treatment of melanoma. Int J Radiat Oncol Biol Phys 1991;20:429 32. Orton CG, Ellis F. A simplification in the use of the NSD concept in practical radiotherapy. Br J Radiol 1973;46:529 37. Arma-Szlachcic M, Ott F, Storck H. Zur strahlentherapie der melanotischen pra -cancerosen. Hautartz 1970; 21:505 8. Braun-Falco O, Lukacs S, Schoefinius HH. Zur behandlung der melanosis circumscripta praecancerosa Dubreuilh. Hautartz 1975;26:207 10. De Groot WP. Provisional results of treatment of the cance reuse circonscrite Dubreuilh by melanose pre Bucky-rays. Dermatologica 1968;136:429 31. Harwood AR, Cummings BJ. Radiotherapy for malignant melanoma: a re-appraisal. Cancer Treat Rev 1981;8:271 82. [18] Christie DRH, Tiver KW. Radiotherapy for melanotic freckles. Australas Radiol 1996;40:331 3. [19] Panizzon RG. Die roentgenweichstrahlentherapie als alternative bei alteren patienten. In: Burg G, Hartmann AA, Konz B, editors. Onkologische dermatologie: fortschritte der dermatologischen und onkologischen dermatologie, BAND 7. Heidelberg: Springer-Verlag; 1992. p. 263 7. [20] Tsang RW, Liu F-F, Wells W, et al. Lentigo maligna of the head and neck: results of treatment by radiotherapy. Arch Dermatol 1994;130:1008 12. [21] Skibber JM, Soong S, Austin L, et al. Cranial irradiation after surgical excision of brain metastases in melanoma patients. Ann Surg Oncol 1996;3:118 23. [22] Hagen NA, Cirrincione C, Thaler HT, et al. The role of radiation therapy following resection of single brain metastases from melanoma. Neurology 1990; 40:158 60. [23] OBrien CJ, Coates AS, Petersen-Schaefer K, et al. Experience with 998 cutaneous melanomas of the head and neck over 30 years. Am J Surg 1991;162:310 4. [24] OBrien CJ, Gianoutsos MP, Morgan MJ. Neck dissection for cutaneous malignant melanoma. World J Surg 1992;16:222 6. [25] Byers RM. The role of modified neck dissection in the treatment of cutaneous melanoma of the head and neck. Arch Surg 1986;121:1338 41. [26] Lee RJ, Gibbs JF, Proulx GM, et al. Nodal basin recurrence following lymph node dissection for melanoma: implications for adjuvant radiotherapy. Int J Radiat Oncol Biol Phys 2000;46:467 74. [27] Ang KK, Byers RM, Peters LJ, et al. Regional radiotherapy as adjuvant treatment for head and neck malignant melanoma: preliminary results. Arch Otolaryngol Head Neck Surg 1990;116:169 72. [28] Ang KK, Peters LJ, Weber RS, et al. Postoperative radiotherapy for cutaneous melanoma of the head and neck region. Int J Radiat Oncol Biol Phys 1994; 30:795 8. [29] OBrien CJ, Petersen-Schaefer K, Stevens GN, et al. Adjuvant radiotherapy following neck dissection and parotidectomy for metastatic malignant melanoma. Head Neck 1997;19:589 94. [30] Corry J, Smith JG, Bishop M, et al. Nodal radiation therapy for metastatic melanoma. Int J Radiat Oncol Biol Phys 1999;44:1065 9. [31] Cooper JS, Chang WS, Oratz R, et al. Elective radiation therapy for high-risk malignant melanomas. Cancer J 2001;7:498 502.
[2]
[3]
[4]
[5] [6] [7]
[8] [9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
Dermatol Clin 20 (2002) 717 725
Vaccines for melanoma

Jean-Claude Bystryn, MD
The Ronald O. Perelman Department of Dermatology and Melanoma Program and Melanoma Immunotherapy Clinic, Kaplan Comprehensive Cancer Center, New York University School of Medicine, 35 East 35th Street, Suite 208, New York, NY 10016, USA
Vaccines are a novel but still experimental treatment for melanoma. The rationale for believing that they may be effective, the relative advantages and disadvantages of the different approaches used to construct melanoma vaccines, and the results obtained to date with these vaccines are summarized in this article. Melanoma vaccine therapy is still experimental, and its effectiveness remains to be established. At present it is available only through clinical trials. These trials are being conducted in a number of institutions, however, so that melanoma vaccines are a realistic option for the adjuvant treatment of patients with melanoma at high risk of progression. Unfortunately, comparative data on the relative effectiveness of different melanoma vaccines are not available, so that it is difficult objectively to select the vaccine that is the most effective. One basis on which to make this decision is the strategy used to construct the vaccine, and its relative advantages and disadvantages in relation to alternate strategies.
The rationale for using vaccines to treat melanoma The rationale for using vaccines to treat melanoma is based on several fundamental observations, which have been reviewed [1]. Melanoma vaccines are intended to stimulate a patients immune system to react more strongly and effectively against his or her
own melanoma, and by so doing destroy the tumor or slow its progression. There are two major reasons for believing that melanoma vaccines may be able to do this. First, there are defense mechanisms in humans that can selectively kill melanoma cells. This is evidenced by the presence of tumor regression in 15% to 20% of primary melanomas, and by the rare but dramatic spontaneous and complete regression of advanced melanoma in some patients. These dramatic events clearly indicate that humans possess protective mechanisms that can selectively destroy melanoma cells. These defense mechanisms seem to be mediated by immune responses against melanoma because of the selectivity with which they act. They can destroy malignant pigmented cells (melanoma cells) without harming adjacent normal pigmented cells (melanocytes). Stimulating these immune defense mechanisms with vaccines should increase resistance to melanoma. Second, there is direct evidence that this can be accomplished, at least in mice. One hundred percent of melanoma vaccine immunized mice can survive challenge with a lethal dose of melanoma cells that invariably kills all nonimmunized animals [2]. The protection is specific for melanoma, because melanoma vaccine immunized mice are not protected against an unrelated tumor. This confirms that the protective mechanisms are immunologic in nature. The hope is that melanoma vaccines will likewise be able to increase resistance to melanoma in humans.
This article is supported in part by National Institutes of Health grant No. 1 R01 CA89270 and by grants from the Rose M. Badgeley Residual trust and the Catherine and Henry Gaisman Foundation. E-mail address: bystryn@nyu.edu
What is required for a melanoma vaccine to work? All cancer vaccines consist of two components: (1) tumor antigens, the component that stimulates
718
J.-C. Bystryn / Dermatol Clin 20 (2002) 717725
immune responses against the cancer; and (2) an adjuvant, the component that increases the ability of the tumor antigens to stimulate immune responses. An adjuvant is necessary, and used with almost all vaccines, because the immune responses induced by the tumor antigens alone are weak. For a cancer vaccine to be effective, it must satisfy a number of requirements. The two most important are that the tumor antigens selected to construct the vaccine must be able to stimulate tumor-protective immune responses, and some of these antigens must be present on the tumor to be treated. Unfortunately, there are problems satisfying both requirements. Although many antigens associated with melanoma have been identified and in some cases purified, those that can stimulate clinically effective, tumor-protective immune responses in humans are still unknown. In addition, the expression of tumor antigens by melanoma cells is heterogeneous. The pattern of antigens expressed by melanoma in different individuals, between different tumor nodules in the same individual, and even between different melanoma cells within the same tumor nodule, is variable. Furthermore, the particular antigens expressed by the residual melanoma cells in an individual patient at the time treatment is instituted are also usually unknown, and whatever this pattern may be it can change during the further evolution of the tumor (ie, melanoma cells in a given individual may gain or lose tumor antigens as the tumor progresses). One does not know what antigens should be used to construct a melanoma vaccine, whether these antigens are expressed by the tumor to be treated, or will still be there once the treatment has begun. A number of other problems complicate the selection of antigens used to construct a melanoma vaccine. Lastly, melanoma vaccines must fulfill a number of other requirements. They should be safe to use and devoid of potentially serious side effects. Their composition should be well characterized, and they must be able to be produced in a reproducible manner. In addition, to retain their potential to provide costeffective therapy, melanoma vaccines should be simple to prepare and administer. The number of antigens required to induce effective protective immunity It is unknown whether clinically effective tumor protective immunity can be induced by immunization to a single antigen, or whether responses to multiple antigens on the cancer cells are required for the cells to be killed. The number of tumor antigens that
should be included in a vaccine is another important consideration. Even if immune responses to a single antigen can kill melanoma cells, it seems logical that stimulating responses to multiple antigens should do so even more effectively. As described later, there is a price to be paid for increasing the number of antigens in a vaccine. HLA restriction of vaccine-induced immune responses Some immune responses to antigens are HLA restricted. HLA antigens, formerly known as transplantation antigens, are present in all individuals. There are a large number of HLA antigens, whose expression varies from individual to individual. Unfortunately, some immune responses can only be stimulated if the antigen is bound to an HLA molecule. These include several responses that play a critical role in mediating tumor protective immunity, such as CD8 + and CD4 + T-cell responses. Furthermore, each antigen or its fragments usually binds to only one of the many different types of HLA molecules. This means that an antigen induces these immune responses only in patients who have the required type of HLA antigen in their tissue. Because the HLA antigens expressed by different individuals are variable, this greatly reduces the number of patients who can benefit from vaccine constructed from a single antigen. Even the most common HLA antigen is expressed by less than half of the population, and many are expressed by only a few percentages of patients to be treated. As a result of HLA restrictions on the ability of tumor antigens to stimulate immune responses, and the variable expression of antigens on melanoma cells, the proportion of patients who can theoretically benefit from vaccines prepared from a single antigen that is universally expressed in all melanomas is no greater than 40%. Depending on the tissue expression of the antigen and the frequency of the HLA antigen that presents it among the population, the proportion of patients who may benefit from the treatment can be much lower. Heterogeneity in patients ability to develop immune responses to the same antigen There is also heterogeneity in the ability of different patients to develop immune responses to the same antigen, which is independent of HLA restriction [3]. Different patients expressing the same HLA phenotype and immunized identically to the same antigen vary in their ability to develop a cellular immune response to that antigen. Some develop a
719
strong response and others do not respond. This is not caused by lack of immune competence on the part of the patient, because patients who do not respond to one antigen respond well to another antigen presented by the same HLA molecule, whereas patients who responded to the first antigen may not respond to the second. There is similar heterogeneity in the ability of individuals to develop antibody responses [4].
The advantages of polyvalent vaccines One strategy partially to circumvent these problems is to construct polyvalent vaccines from a mixture of melanoma-associated antigens, which are frequently expressed among melanoma, whose pattern of expression is complimentary, and which are presented by HLA that are commonly found among the patient population to be treated. This approach increases the proportion of patients who can be treated by the vaccine. Furthermore, the greater the number of antigens in the vaccine, the greater the chances that it contains the unknown antigens that can stimulate tumor protective immunity, that at least one of these antigens is expressed by the patients tumor, and that the antigen is able to stimulate an immune response in the patient. Furthermore, stimulation of immune responses against multiple targets on the melanoma cells should increase the chances of killing these cells. Reflective of this approach is that the newer melanoma vaccines are now polyvalent, that is they contain multiple melanoma antigens rather than only a single antigen.
ments. None satisfies all requirements. All approaches can be grouped into one of three categories, which differ in the purity of the tumor antigens that are used to prepare the vaccine. The vaccine antigens can be (1) nonpurified, (2) partially purified, or (3) pure. Each of these procedures has certain advantages and disadvantages, as illustrated in Table 1 and discussed in greater detail elsewhere [1]. There is no ideal way of constructing cancer vaccines. Each available approach has certain advantages and disadvantages, and the treating physician must decide which of these combinations seems most advantageous. The essential difference between these different vaccine design strategies is in the number and the concentration of tumor antigens in the vaccine. The basic dilemma is that whereas nonpurified vaccines contain multiple tumor antigens and are more likely to contain relevant antigens, these antigens account for only a very small fraction of the material in the vaccine. Conversely, vaccines prepared from pure antigens contain a very high concentration of antigen, but it is uncertain whether the antigen selected to prepare the vaccine is the correct one, or whether immune responses stimulated against a single antigen are sufficiently potent to result in effective tumor kill. Partially purified vaccines strike a middle course between these two polar approaches. Nonpurified vaccines The traditional method to construct melanoma vaccines is to prepare them from whole melanoma cells or nonpurified extract of these cells. The melanoma cells are often pooled from different donors to increase the number of melanoma antigens in the vaccine. Examples of such vaccines, which are in advanced clinical trials, include CancerVax, a polyvalent vaccine prepared from several lines of intact, irradiated, melanoma cells combined with bacille rin (BCG) as an adjuvant [5]; Melacine, Calmette-Gue
Strategies used to construct melanoma vaccines and their relative advantages and disadvantages Multiple strategies are used to prepare melanoma vaccines that address the previously listed require-
Table 1 Relative merits of different melanoma vaccine design strategies Desirable vaccine attribute Effectiveness Presence relevant antigens Presence multiple antigens Immunogenic potency of antigens Elegance of preparation Purity Reproducible manufacture Quality control Safety Nonpurified vaccines +++ +++ +++ + + + ++ Partially pure vaccines ++ ++ ++ ++ ++ ++ ++ Pure vaccines + + + +++ +++ +++ ++
720
which is a mechanical lysate of two melanoma cell lines admixed with Detox as an adjuvant [6]; whole, autologous, melanoma cells conjugated to a hapten [7]; and lysates of cells infected with vaccinia or other nonpathogenic viruses [8]. More recently, the use of intact tumor cells as a vaccine has permitted the cells to be engineered genetically to express certain cytokines, such as interleukin (IL)-2 or other molecules or antigens that can increase their ability to stimulate immune responses. The major advantage of nonpurified vaccines is that they contain a wide variety of tumor antigens. They are more likely to contain the as yet unidentified antigens that are essential for vaccine activity, and they obviate the need to identify and purify each individual such antigens. Because the vaccines contain a broad range of tumor antigens, particularly if prepared from a pool of melanoma cells, they have an increased chance of containing relevant antigens expressed by the individual tumor to be treated. For the same reason the vaccines are more likely to contain antigens that are recognized in the context of the patients HLA phenotype or that can trigger immune responses in a particular patient. Vaccines prepared from whole tumor cells may be inherently more immunogenic, because antigens tend to lose potency as they are purified. Lastly, the multiple antigens in nonpurified vaccines have the potential to stimulate antitumor immune responses to multiple targets on tumor cells, which is more likely to result in cell kill and clinical effectiveness. The major drawback to this approach is that the active antigens in the vaccine are present in a very low concentration and constitute only a small fraction of the material in the vaccine. The presence of large quantities of irrelevant material increases the technical difficulty of producing a reproducible and wellcharacterized vaccine formulation. In addition, the nonantigen material in the vaccine may potentially decrease its effectiveness because of the presence of suppressive factors or by competitive inhibition of antitumor immune responses, and may potentially increase toxicity. Vaccines prepared from pure antigens Scientifically more elegant is to construct the vaccine from a single or limited number, of highly purified antigens. This has become feasible as a result of the identification and purification of melanoma-associated antigens that are recognized by antibodies or T cells. The purified vaccine in the most advanced stage of clinical testing is a GM-2 ganglioside vaccine, which is conjugated to Keyhole
Limpet Hemocyanin (KLH) and administered with the adjuvant QS-21 [9]. Other purified antigen vaccines are still in early phases of clinical trials. Most of these are prepared from peptides derived from the melanoma-associated antigens MAGE-1, MAGE-3, Melan A/Mart-1, tyrosinase, or gp100, which are recognized by CD8 + T cells in vitro. The use of anti-idiotype monoclonal antibodies is an alternate approach to stimulate responses to tumor using a purified protein [10]. The major advantage of this approach is that pure antigen vaccines are characterized and prepared more easily in a reproducible manner, and that the concentration of the antigens in the vaccine is high. From a research perspective, the immune responses that they induce can be measured more easily. Unfortunately, because the melanoma antigens actually responsible for inducting clinically beneficial tumor protective responses have not yet been identified, the precise antigens that should be selected to formulate such vaccines are still unknown. In addition, many of the antigens used are HLA-restricted. That, together with the less than universal expression of these antigens in different melanomas, means that such vaccines cannot be used in many patients with melanoma. Lastly, it is still unclear whether immune responses induced to a single antigen or peptide are sufficiently potent to result in effective tumor cell kill. These problems can be minimized to a certain degree by preparing vaccines from a cocktail of several purified antigens that have been purified. Furthermore, at a practical level it is very difficult to actually create such a cocktail because of intellectual property issuesthe rights to each antigen belong to different persons. Vaccines prepared from partially purified antigens The third design strategy attempts to balance the advantages and disadvantages of the two prior approaches. It does so by preparing vaccines from tumor cell extracts that are enriched in the cellular elements most likely to contain tumor antigens relevant for vaccine therapy and depleted of material likely to be irrelevant. The only example of this approach currently in clinical trial is a polyvalent, shed melanoma antigen vaccine prepared from cell-surface antigens released into culture medium by melanoma cells [11]. This method exploits a natural phenomenon, the rapid release or shedding of surface material by cells in culture. Because molecules on the cell surface are released much more rapidly than the bulk of unrelated cellular material, which is in the cytoplasm or the nucleus, the shed material is partially purified and is
721
enriched in surface antigens. To increase the representation of melanoma antigens in the vaccine, the cell lines used for vaccine production are selected based on each expressing different patterns of cell-surface melanoma-associated antigens. The shed antigen vaccine that has been prepared in this manner contains multiple melanoma antigens, including MAGE-1, MAGE-3, MelanA/Mart-1, tyrosinase, gp100, S100, TRP-1, and numerous immunogenic antigens ranging in molecular weight from 30 to 150 kd, which are expressed by melanoma in vivo [3,12]. The advantage of this approach is that it retains the most critical element required for vaccine effectiveness (ie, a cocktail of antigens), while being much purer than vaccines prepared from whole tumors or their extracts. Polyvalent vaccines that contain a large number of tumor antigens are more likely to stimulate tumor protective immunity, to circumvent the antigenic heterogeneity and HLA polymorphism of patients with melanoma, and stimulate responses to multiple targets on the melanoma cells.
of a cytokine that can up-regulate immune responses, such as IL-2 or granulocyte-macrophage colony stimulating factor. This has the advantage of greatly prolonging the half-life of the cytokine and of delivering the cytokine together with the antigens to draining lymph nodes where stimulation of immune cells occurs. This approach has been shown to increase markedly vaccine-induced cellular immune responses in humans [13]. Complicating the selection of the best adjuvants is that different adjuvants potentiate humoral and cellular immune responses differently. Furthermore, their effect varies depending on the antigen with which they are coupled, so that the results obtained with one vaccine-adjuvant combination do not necessarily translate to another vaccine coupled with the same adjuvant. Clearly, adjuvants are required to improve the effectiveness of vaccines for melanoma, a large choice is available, but it is unclear which is the best. These choices, however, further complicate the selection of the vaccine that one believes may be the most effective.
Adjuvants Results of clinical trials of melanoma vaccines The inability of most melanoma antigens to consistently induce strong antimelanoma immune responses is another problem. Thus, the development of powerful adjuvants is another critical element in the construction of effective cancer vaccines. A large number of different adjuvants is currently being investigated. These range from modifications of the physical or biochemical properties of the antigens; classic adjuvants, such as alum, BCG, or QS-21; and includes slow-release vehicles, such as liposomes; immunomodulators, such as IL-1, IL-2, IL-12, or granulocyte-macrophage colony stimulating factor; or antigen-presenting cells, such as dendritic cells. Other approaches include coupling the antigen to strongly immunogenic molecules, such as KLH; expressing the antigen on the surface of viral particles, such as vaccinia or on that of inert beads; or transfecting tumor cells to express molecules that it is hoped will increase their immunogenicity. Little is known about the relative effectiveness of these approaches. One of the most effective approaches is to present the antigen on dendritic cells. Unfortunately, this approach is limited by the need to obtain the dendritic cells from the patient to be treated, which means that a custom vaccine needs to be prepared for each individual patient. This greatly increases the cost and the time required to prepare such vaccines. Another promising approach is to encapsulate the vaccine into liposomes together with a small amount Taking into account the variety of different approaches that can be used to select the antigens used to prepare melanoma vaccines and the equally large number of adjuvants used to enhance the activity of these vaccines, it is not surprising that a large number of melanoma vaccines are currently in clinical trial. A partial listing of these vaccines is provided in Table 2. Little solid information is available on the clinical effectiveness of these vaccines. Most of the information available relates to nonpurified and partially purified vaccines because several of these are already in advanced clinical trials. Three criteria are used to evaluate the clinical activity of tumor vaccines: (1) safety; (2) immunologic activity (ability to stimulate immune responses against melanoma); and (3) clinical effectiveness (ability to slow the progression of melanoma). In general, melanoma vaccines have proved very safe to use. There has been little toxicity in the several thousand patients who have been treated to date. The toxicity that has been reported is usually caused by the adjuvant rather than to the vaccine. Vaccines are clearly less toxic than the current Food and Drug Administration (FDA) approved therapy for resected melanoma at high risk of recurrence. This is interferon alfa-2b, which causes toxicity in over two thirds of patients [14].
722 Table 2 Melanoma vaccines in clinical trials Vaccine Nonpurified Whole cells Cell lysate (Melacine) Whole cells Whole cells Whole cells Semipure Shed antigens Pure GM-2 ganglioside MAGE-1 peptide MAGE-3 peptide MART-1 peptide gp100 Autologous peptides Anti-R24 Anti-HMW-MAA
Adjuvant BCG Detox IL-2/TNF gene IFN gene Hapten Various KLH + QS21 Poxvirus IFA IFA/vaccinia/adeno Vaccinia/adenovirus Heat shock protein BCG KLH + BCG
Investigator Morton Mitchell Rosenberg Seigler Berd Bystryn Livingston Therion Rosenberg Rosenberg Rosenberg Srivastava Ferrone IDEC
Phase trial III III I I III III III I I I I I I I
Abbreviations: BCG, bacille Calmette-Gue rin; IFA, incomplete points freunds adjuvant; IFN, interferon; FL, interleukin; KLH, ; TNF, tumor necrosis factor.
Some melanoma vaccines are immunogenic in humans, and increase antibody or cellular immune responses to melanoma [5,7,15 20] The frequency with which they do so and the type of response they stimulate are related to the nature of the vaccine and the manner in which it is administered. For example, it is difficult to demonstrate immune responses to peptide-based vaccines, ganglioside vaccines stimulate antibody but not cellular responses, and cellular and protein-based vaccines can stimulate both types of responses. Potent adjuvants, such as dendritic cells, IL-2 liposomes, and QS21, can markedly increase the frequency with which the same vaccine stimulates antibody or cellular responses. The clinical effectiveness of melanoma vaccines remains unknown, because no large-scale phase III randomized study demonstrating a survival advantage for vaccine-treated patients has yet been completed. A number of vaccines seem effective, however, based on comparisons with historic controls, or in small concurrently controlled randomized trials. Some vaccines have been reported to induce regression of established tumors, but because similar remission rates occur with most other modalities used to treat melanoma, the significance of this observation is unclear. The vaccine formulations that have advanced to randomized trials are listed next. All are safe, because they have been administered to numerous patients
with minimal toxicity. All can stimulate antibody responses, and all but the GM-2 vaccine have been reported to stimulate cellular immune responses. 1. CancerVax, an irradiated, whole melanoma cell, polyvalent, vaccine developed by Morton at the John Wayne Cancer Center (Santa Monica, CA). The recurrence-free and overall survival of stage III and IV patients treated with this vaccine is longer than that of large numbers of historic patients treated by the same investigator [21]. This vaccine is currently in a national phase III trial and is available through a number of institutions. 2. A polyvalent, partially purified, shed antigen vaccine developed by Bystryn and associates at the NYU Medical Center (New York, NY). This vaccine contains multiple melanomaassociated antigens including MAGE-1, MAGE-3, MelanA/MART-1, gp100, S100, tyrosinase, TRP-1, TRP-2, and a number of other antigens with molecular weight ranging from 30 to 150 kd. This vaccine stimulates antibody responses to multiple melanomaassociated antigens, which are expressed in vivo by melanoma nodules [12]; cellular responses to a patients own melanoma in vivo [22]; and CD8 + T-cell responses to multiple peptides derived from MAGE-1, MAGE-3, MART-1, tyrosinase, and TRP-1,
723
which are presented by the class I HLA phenotypes most common among melanoma patients [3]. The recurrence-free and overall survival of patients with stage III or early stage IV disease treated with this vaccine is approximately 50% longer than that of similar historic controls [23]. The most compelling evidence that it is effective is that in doubleblind placebo-controlled trial in which the recurrence-free survival of melanoma vaccine treated patients was over twice as long as that of patients treated with a placebo vaccine, and their overall survival was 40% longer [11]. The difference in recurrence-free survival was statistically significant after Cox multivariate analysis. The results must be interpreted cautiously, however, because they are based on a small number of patients. 3. Melacine, a vaccine prepared by Dr. Malcolm Mitchell, from the lysate of two melanoma cell lines admixed with Detox as an adjuvant. The clinical outcome of patients with stage IV melanoma treated with the vaccine or with standard chemotherapy was similar, but toxicity with vaccine treatment was less, resulting in improved quality of life [6]. Furthermore, in a randomized trial in resected melanoma, the recurrence-free survival of vaccine-treated patients with certain HLA phenotypes was prolonged compared to the control group. 4. GM-2 vaccine prepared from a purified ganglioside by Dr. Phillip Livingston. An initial randomized study of this vaccine has been conducted in American Joint Committee on Cancer stage III melanoma [19]. The patients were pretreated with cyclophosphamide and immunized to either vaccine plus BCG or to BCG alone. There was no significant difference in disease-free or overall survival between the two groups, but an encouraging trend in favor of the vaccine-treated patients was present. Outcome was significantly better in those patients who developed an antibody response to the vaccine [19]. Conjugating the GM-2 to KLH and giving the construct with QS-21 as an adjuvant significantly augments antibody responses, suggesting that this may provide a more effective vaccine. In a large randomized trial of this modified vaccine conducted by the Eastern Clinical Oncology Group, however, the recurrence-free survival of vaccine-treated patients was poorer than that of patients treated conventionally with high doses of interferon alfa [9]. This trial may have
been interrupted early, however, before the full benefits of vaccine treatment became evident. Numerous other melanoma vaccines are in clinical trials in the United States and Europe. A list of melanoma vaccines currently in clinical trial, and a description of the entry requirements for these trials, can be obtained by contacting the National Cancer Institute at 301-496-4000.
Who is a candidate for melanoma vaccine treatment? Because vaccines are still experimental, there is no standard criteria for determining whether or not a patient is a candidate for melanoma vaccine treatment. The eligibility and exclusion criteria for vaccine treatment is set by the investigators conducting a particular trial, so that it is necessary to contact each trial to determine what these are. Generally, the types of patients most often treated with vaccines are those with melanoma that has been resected but has a high chance of recurrence (ie, patients with American Joint Committee on Cancer stage IIB [primary lesions 4 mm or thicker] or with stage III [metastatic to regional nodes] disease). A few trials are being conducted in patients with thinner primary melanoma (primary lesions 1.5 to 4 mm in thickness) and with disseminated stage IV disease, either resected or with limited tumor load. It is generally believed that vaccines will be less effective in patients with advanced disease. Conversely, it is believed that vaccines should not be used in patients with melanoma at low risk of recurrence, because the treatment is still experimental and the full spectrum of potential side effects remains to be determined.
What to tell patients Before initiating therapy with vaccines it is important to discuss with patients the experimental nature of the treatment, its potential risks, and the lack of solid data about effectiveness. The specific toxicities that need to be discussed and the data on potential effectiveness vary depending on the particular vaccine that is offered to the patient. The FDA now requires that patients be informed about the highly unlikely possibility of contracting bovine spongiform encephalopathy or Creutzfeldt-Jakob disease if any of the reagents used to prepare the vaccine (such as the fetal calf serum or trypsin often used in the preparation of whole-cell vaccines or the amino
724
acid used to prepare peptide vaccines) are of ruminant or human origin. Alternate treatment options also must be discussed. This is particularly true for patients in the adjuvant setting (ie, those with thick primary melanoma or disease metastatic to regional nodes) because high doses of interferon alfa-2b is an FDA-approved treatment for such patients. The patients need to be informed that interferon is an FDA-approved treatment for their disease, that there are data from randomized trials indicative of effectiveness, but that improvement in recurrence-free and overall survival is limited and not fully accepted by some oncologists, and that the treatment is associated with frequent and sometimes severe side effects. Vaccines are an option for patients who refuse interferon therapy or who have completed interferon therapy. The chance of vaccine being effective in a patient who has failed chemotherapy and has widespread disease seems unlikely at this stage.
the availability of fresh autologous tumor tissue. Lastly, selection among trials must also be based on what trials are physically accessible to the patient, although this range can be very wide for patients with the financial means to travel to a distant facility.
Summary Melanoma vaccines are now an accepted but still experimental treatment for patients who have been rendered clinically free of disease by surgical resection but are at high risk of recurrence and in selected patients with advanced but still limited disease. In general, there seems to be a correlation between the ability of melanoma vaccines to stimulate antimelanoma cellular or antibody immune responses and improved clinical outcome. Accordingly, a number of strategies are now being pursued to improve the clinical effectiveness of this first generation of vaccines by improving their ability to stimulate antimelanoma immunity. To establish the true effectiveness of vaccines in the treatment of malignant melanoma, several large, prospectively randomized phase III studies are currently being conducted.
What vaccine should be recommended? The basic issues in selecting a particular vaccine treatment include (1) the relative effectiveness of the different vaccine formulations, (2) the entry and exclusion criteria for a particular trial, and (3) the trials that are actually available within the patients geographic range. There is little hard data on the actual effectiveness of individual vaccines, and no data at all on the relative effectiveness of the vaccines to each other. Data supportive of effectiveness that is based on randomized trials are available for only three vaccines: (1) Melacine, (2) the GMK vaccine, and (3) the shed antigen vaccine developed by Bystryn. Although randomized trial data are not yet available, there is an unusually large historic database to support the effectiveness of Cancervax. Data supporting the clinical effectiveness of other vaccines are either lacking or are based on small historically controlled studies that must be interpreted with a great deal of caution. The effectiveness of vaccines at present is thus evaluated predominantly based on what theoretically seems to be the most effective approach (eg, does the vaccine contain multiple antigens and does it actually induce antitumor immune responses) and least likely to cause toxicity. Vaccines constructed from viral particles or admixed with strong adjuvants, such as BCG, are more likely to cause toxicity. Selection among trials must also be based on the restrictions of the trials: the stage of the disease; the time that has elapsed since surgery; whether the trial is intended only for patients with a particular HLA phenotype, such as HLA-A2; and whether it requires
References
[1] Bystryn J-C, Shapiro RL, Oratz R. Cancer vaccines: clinical applications: Partially purified tumor antigen vaccines. In: DeVita V, Hellman S, Rosenberg SA, editors. Biologic therapy of cancer. 2nd edition. Philadelphia: JB Lippincott; 1995. p. 668 79. [2] Bystryn J-C. Antibody response and tumor growth in syngeneic mice immunized to partially purified B16 melanoma associated antigens. J Immunol 1978;120: 96 101. [3] Reynolds SR, Celis E, Sette A, Oratz R, Shapiro RL, Johnston D, et al. HLA-independent heterogeneity of CD8 + T cell responses to MAGE-3, Melan A/MART-1, gp100, tyrosinase, MC1R and TRP-2 in vaccine-treated melanoma patients. J Immunol 1998;161:6970 6. [4] Miller K, Abeles G, Oratz R, Zeleniuch-Jacquotte A, Cui J, Roses DF, et al. Improved survival of melanoma patients with an antibody response to immunization to a polyvalent melanoma vaccine. Cancer 1995;75: 495 502. [5] Morton DL, Barth A. Vaccine therapy for malignant melanoma. CA Cancer J Clin 1996;46:225 44. [6] Mitchell MS, Rechtman DJ, Von Eschen KB. A randomized phase III trial of Melacine versus combination chemotherapy in patients with disseminated melanoma. Can J Infect Dis 1995;6(suppl C):347C. [7] Berd D, Maguire Jr. HC, Schuchter LM, Hamilton R, Hauck WW, Sato T, et al. Autologous haptenmodified
J.-C. Bystryn / Dermatol Clin 20 (2002) 717725 melanoma vaccine as postsurgical adjuvant treatment after resection of nodal metastases. J Clin Oncol 1997; 15:2359 70. Hersey P. Evaluation of vaccinia viral lysates as therapeutic vaccines in the treatment of melanoma. Ann N Y Acad Sci 1993;690:167 77. Kirkwood JM, Ibrahim JG, Sosman JA, Sondak VK, Agarwala SS, Ernstoff MD, et al. High dose interferon alfa-2b significantly prolongs relapse-free and overall survival compared with the GM2-KLH/QS-21 vaccine in patients with resected stage IIB-III melanoma: results of intergroup trial E1694/S9512/C509801. J Clin Oncol 2001;19:2370 80. Ferrone S. Human tumor associated antigen mimicry by anti-idiotypic antibodies. Ann N Y Acad Sci 1993; 690:214 21. Bystryn J-C, Zeleniuch-Jacquotte A, Oratz R, Shapiro RL, Harris MN, Roses DF. Double-blind trial of a polyvalent, shed-antigen, melanoma vaccine. Clin Cancer Res 2001;7:1882 7. Applebaum J, Reynolds S, Knispel J, Oratz R, Shapiro RL, Bystryn J-C. Identification of melanoma antigens that are immunogenic in humans and expressed in vivo. J Natl Cancer Inst 1998;90:146 9. Oratz R, Shapiro RL, Johnston D, Harris MN, Roses DF, Zeleniuch-Jacquotte A, et al. IL-2 liposomes markedly enhance the activity of a polyvalent melanoma vaccine. Proc Am Assoc Clin Oncol 1996;15:436. Kirkwood JM, Strawderman MH, Ernstoff MS, Smith TJ, Borden EC, Blum RH. Interferon alfa-2b adjuvant therapy of high-risk resected cutaneous melanoma: the Eastern Cooperative Oncology Group Trial EST 1684. J Clin Oncol 1996;14:7 17. Rosenberg SA, Yang JC, Schwartzentruber DJ, Hwu P, Marincola FM, Topalian SL, et al. Immunologic and therapeutic evaluation of a synthetic peptide vaccine
725
[16]
[8]
[9]
[17]
[18]
[10]
[11]
[19]
[12]
[20]
[13]
[21]
[22]
[14]
[23]
[15]
for the treatment of patients with metastatic melanoma. Nat Med 1998;4:321 7. Reynolds SR, Oratz R, Shapiro RL, Fotino M, Hao P, Vukmanovic S, et al. Stimulation of CD8 + T cell responses to MAGE-3 and MELAN A/MART-1 by immunization to a polyvalent melanoma vaccine. Int J Cancer 1997;72:972 6. Miller K, Abeles G, Oratz R, Zeleniuch-Jacquotte A, Cui J, Roses D, et al. Improved survival of melanoma patients with an antibody response to immunization to a polyvalent melanoma vaccine. Cancer 1995;75: 495 502. Takahashi T, Johnson TD, Nishinaka Y, Morton DL, et al. IgM anti-ganglioside antibodies induced by melanoma cell vaccine correlate with survival of melanoma patients. J Invest Dermatol 1999;112:205 9. Livingston PO, Wong GYC, Adluri S, Tao Y, Padavan M, Parente R, et al. Improved survival in stage III melanoma patients with GM2 antibodies: a randomized trial of adjuvant vaccination with GM2 ganglioside. J Clin Oncol 1994;12:1036 44. Mitchell MS. Perspective on allogeneic melanoma lysates in active specific immunotherapy. Semin Oncol 1998;25:623 35. Chan AD, Morton DL. Active immunotherapy with allogeneic tumor cell vaccines: present status. Semin Oncol 1998;25:611 22. Oratz R, Cockerall C, Speyer JL, Harris MN, Roses DF, Bystryn J-C. Induction of tumor-infiltrating lymphocytes in malignant melanoma metastases by immunization to melanoma antigen vaccine. J Biol Res Modif 1989;8:355 8. Bystryn J-C, Oratz R, Roses D, Harris M, Henn M, Lew R. Relation between immune response to melanoma vaccine immunization and clinical outcome in stage II malignant melanoma. Cancer 1992;69:1157 64.
Dermatol Clin 20 (2002) 727 733
Unusual melanoma types

Jason K. Rivers, BSc, MD*
Division of Dermatology, University of British Columbia, 835 West 10th Avenue, Vancouver, BC, Canada V5Z 4E8 The British Columbia Cancer Agency, 600 West 10th Avenue, Vancouver, BC, Canada, 5Z4 E6
Melanoma (MM) is the most rapidly increasing cancer in white populations, yet until recently there has been no significant improvement in diagnostic accuracy [1]. To address this problem, tools have been developed to aid in the diagnosis of MM. Some of these include epiluminescence microscopy; computer digital image analysis; and more recently, multispectral digital dermoscopy. Epiluminescence microscopy has been shown to increase the accuracy of diagnosis, in experienced hands, and can help to distinguish between certain benign and malignant growths [2 4]. Computerized digital imaging provides an automated analysis of digital epiluminescence microscopy images. This may be useful for the follow-up of patients with multiple atypical nevi [5,6]. Multispectral digital dermoscopy is a fully automated device that can differentiate MM from dysplastic nevi and other melanocytic nevi. With this technique, 10 dermoscopic images are acquired in the visible and near-infrared spectrum. In one recent study [7], the technique achieved a sensitivity and specificity of 100% and 84%, respectively. Even with these tools, however, there remain a number of unusual types of MM that are difficult to diagnose and manage. An awareness of these less common forms of MM and their natural histories aids the physician to diagnose them early and treat them appropriately, thereby maximizing the chance of cure while minimizing patient morbidity.
Amelanotic MM Some MM produce little if any discernible pigment (Fig. 1) and this often leads to potential delays in diagnosis. Amelanotic MM represents 1.8% to 8.1% of all MM [8] and these tumors can be confused with a number of benign and malignant lesions (Table 1). Although any subtype of cutaneous MM may be amelanotic, it is more common for this to occur in subungual and desmoplastic lesions [8]. MM is rare in those who have oculocutaneous albinism, but in at least a few of these individuals, amelanotic MM has been described [9,10]. Amelanotic MM can recur after cryotherapy for pigmented lentigo maligna [11], and this stresses the importance of complete surgical excision of any MM whenever possible. Bono et al [12] recently described the clinical and dermatoscopic features of early amelanotic MM. They studied 15 biopsy-confirmed lesions with a thickness of less than 1 mm. Although, the clinical features were only diagnostic in 40% of cases, these authors noted that typical early lesions presented as an asymmetric macule that was uniformly pinkish or reddish in color. Many lesions, however, did have faint pigment at the margin. By dermatoscopy, they identified as a constant feature the presence of small red dots, evenly distributed or grouped on a whitish or pink-red background. These features were not of themselves, however, diagnostic of amelanotic MM. Near-infrared confocal scanning laser microscopy is a new imaging modality that holds promise for the in vivo diagnosis of amelanotic MM [13]. In one case, an abnormal intraepithelial melanocytic proliferation was clearly different from normal skin, and this was later confirmed using routine histology [13]. In animal models photodynamic therapy using topical a-amino-
* 835 West 10th Avenue, Vancouver, BC, Canada, V52 4E8. E-mail address: jrivers@interchange.ubc.ca (J.K. Rivers).
728
J.K. Rivers / Dermatol Clin 20 (2002) 727733
Fig. 1. Amelanotic melanoma on the back of a 35-year-old woman; Breslow thickness 0.58 mm.
levulinic acid has been shown to delineate amelanotic MM from normal surrounding skin [43]. It remains to be seen if this modality is of diagnostic and therapeutic benefit in humans. The prognosis of primary cutaneous amelanotic MM is similar to its pigmented counterpart and is primarily dependent on tumor thickness [8]. These tumors, however, are often advanced because of a delay in diagnosis. In one series, the average delay in diagnosis was 13 months and many lesions had been treated with various modalities before a definite diagnosis was ascertained [15].
Desmoplastic melanoma The first report of desmoplastic MM, a tumor with dermal spindle cells and abundant collagen, appeared in the literature in 1971 [16]. Since then several large series have been published that highlight the clinical and histologic features of this unusual form of MM [17 20]. These lesions generally occur in older individuals and are more frequent in men. Most often, they are found on the head and neck region in sun-damaged skin, although other sites of involvement have been
described [21]. Often these MM are amelanotic (40% to 50%); seen in association with lentigo maligna (Fig. 2); and show a propensity to local recurrence with a poor prognosis. Recently, Quinn et al [18] described their experience with 280 patients from a tertiary referral center in Australia. The male to female ratio was 1.75:1, the median age of the patients was 61 years, almost half of the cases were amelanotic, and the median tumor thickness was 2.25 mm. The local recurrence rate was 11% and this was more likely to be increased when the surgical margin was less than 1 cm and neurotropism was evident. Almost 80% of recurrences occurred within 2 years of presentation, which supports the finding that this type of MM recurs sooner than after other MMs [22]. The 5-year survival rate was 75%, which was similar to patients with other types of cutaneous MM. Skelton et al [19] reported on the histology of 128 cases of desmoplastic MM and noted that the tumors were composed of elongated and fusiform spindle cells that infiltrated the subcutaneous tissue (Fig. 3). Abundant collagen was present within the tumors and scattered lymphoid aggregates and plasma cells were evident around tumor cells. The tumors demonstrated variation of nuclear atypia but were characterized by low cellularity and modest mitotic activity. In some cases, nests of more epithelioid-like cells characteristic of a typical MM were admixed within the desmoplastic MM. Because the histologic diagnosis of desmoplastic MM is often not straightforward and may be missed or confused with benign or less aggressive tumors [20], immunohistochemistry has been used to aid in the diagnosis. Desmoplastic MM generally expresses S-100 protein but these tumors are almost always negative for the melanocyte specific markers HMB-45 and melan-A [18,19,23]. Unfortunately, S-100 ex-
Table 1 Clinical misdiagnosis: amelanotic melanoma Benign lesions Intradermal nevus Verruca vulgaris Pyogenic granuloma Dermatitis Malignant lesions Basal cell carcinoma Keratoacanthoma Bowens disease Atypical fibroxanthoma Fig. 2. Desmoplastic melanoma arising in association with lentigo maligna melanoma.
Adapted from Koch SE, Lange JR. Amelanotic melanoma: the great masquerader. J Am Acad Dermatol 2000;42: 731 734; with permission.
729
Schwann cells, and smooth muscle cells. This lack of sensitivity and specificity may limit the use of microphthalmia transcription factor, especially in reexcisions, when immunohistochemistry is often needed to distinguish an inflamed scar from tumor. Reasons for frequent tumor recurrence include clinical misdiagnosis and the confusing histology that may eventuate in the setting of residual disease. Positron emission tomography may be of benefit in monitoring patients with desmoplastic MM [26], whereas wide local excision of primary or residual disease is the treatment of choice for most lesions [17,18].
Malignant blue nevus Malignant blue nevus is a rare form of MM that frequently metastasizes [27]. This tumor tends to occur in whites and is usually seen on the scalp. Other sites have been reported, however, such as the face, trunk, and foot [28]. The average age at diagnosis is 45 years and most tumors have arisen in men. The diagnosis should be suspected in a lesion with a diameter greater than 2 cm in a solitary nodule (Fig. 4), or in the presence of multiple lesions in a multinodular or plaque form. A history of sudden change or rapid growth points to the diagnosis. This malignancy needs to be distinguished from a benign blue nevus with satellites (Fig. 5) [29], and a cellular blue nevus with nevus-cell aggregates in the regional lymph nodes [30]. By histology, the epidermis is uninvolved and tumor cells are located within the dermis and subcutaneous fat (Fig. 6). A variable degree of melanin pigment may be evident and cellular pleomorphism is often present. Mitoses are usual (in contrast to
Fig. 3. Desmoplastic melanoma. Lentigo maligna epidermal component with spindle-shaped cells surrounded by mature collagen in the dermis (hematoxylin and eosin, original magnification 10).
pression has been detected in a high percentage of cutaneous scars and this may complicate the diagnosis of desmoplasticMM[23].Thisbecomesespeciallyimportant in the setting of residual disease in incompletely excisedormarginallyexcised tumors. With this in mind, newer more sensitive and specific melanocyte markers have been explored. Koch et al [24] found that microphthalmia transcription factor and melanoma cell adhesion molecule expression could help to distinguish desmoplastic MM from other benign and malignant spindle cell proliferations. They concluded that microphthalmia transcription factor (a nuclear regulatory protein essential for the embryologic migration and survival of melanocytes, and for the expression of the tyrosinase gene promoter) should be part of the initial immunohistochemical panel for the work-up of such cases, whereas melanoma cell adhesion molecule is best suited for the confirmation of S-100 positive, microphthalmia transcription factor negative suspected desmoplastic MM. Others [25], however, have documented microphthalmia transcription factor expression in a number of normal cells including melanocytes, macrophages, lymphocytes, fibroblasts,
Fig. 4. Malignant blue nevus on the nasal ala of a 70-yearold Asian man.
730
Fig. 5. Benign blue nevus with satellites arising on the sole. (Courtesy of David McLean, MD, Vancouver, Canada.)
benign blue nevi) and in some cases tumor necrosis is also apparent. Biopsy specimens taken simultaneously from the same lesion may show different histology, however, ranging from benign to malignant [27,28]. The manner in which the biopsy is taken is critical; punch or shave biopsies may be misleading and should be avoided. Only complete excision of the lesion should be done to confirm the diagnosis.
Fig. 7. Scleral pigmentation seen in nevus of Ota (From: Rivers JK, Bhayana S, Martinka M. Dural Melanoma associated with ocular melanosis and multiple blue nevi. J Cutan Med Surg 2001;5:381 5; with permission).
Nevus of Ota associated melanoma The first definitive description of the nevus of Ota was made in 1939 and it was originally termed nevus fuscoceruleus ophthalmomaxillaris and categorized as a congenital benign dermal melanocytosis [31]. It usually occurs as a flat or slightly raised blue-black or slate-gray unilateral discoloration most often in the distribution of the first and second branches of the trigeminal nerve. Other areas, such as the pharynx, palate, tympanic membrane, and nasal mucosa, may
also have melanosis [14,32]. Up to 77% of cases may have ocular involvement with pigment being distributed in the sclera, conjunctiva, iris, cornea, choroid, or optic disk (Fig. 7). Glaucoma may complicate this condition and has been reported to occur in 10.3% of patients in one series [33]. As of 1998, only 10 cases of cutaneous MM developing in a nevus of Ota had been reported [14]. The most common mode of presentation was of a new subcutaneous nodule that was suggestive of a more benign process (eg, angioma or dermoid cyst). Primary intracranial MM has been reported in several cases, and in almost half of these cases, dural attachment was documented [34 36]. Primary dural MM may follow a less aggressive clinical course with a median postoperative survival of 4 years [35], and surgery without radiotherapy may be the preferred approach to management of these rare intracranial tumors.
Fig. 6. Malignant blue nevus. Note sparing of the epidermis and cellular pleomorphism (hematoxylin and eosin, original magnification 40).
Fig. 8. Subungual melanoma arising from the thumb. Amelanotic lesions may be confused with other benign and malignant processes.
731
Fig. 9. In situ subungual melanoma of the thumb. Note the absence of Hutchinsons sign.
Suspicious changes within a nevus of Ota should undergo biopsy, especially the sudden appearance of a subcutaneous nodule. Any neurologic or ophthalmologic signs or symptoms should be investigated because they may herald a malignancy.
Subungual melanoma Subungual MM is uncommon representing 0.7% to 3.5% of all MM [37,38]. Between 15% and 65% of these tumors may be amelanotic (Fig. 8), accounting for the delay in diagnosis compared with cutaneous MM [38]. This tumor usually affects individuals in the fifth to seventh decades of life (range 12 months to 90 years). Although the overall rate for MM is low in people with dark skin, subungual MM is overrepresented in these groups: 15% to 20% of all MM in African-Americans, 10% to 31% in Asians, and 33% in Native Americans [37]. Subungual MM often presents as a brown to black longitudinal nail band (melanonychia stria). The band is often wider than 5 mm and the border may be blurred or irregular (Fig. 9). Hutchinsons sign (periungual spread of pigment into the proximal or lateral nail folds) is not always present, and may be seen in conjunction with benign processes [37,39]. In one
Table 2 Modified ABC rule for the detection of subungual melanoma A B C D E F
small series [39], the dermatoscopic findings of hyponychial pigmentation of subungual MM and benign melanocytic nevi were different. In the former, pigment was distributed in a haphazard fashion over the entire surface, whereas in the latter, it had a pattern similar to acquired plantar melanocytic nevi (linear pigmentation in the furrows or across the skin marks). In children, subungual MM is rare and observation is usually recommended in longitudinal melanonychia except when there are significant changes with time [40]. This tumor is most likely to occur on the hallux, thumb, or index finger, but any digit can be involved. Although subungual MM generally only affects one digit, multiple MM can develop synchronously or metasynchronously. Levit et al [38] have suggested a modification of the ABCDs of melanoma to help in the diagnosis of subungual MM (Table 2). The histology of most subungual MM is acral lentiginous in type, with superficial spreading being the other form most often encountered. The diagnosis of subungual MM may be difficult because the features of MM in the radial growth phase may be subtle [37,41]. Most tumors display a lentiginous pattern with pleomorphic, often dendritic atypical melanocytes arranged in single cells over nests in the basal and suprabasal epithelial layers [37]. Immunohistochemical stains with S-100 and HMB-45 may help in arriving at the correct diagnosis. Serial and step sections are highly recommended. As for other MM, the prognosis is dependent on tumor thickness. Unfortunately, there is often a delay in diagnosis especially for toenail MM [38]. To circumvent this problem, Haneke and Baran [37] have suggested that any acquired longitudinal melanonychia and periungual pigmentation should be removed totally and examined by serial sections. For histologically confirmed subungual MM, the primary mode of therapy remains amputation [42]. Wide resection of the lesion with a free margin of 0.5 to 1 cm should only be considered for in situ disease [42], bearing in mind that this approach has not been studied long term.
Age: peak 5th to 7th decades Band (nail band): brown to black pigment, breadth > 3 mm, border irregular/blurred Change: rapid increase in size/growth of nail band. Failure of nail dystrophy to improve with adequate treatment Digit: thumb > hallux > index finger; single digit > multiple digits Extension: pigment spread to involve the proximal or lateral nail fold (Hutchinsons sign) Family: history of MM or atypical nevus syndrome
Modified from Levit EK, Kagen MH, Scher RK, et al. The ABC rule for clinical detection of subungual melanoma. J Am Acad Dermatol 2000;42: 269 274; with permission.
732
J.K. Rivers / Dermatol Clin 20 (2002) 727733 [18] Quinn MJ, Crotty KA, Thompson JF, et al. Desmoplastic and desmoplastic neurotropic melanoma: experience with 280 patients. Cancer 1998;83:1128 35. [19] Skelton HG, Smith KJ, Laskin WB, et al. Desmoplastic malignant melanoma. J Am Acad Dermatol 1995;32: 717 25. [20] Wharton JM, Carlson JA, Mihm MC. Desmoplastic malignant melanoma: diagnosis of early clinical lesions. Hum Pathol 1999;33:537 42. [21] Chan GSW, Choy C, Ng WK, et al. Desmoplastic malignant melanoma on the buttock of an 18-year-old girl: differentiation from desmoplastic nevus. Am J Dermatopathol 1999;21:170 3. [22] Schuchter L, Schultz DJ, Synnestvedt M, et al. A prognostic model for predicting 10-year survival in patients with primary melanoma. Ann Intern Med 1996;125: 369 75. [23] Robson A, Allen P, Hollowood K. S100 expression in cutaneous scars: a potential diagnostic pitfall in the diagnosis of desmoplastic melanoma. Histopathology 2001;38:135 40. [24] Koch MB, Ie-Ming S, Weiss SW, et al. Microphthalmia transcription factor and melanoma cell adhesion molecule expression distinguish desmoplastic/spindle cell melanoma from morphologic mimics. Am J Surg Pathol 2001;25:58 64. [25] Busam KJ, Iversen K, Coplan KC, et al. Analysis of microphthalmia transcription factor expression in normal tissues and tumors, and comparison of its expression with s-100 protein, gp100, and tyrosinase in desmoplastic malignant melanoma. Am J Surg Pathol 2001;25:197 204. [26] Hannah A, Feigen M, Quong G, et al. Use of (18F)fluorodeoxyglucose positron emission tomography in monitoring response of recurrent neurotrophic desmoplastic melanoma to radiotherapy. Otolaryngol Head Neck Surg 2000;122:304 6. [27] Goldenhersh MA, Savin RC, Barnhill RL, et al. Malignant blue nevus: case report and literature review. J Am Acad Dermatol 1988;19:712 22. [28] Ozgur F, Akyurek M, Kayikcioglu A, et al. Metastatic malignant blue nevus: a case report. Ann Plast Surg 1997;39:411 5. [29] del Rio E, Vazquez Veiga HA, Suarez Penaranda JM. Blue nevus with satellitosis mimicking malignant melanoma. Cutis 2000;65:301 2. [30] Sterchi JM, Muss HB, Weidner N. Cellular blue nevus simulating metastatic melanoma: report of an unusually large lesion associated with nevus-cell aggregates in regional lymph nodes. J Surg Oncol 1987;36:71 5. [31] Ota M. Nevus fuscoceruleus ophthalmomaxillaris. Jpn J Dermatol 1939;46:369 72. [32] Page DG, Svirsky JA, Kaugars GE. Nevus of Ota with associated palatal involvement. Oral Surg Oral Med Oral Pathol 1985;59:282 4. [33] Teekhasaenee C, Ritch R, Rutnin U, et al. Glaucoma in oculodermal melanocytosis. Ophthalmology 1990;97: 562 70. [34] Hartmann LC, Oliver GF, Winkelmann RK, et al. Blue
References
[1] Grin CM, Kopf AW, Welkovich B, et al. Accuracy in the clinical diagnosis of malignant melanoma. Arch Dermatol 1990;126:763 6. [2] Binder M, Kittler H, Steiner A, et al. Reevaluation of the ABCD rule for epiluminescence microscopy. J Am Acad Dermatol 1999;40:171 6. [3] Steiner A, Pehamberger H, Wolff K. In vivo epiluminescence microscopy of pigmented skin lesions. II. Diagnosis of small pigmented lesions and early detection of malignant melanoma. J Am Acad Dermatol 1987;17:584 91. [4] Westerhoff K, McCarthy WH, Menzies SW. Increase in the sensitivity for melanoma diagnosis by primary care physicians using skin surface microscopy. Br J Dermatol 2000;143:1016 20. [5] Kittler H, Pehamberger H, Wolff K, et al. Follow-up of melanocytic skin lesions with digital epiluminescence microscopy: patterns of modifications observed in early melanoma, atypical nevi, and common nevi. J Am Acad Dermatol 2000;43:467 76. [6] Seidenari S, Pellacani G, Giannetti A. Digital videomicroscopy and image analysis with automatic classification for detection of thin melanomas. Melanoma Res 1999;9:163 71. [7] Elbaum M, Kopf AW, Rabinovitz HS, et al. Automatic differentiation of melanoma from melanocytic nevi with multispectral digital dermoscopy: a feasibility study. J Am Acad Dermatol 2001;44:207 18. [8] Koch SE, Lange JR. Amelanotic melanoma: the great masquerader. J Am Acad Dermatol 2000;42:731 4. [9] Perry PK, Sliverberg NB. Cutaneous malignancy in albinism. Cutis 2001;67:427 30. [10] Streutker CJ, McCready D, Jimbow K, et al. Malignant melanoma in a patient with oculocutaneous albinism. J Cutan Med Surg 2000;4:149 52. [11] Mckenna DB, Cooper EJ, Kavanagh GM, et al. Amelanotic malignant melanoma following cryosurgery for atypical lentigo maligna. Clin Exp Dermatol 2000;25: 600 4. [12] Bono A, Maurichi A, Moglia D, et al. Clinical and dermatoscopic diagnosis of early amelanotic melanoma. Melanoma Res 2001;11:491 4. [13] Busam KJ, Hester K, Charles C, et al. Detection of clinically amelanotic malignant melanoma and assessment of its margins by in vivo confocal scanning laser microscopy. Arch Dermatol 2001;137:923 9. [14] Patel BCK, Egan CA, Lucius RW, et al. Cutaneous malignant melanoma and oculodermal melanocytosis (nevus of Ota): report of a case and review of the literature. J Am Acad Dermatol 1998;38:862 5. [15] Ariel IM. Amelanotic melanomas: an analysis of 77 patients. Curr Surg 1981;38:151 5. [16] Conley J, Lattes R, Orr W. Desmoplastic malignant melanoma (a rare variant of spindle cell melanoma). Cancer 1971;28:914 36. [17] Payne WG, Kearney R, Wells K, et al. Desmoplastic melanoma. Am Surg 2001;10:1004 6.
J.K. Rivers / Dermatol Clin 20 (2002) 727733 nevus and nevus of Ota associated with dural melanoma. Cancer 1989;64:182 6. Rivers JK, Bhayana S, Martinka M. Dural melanoma associated with ocular melanosis and multiple blue nevi. J Cutan Med Surg 2001;5:381 5. Theunissen P, Spincemaille G, Pannebakker M, et al. Meningeal melanoma associated with nevus of Ota: case report and review. Clin Neuropathol 1993;12: 125 9. Haneke E, Baran R. Longitudinal melanonychia. Dermatol Surg 2001;27:580 4. Levit EK, Kagen MH, Scher RK, et al. The ABC rule for clinical detection of subungual melanoma. J Am Acad Dermatol 2000;42:269 74. Kawabata Y, Ohara K, Hino H, et al. Two kinds of Hutchinsons sign, benign and malignant. J Am Acad Dermatol 2001;44:305 7.
733
[35]
[36]
[37] [38]
[39]
[40] Goettmann-Bonvallot S, Andre J, Belaich S. Longitudinal melanonychia in children: a clinical and histological study of 40 cases. J Am Acad Dermatol 1999;41: 17 22. [41] Tomizawa K. Early malignant melanoma manifested as longitudinal melanonychia: subungual melanoma may arise from suprabasal melanocytes. Br J Dermatol 2000;143:431 4. [42] OLeary JA, Berend KR, Johnson JL, et al. Subungual melanoma: a review of 93 cases with identification of prognostic variables. Clin Orthop 2000;378:206 12. [43] Pahernik S, Langer S, Botzlar A, et al. Tissue distribution and penetration of 5-ALA induced fluorescence in an amelanotic melanoma after topical application. Anticancer Res 2001;21:59 63.
Dermatol Clin 20 (2002) 735 747
Computer-aided melanoma diagnosis

Marek Elbaum, PhD
Electro-Optical Sciences, Inc., 1 Bridge Street, Suite 15, Irvington, NY 10533, USA
This article presents a snapshot of fast-moving research into the application of digital imaging technology to in vivo computer-aided diagnosis of early melanoma. It is written from the perspective of a researcher with a keen interest in bringing this technology to physicians offices. The primary emphasis is on reviewing the existing literature, supplemented by unpublished material relating to a particular system and as yet unpublished research studies. Early detection of melanoma (Breslow thickness [BT] [1] < 0.75 mm) permits complete cure, in almost all patients, through prompt excision of the neoplasm. The physicians arsenal for early melanoma detection consists of visual inspection (clinical view), guided by the ABCD rules [2], and dermoscopy (dermoscopic view) with its associated rules [3 5]. Dermoscopic images of lesions contain many more details of neoplasm features than do clinical images. Mastery of dermoscopic image patterns and their classification may lead to substantial improvement in diagnostic accuracy relative to that achieved through clinical imaging [6]. Only 23% of US dermatologists use dermoscopy (Tripp et al. Management of dysplastic nevi: a survey of Fellows of the American Academy of Dermatology and review of the literature, unpublished data), however, because its mastery requires training and its use is not reimbursed by most third-party payers. The relative merit of these
This work was supported in part by grants CA/AR60229, CA74628, and CA90029 from the National Cancer Institute to Electro-Optical Sciences, Inc., and by the Christopher Columbus Fellowship Foundation Award for 1998 to Dr Marek Elbaum, which support is gratefully acknowledged. The opinions expressed herein are those of the author and not necessarily those of any other person or organization. E-mail address: elbaum@eo-sciences.com (M. Elbaum).
two diagnostic methods remains controversial, given the lack of comprehensive multisite studies designed to compare directly the diagnostic performance of physicians who use only clinical view with those who use a combination of clinical and dermoscopic imaging on the same group of patients. Early detection of malignant melanoma is challenging, because there are a great many benign simulants (atypical moles) that cannot be differentiated reliably from melanoma, neither through clinical nor through dermoscopic methods [7,8]. Neoplasms that look suspicious to the physician undergo biopsy and are subjected to histopathologic evaluation. The pathology reports indicate that most such biopsies are of benign lesions. Determination of the degrees of atypia that warrant biopsy is highly subjective and dependent on the physicians training. Biopsy of all atypical moles in searching for melanoma is impractical, and cannot completely reduce the patients risk of developing melanoma [9,10]. Consequently, the motivation for development of computer-aided methods has been to assist physicians in improving the sensitivity and specificity in differential diagnosis of melanoma, and to provide new imaging modalities that enable in vivo histopathology. In all current computer-aided systems for diagnosis of melanoma, a solid state or optoelectronic scanning imaging probe is used to capture lesion images in vivo. Depending on the application, the probe is specialized for capturing images from fullbody view, clinical view, dermoscopic view, or histopathologic view. The fields of view for such probes are many dozens of centimeters for full-body view, up to 3 cm for clinical and dermoscopic views, and millimeters for histopathologic view. Their respective spatial resolutions across the lesion are millimeters for full-body probes, 5 to 50 mm for clinical and dermo-
736
M. Elbaum / Dermatol Clin 20 (2002) 735747
scopic probes, and a few to submicrometers for scanning probes. Probes for full-body and clinical views operate in visible light, as do all but two (MelaFind, Electro-Optical Sciences, Inc., Irvington, NY [8,11], and SIAscope [Astron Clinica, Cambridge, England] [12]) of the probes for dermoscopic view. Probes for histopathologic view typically operate in the near-infrared [13]. To facilitate diagnosis, computers are used to perform operations of various kinds on the digital images captured with these probes. This article reviews the status of the following six computeraided noninvasive diagnostic methods for early detection of melanoma: 1. Teledermatology: Uses images in clinical and dermoscopic views to assist primary care physicians (PCPs) in gaining fast access to a dermatologist for consultations, for education of patients, for training physicians, for consensus consultations, and for cost-effective point-of-care diagnostic service. 2. Mole mapping and monitoring: Uses images in full-body view to facilitate tracking by dermatologists of changes in lesions over time. 3. Automated diagnosis of melanoma: All but one [12] use dermoscopic images in the visible region and assist dermatologists in the image analysis process for differential diagnosis of melanoma. 4. Computer vision for automatic diagnosis of melanoma: Uses dermoscopic images in the visible and infrared to provide an objective diagnostic adjunct in the detection of melanoma. 5. Computer vision for noninvasive measurement of BT of melanomas: Uses dermoscopic images in the visible and infrared to measure accurately the BT in near-real time without biopsy. 6. Confocal microscopy for histopathology: Creates images with subcellular resolution to enable in vivo pathologic examination.
Teledermatology applications reported in the literature comprise: Referral of patients by PCPs to dermatology centers of excellence at a distance. Education of physicians in differential diagnosis of melanoma by providing Internet access to databases containing images of pigmented skin lesions (PSL), which are discussed by experts. Point-of-care diagnosis, a virtual laboratory, wherein a physician captures the images of a PSL with a digital camera, the image information is processed and reduced to a feature template that is sent by telephone line to a server at a different location, where this template is automatically compared with the distribution of templates derived from images of melanoma and nonmelanoma. The result of this comparison (lesion classification) is provided to the physician in a form suitable to assist in diagnosis (ie, the lesion is within the range of melanomas or the lesion is outside the range of melanomas). Patient referral by teledermatology can be conducted in real time, using video conferencing equipment including computers, or by store-and-forward methods, where computers are used for storing lesion images and the patients clinical history and for retrieval of these data [14]. Eedy and Wootton [15] have provided a comprehensive review of this application of teledermatology. The potential of teledermatology for physician education and for consultations among physicians has been demonstrated recently by a visual consensus meeting among 40 leading dermoscopists from Europe, the United States, Australia, and Japan [16]. This consensus conference on dermoscopy of PSLs was conducted by Internet from July 14 through November 8, 2000. This virtual conference, historic in the annals of dermatology, brought together internationally known experts in dermoscopy to assess the relative diagnostic merit of four different methods for analysis of dermoscopic images: (1) modified pattern recognition, (2) ABCD rules of dermoscopy, (3) Menzies method, and (4) the seven-point checklist [3 5,17 22]. The results of the study were based on 128 PSLs obtained from the archives of two European, one American, and one Australian clinical center [16]. None of the selected lesions had dimensions exceeding 14 10 mm, and each was imaged clinically and dermoscopically using Dermophot (Heine Optotechnik, Herrsching, Germany) with 10-fold magnifica-
Teledermatology The general paradigm is for a physician or an assistant to capture one or more digital images of a neoplasm at one office and then transmit the images to an office at another location, where the images are stored and then retrieved and displayed on a monitor for clinical evaluation. Rapid progress in digital image technology and high-bandwidth transmission has made teledermatology feasible and affordable.
737
tion. Good quality color film slides, the relevant clinical data, and histopathologic specimens were available for each of the 128 lesions. The color slides were converted to digital format using a Kodak photo CD system. All of the photo CD images (RGB [red/green/blue], 768 512, 72 dots per inch) were converted to JPEG format. Each image was processed in a standard manner (using Adobe Photoshop [Adobe Systems, San Jose, California]) to optimize its color, brightness, and contrast. A panel of six dermopathologists evaluated the histopathologic sections. According to these panelists, there were 33 melanomas; 70 different types of benign melanocytic lesions, including 36 melanocytic nevi; 10 basal cell carcinoma; and 15 other nonmelanocytic PSLs. The best results were obtained with the modified pattern analysis (sensitivity 83.7% and specificity 83.4%). The two algorithmic methods yielded similar sensitivity but lower specificity (11.9% and 13.4%, respectively). Through dermoscopy, 46.4% of the PSLs were judged to be benign and not to be excised [16].
Mole monitoring There are millions of high-risk melanoma patients with many (sometimes hundreds) abnormally appearing nevi. One cannot predict which of these benign lesions will progress to melanoma. Excision of all of these lesions is not practical, involves unnecessary surgery, and does not relieve the patient from the need for follow-up visits. It is generally agreed that high-risk patients with multiple atypical moles should have skin examinations several times a year [23,24]. For example, Rhodes at al [24] reported that a changing mole has 400 times higher relative risk factor for development of melanoma and a congenital mole has a relative risk factor of 21. In most patients with growing and changing moles, excision is the appropriate therapy. Kittler et al [25] used digital dermoscopy to monitor changes on 1862 melanocytic lesions from 202 patients over a median interval of 12.6 months. Approximately 4% of these lesions (75) were excised because they showed substantial modifications over time (enlargement, changes in shape, regression, color changes, or appearance of dermoscopic structures that are characteristic of melanoma). Eight changing lesions, on eight different patients, were histologically diagnosed as early melanomas. This study demonstrated the benefits to patients from monitoring changes of melanocytic lesions over
time. If, instead of monitoring all of the 1862 lesions, they had all been are excised, there would have been 232 excisions of benign lesions for each detected melanoma, and 194 patients out of 202 would have undergone an average of eight unnecessary surgical procedures. Practically, however, such monitoring can be tedious. Computers can be used for acquisition and surveillance with digital dermoscopes, as reported by Stolz et al [23]. Their computerized acquisition and surveillance system uses a skin-surface-microscope television camera. This system allows a comparison of current and previous images and measures the lesions area in 1 minute. The system allows rapid documentation of PSLs and shows simultaneously lesion images that have been captured at different times. Such computerized acquisition and surveillance systems can facilitate research and training of students in identifying those changes in dermoscopic images of lesions that facilitate early detection of melanoma, while keeping the specificity high. It is hoped that, as a result of this research, a set of rules will be developed to identify those changes that are indicative of risk of melanoma. Such results will no doubt stimulate research into automated methods for monitoring PSLs over time.
Automated analysis of dermoscopic images to assist human vision in differential diagnosis of melanoma Research into the applicability of automated methods of image analysis and image classification has been motivated by the desire to improve the sensitivity and specificity in differential diagnosis of early melanoma and atypical moles beyond that achieved solely through the physicians vision applied to clinical and dermoscopic images. In an automated method, the physician supervises what the computer does (unlike the fully automatic machine vision methods discussed in the next section). Of the six computer-aided methods reviewed in this article, automated analysis of dermoscopic images is the earliest and has been the most popular among researchers for two decades. Such methods comprise one or more of the following types of computer-assisted analysis of digital dermoscopic images of cutaneous lesions: image enhancement, image segmentation, feature extraction and quantification, and lesion classification [8,26]. In an automated analysis, a physician may decide that the captured image is of poor quality and reject it, whereas another physician may declare the same image to be acceptable for processing. As another example, the
738
physician may decide whether a border between the lesion and surrounding skin as drawn through the use of a mathematical (image segmentation) algorithm matches the border as perceived through his or her vision. If it does not, the physician can either reject the segmentation or modify it to his or her liking. Consequently, these automated methods lead to subjective diagnostic decisions. This field has been reviewed recently by Fleming [26], with emphasis on segmentation, feature extraction, and lesion classification, and by Menzies [27] from a broader perspective. Spectrophotometric intracutaneous analysis (SIA) is another automated technique for imaging PSL and assisting the diagnosis of melanoma [12]. This technique differs from the others described previously, in that eight images of a PSL are taken, each in a different spectrally narrow illumination band between 400 and 1000 nm (SIAscope). These images are processed with algorithms that the authors represent as describing the transport of light through the skin and the neoplasm of interest. The SIA algorithms are intended to provide three-dimensional distributions of chromophores (eg, melanin, blood, and collagen) within the papillary dermis and to locate the melanin relative to the dermoepidermal junction. The technique is said to create reproducible patterns with reliable features [12]. Using a combination of such features, two experimenters, blinded as to the histopathologic diagnosis of the lesion, achieved 82.7% sensitivity and 80.1% specificity on 348 PSLs that included 52 melanomas. The diagnostic merit of this technique relative to clinical and dermoscopic evaluation is yet to be determined by direct comparison. The present author shares the conclusion drawn by Menzies [27] that it is difficult to compare results reported by different groups using automated methods for diagnosis of melanoma. This difficulty arises
from lack of standardization of image capture hardware and of software for image analysis, and the diversity of clinical protocols. A number of research groups in collaboration with industry have developed automated diagnostic systems for detection of melanoma (eg, Fleming [26] and Menzies [27]), some of which are already on the market (eg, DermoGenius Ultra [Rodenstock Pra zisionsoptick, Munich, Germany], MicroDerm [Visiomed AG, Bochum, Germany], and MoleMax II [Derma Instruments, Vienna, Austria]). None of these systems, however, has received clearance from the US Food and Drug Administration (FDA) for marketing in the United States. Each of them uses a custom-designed imaging probe for capturing dermoscopic images in the visible region, and custom software is used for analysis of images captured with that probe. There are no reports in the literature of objective studies on the relative merits of these and other systems, many of which are still in the development stage.
Machine vision as an automatic diagnostic adjunct for early detection of melanoma In a computer machine vision method, formal mathematical rules (algorithms) are used to differentiate lesions, based on images captured with an electronic imaging probe. It is a quantitative and automatic method for the objective differentiation of melanoma from other look-alike pigmented lesions. As such, computer vision is a new paradigm for melanoma detection, in which a diagnostic score for the lesion is independent of the human perceptual apparatus. MelaFind is an example of such a computer vision system. All of the steps, including image
Fig. 1. Functional diagram of technology platform.
739
Fig. 2. MelaFind on-line service as a computer vision system.
capture, image segmentation, feature calculations, and lesion classification, are under computer control without interference by the physician. Development of machine vision The development process for machine vision is iterative, as sketched in the functional diagram in Fig. 1. A light-based sensor is used to acquire images of the PSL of interest. The sensor provides appropriate multispectral illumination and creates digital images. These images are stored in a database, together with associated pathology reports. Automatic image processing algorithms are applied to the images, including segmentation and feature estimation. Segmentation separates the lesion from the surrounding tissue. Feature estimation quantitatively characterizes the organization of the tissue. For example, melanoma tends to be more heterogeneous than benign lesions. The computer search engine rapidly selects the best combination of features for classification (differentiation of cancerous from benign lesions), through an iterative process in which the outcome of the classifier is matched against the pathology report (the gold standard). If the correctness of the match does not meet the desired performance standard, the search engine selects other features or the sensor is redesigned to modify the spectral illumination. Once the performance of the match predictably meets the design goals, the result is the computer vision system, which consists of the sensor and its associated expert software (image processing algorithms and classifier). Automatic point-of-care diagnostic service The flow diagram in Fig. 2 shows how the previously described technology lends itself to devel-
opment of an on-line computer vision system that can be used as a point-of-care service. The handheld imaging probe The MelaFind probe is a hand-held smart electro-optical device, as shown in Fig. 3. It consists of hardware for capturing images with illumination in various spectral bands, and software for analyzing these images. The illumination of the lesion and the capture of its multispectral image representation are under microprocessor control. The probe uses 10 distinct spectral bands, distributed over the visible and near-infrared portions of the electromagnetic spectrum, from 430 (blue) through 950 nm (nearinfrared) [8]. During imaging, the MelaFind probe creates a sequence of 10 dermoscopic images of lesions in less than a second. Each image is obtained with a different spectral illumination band. Consequently, each image in the sequence is obtained with light emerging from different depths of the lesion and surrounding skin, the depth increasing with the wavelength. These images are distinct spatial maps of absorption and scattering by a lesion, and collectively they provide unique information about the lesion border, size, and morphology. Because MelaFind images in both the visible and infrared regions, it provides information not available to the naked eye [8]. The MelaFind probe informs the operator whether the captured images meet the quantitative standards (no air bubbles or hair blocking the lesion, level of illumination adequate for the patients skin type, and so forth) necessary for further machine-vision processing. This is accomplished through a group of image-processing algorithms resident in the probe. Another group of algorithms embedded in the probe quantitatively characterizes the tissue pattern organi-
740
first examined to verify the state of calibration of the probe. If the data are deemed acceptable, the classifying algorithms are activated. The classifier assigns a numeric score to the imaged lesion, and compares the score with those in the database that was used to train the classifier. The server then returns these findings to the clinician and records the service performed to the physicians account. Because the diagnostic service is on-line continuously, the service is available to the examining physician continually, unlike a hospital pathology department, which typically returns its report of a lesion biopsy 1 to 7 days after receipt of tissue samples. The knowledge base The diagnostic performance of the system can be improved as the knowledge base increases (as indicated in Fig. 2). The keys to such improvement are the continually increasing database and the computer search engine that automatically seeks more reliable classifiers as the size of the database increases. Current status of MelaFind After 4 years of clinical studies with MelaFind prototypes [7,8,11,28], commercial systems are being tested under identical protocols at the first 15 of a projected 28 clinical sites across the United States (Table 1), with the goal of generating the data needed to support a market clearance application to the FDA. These clinical trials are to demonstrate at least one of the following three hypotheses concerning MelaFind: 1. Sensitivity and specificity higher than that of PCPs. 2. Sensitivity at least as high as that of dermatologists, but with higher specificity. 3. BT accuracy at least as high as that achieved by pathologists through current practice (using step-sectioning of the tissue every 2 to 3 mm). If these clinical trials succeed, the proposed labeling for MelaFind will be as a diagnostic adjunct for noninvasive detection of melanoma and for estimating BT.
Fig. 3. The MelaFind 100 System. (A) System components arranged in carrying case. (B) Handheld probe applied to patient.
zation [8]. First, the lesion boundary is demarcated. Next, the features are extracted and quantified; these represent a compressed form of essential information for lesion classification. The compression from megabytes to kilobytes allows the numerical data to be transmitted to the remote server by ordinary low bandwidth telephone lines. In the portable MelaFind 100 system (see Fig. 3), the handheld probe is connected to a laptop computer during image acquisition for database enhancement. This system provides reproducible results that are independent of the position or orientation of the lesion in the field of view. For example, measurements of lesion area, diameter, and perimeter, determined automatically from the images, had relative errors of only about 6%, 3%, and 4%, respectively, when each lesion was imaged independently 10 times [11]. The server The numerical data from the probe are transmitted to a remote server. At the server, the incoming data are
Machine vision for in vivo automatic measurement of BT The BT of a cutaneous melanoma is one of the most accurate and reliable indicators of prognosis of this cancer and is widely used for deciding on surgical margins, lymph node biopsy, and patient
M. Elbaum / Dermatol Clin 20 (2002) 735747 Table 1 Advisors and clinical collaborators in MelaFind testing Last Name Medical directors Cognetta Rabinovitz Technical-director Gutkowicz-Krusin Science advisory board Callen Kopf Mihm Rigel Sober Clinical collaborators Callen Cognetta Dinehart Duvic Elgart Friedman Goldman Grin Gross Halpern Lee Levine Monheit Moy Nestor Otley Peck Polsky Rabinovitz Rao Robins Rogers Schwartz Thomas Wolfe Dermatopathologists Mihm, Jr Prieto Googe King First Name Armand Harold Dina Jeffrey Alfred W Martin (Chair) Darrell Arthur Jeffrey Armand Scott Madeline George Robert Mitchel Caron Kenneth Allan Peter Norman Gary Ronald Mark Clark Gary David Harold Babar Perry Gary Jennifer Nancy Jonathan Martin Victor Paul Roy Degree MD MD PhD MD MD MD MD MD MD MD MD MD MD MD MD MD MD MD MD MD MD MD MD MD MD MD, PhD MD MD MD MD MD MD MD MD MD MD MD Location Dermatology Associates of Tallahassee, Tallahassee, FL Skin and Cancer Associates, Plantation, FL Electro-Optical Sciences, Inc., Irvington, NY University of Louisville / Associates in Dermatology, Louisville, KY NYU Medical Center, New York, NY Massachusetts General Hospital, Boston, MA Rigel Dermatology Group, New York, NY Massachusetts General Hospital, Boston, MA
741
University of Kentucky / Associates in Dermatology, PLCC, Louisville, KY Dermatology Associates of Tallahassee, Tallahassee, FL Bressinck-Gibson-Parker-Dinehart-Sangster Dermatology, PA, Little Rock, AR University of Texas, MD Anderson Cancer Center, Houston, TX University of Miami School of Medicine, Miami, FL Private Practice, New York, NY Dermatology Associates, La Jolla, CA University of Connecticut Health Center, Farmington, CT Skin Surgery Medical Group, Inc., San Diego, CA Memorial Sloan-Kettering Cancer Center, New York, NY University of Minnesota, Minneapolis, MN University of Arizona, Tucson, AZ Dermatology Associates, Birmingham, AL UCLA / Private Practice, Los Angeles, CA Skin and Cancer Associates, Aventura, FL Mayo Clinic Rochester, Rochester, MN Washington Cancer Institute, Washington Hospital Center, Washington, DC NYU Medical Center, New York, NY Skin and Cancer Associates, Plantation, FL Robert Wood Johnson Medical School, New Brunswick, NJ NYU Medical Center, New York, NY New England Medical Center, Tufts University, Boston, MA University of Michigan, Ann Arbor, MI University of North Carolina, Chapel Hill, Chapel Hill, NC Burgoon Mackay and Schuler, Plymouth Meeting, PA Massachusetts General Hospital, Boston, MA University of Texas, MD Anderson Cancer Center, Houston, TX Knoxville Dermapathology Laboratory, Knoxville, TN Knoxville Dermapathology Laboratory, Knoxville, TN
follow-up strategies [1,29]. For cutaneous melanomas, both the diagnosis of the lesion and the BT determination require a biopsy. At present, there are no established tools to allow physicians to determine the tumor thickness before obtaining a biopsy. Current standard practice in the measurement of the BT is to examine step-sections of the biopsy taken
every 2 to 3 mm. Such standard histologic sections sample only a small part of the lesion, however, and this may lead to a significant underestimate of the BT [30]. Measurements of the BT based on sequential serial sections, each about 5 mm thick, were compared with those based on step-sections by Solomon et al [30] for 19 thin ( < 0.76 mm in thickness) melano-
742
mas. When such serial sections were examined, the increase in the measured thickness ranged from 4% to 115% compared with the step-section measurements alone. The effect of undersampling on the measured values of BT could be very significant, at least for thin lesions. The effect of undersampling on BT determination for thick lesions has not been discussed in the literature. One noninvasive method for the clinical estimation of melanoma thickness is palpation. A study by ODonnell et al [31], however, determined that this method is not reliable. Several groups have investigated the use of high-frequency ultrasound for noninvasive preoperative evaluation of melanomas [32 34]. One difficulty in measuring the tumor thickness with ultrasound was the inability to distinguish between the neoplasm and lymphocytic infiltrate. In addition, many malignant melanomas are associated with benign melanocytic nevi, and ultrasound does not distinguish between them. In part because of these difficulties, whereas the reported correlations between the ultrasound and histologic thickness are high, there are significant differences in the actual values obtained. Gutkowicz-Krusin of Electro-Optical Sciences, Inc., has recently developed software (MelaMeter) that can noninvasively estimate the BT through analysis of multispectral images of the lesion acquired with a
MelaFind imaging probe. The algorithms in the software differentiate between invasive and in situ melanomas and estimate the BT of invasive melanomas. The sequential flow of image data is shown schematically in Fig. 4 (proposal to National Cancer Institute for Grant No. CA90029). The database used in a study of the feasibility of this approach consisted of 39 invasive (superficial spreading) melanomas and 31 in situ cutaneous melanomas acquired at three medical centers. Only lesions for which the two study dermatopathologists had a concordant diagnosis of melanoma were included in the study. BT was determined histologically for all the invasive melanomas from the 2-mm step-sections by one of the two dermatopathologists. The BT ranged from 0.15 to 8.4 mm, and the median was 0.45 mm. For each lesion, before obtaining a biopsy, digital images were acquired with a MelaFind prototype. A 12-parameter linear classifier (M) with a threshold value of unity was developed to differentiate invasive from in situ melanomas. The scores for all the lesions in the data base are shown in Fig. 5 (unpublished report to National Cancer Institute on Grant No. CA90029-01). Lesions with scores greater than or equal to unity are classified as invasive melanomas; those with scores less than unity, as in situ melanomas. Although resubstitution test leads to 100% classification accuracy for all lesions, the limited
Fig. 4. Relationship between MelaFind and MelaMeter.
743
Fig. 5. A 12-parameter linear classifier to differentiate invasive from in situ melanoma. This classifier was trained on a database of 39 invasive and 31 in situ melanomas. A lesion is declared to be invasive melanoma if the classifier score equals or exceeds the threshold value of 1, and in situ melanoma otherwise. This classifier correctly diagnosed all the lesions in the database. At 95% confidence level, the classification accuracy is greater than 93% for the invasive and greater than 91% for the in situ melanomas.
size of the database results in uncertainties in classification accuracy. At the 95% confidence level, the classification accuracy is greater than 93% for the invasive and greater than 91% for the in situ melanomas. An estimator (Q), which is a linear combination of eight lesion parameters, was found to correlate very well with the BT for all 39 invasive melanomas, as shown in Fig. 6 (unpublished report to National Cancer Institute on Grant No. CA90029-01). The correlation of 0.99 is highly significant ( P < 10 40). The correlation is high because of good agreement with the
histologic BT for thick lesions (> 1 mm). Another way of assessing the quality of the estimator is by determining relative errors, defined as the difference between the computer-estimated and the histologic thickness, divided by the latter, which serves as the gold standard. For all 39 invasive melanomas, the average relative error in estimating BT is 38%, whereas for the eight lesions with BT greater than 1 mm, the average relative error is only 11%. A different estimator (Q1), which is another linear combination of eight lesion parameters, was found to provide a much better estimate of the BT for thin lesions, as shown in Fig. 7. The average relative error in determining BT was 22%, with the largest errors
Fig. 6. Estimator Q (a linear combination of eight lesion parameters) for invasive melanoma versus Breslow thickness. Correlation is 0.99 with P < 10 40. This correlation is dominated by the thickest lesions (Breslow thickness > 1 mm).
Fig. 7. Estimator Q1 (a linear combination of eight lesion parameters) for invasive melanoma versus Breslow thickness for thin lesions (Breslow thickness < 1 mm). Correlation is 0.93 with P = 5x10 19.
744
Fig. 8. Blind test of MelaMeter estimates of Breslow thickness on 19 melanomas versus that provided by histopathology on standard step-sections through the lesions. Correlation is 0.74 with P = 3x10 4. The thickness of 6 of the 10 in situ melanomas was correctly estimated as 0 mm.
coming from very thin melanomas (BT < 0.2 mm). The average error in the range from 0.2 to 1 mm was about 15%. In a subsequent study (unpublished report to National Cancer Institute on Grant No. CA90029-01) the BT estimates provided by this method were compared with the BT provided by dermatohistopathology on standard step-sections (Fig. 8). Of the 10 lesions diagnosed as melanomas in situ by standard histology, MelaMeter determined that six are in situ (BT = 0 mm), whereas the other four had BT ranging from 0.21 to 1.54 mm. Of the nine invasive melanomas by standard histology, the difference between the MelaMeter BT and standard histologic BT ranged from -0.63 to 1.22 mm. The significance of these results can be understood through Fig. 9, which demonstrates the errors caused by undersampling in standard histologic BT. The errors in the BT values estimated noninvasively through MelaMeter are not greater than the errors in the BT furnished by standard histopathology.
laser), which illuminates a small spot within the biologic specimen. Light scattered back from the spot is detected with contrast caused by naturally occurring variations of the refractive index of tissue microstructures. The scattered light is then imaged onto a detector through a pinhole spatial filter that rejects light reflected from out-of-focus portions of the object (confocal means that the light source, illuminated spot, and detector are located in conjugate focal planes). In CSLM, a computer controls image capture, scanning the spot of light so as to form a high-contrast image of a thin section (up to 3 mm thick, with approximately 1-mm transverse resolution). By serially scanning at different depths, images of different thin sections can be obtained, and stored for later retrieval. Busam et al [35] have reported using CSLM to detect clinically amelanotic malignant melanoma and assessment of its margins in two lesions (from two patients). Based on this initial result, the authors conclude that CSLM technique may aid in the early detection of clinically barely visible or nonpigmented melanomas and may facilitate preoperative noninvasive assessment of their margins. In another study, Langley et al [36] used nearinfrared CSLM to image 40 PSLs, and compared these images with the histopathology. They found that nuclear, cellular, and architectural details in the
Confocal microscopy for histopathologic diagnosis of melanoma Near-infrared confocal scanning laser microscopy (CSLM) is a new imaging modality for in vivo microscopic analysis of skin lesions [13]. The confocal digital imager uses a point source of light (from a
Fig. 9. Thickness values for 14 melanomas, as determined from step-sections every 0.5 mm, compared with those determined from standard sections (every 2 to 3 mm). The mean difference between fine-section Breslow thickness and standard histologic Breslow thickness was 0.16 mm (standard deviation 0.47 mm). These are not significantly different from the 0.19-mm mean difference and 0.57-mm standard deviation between MelaMeter Breslow thickness and standard histologic Breslow thickness.
745
epidermis and superficial dermis had high resolution and contrast. The authors conjecture that CSML may be useful to noninvasively discriminate benign and malignant lesions in situ.
Discussion In the first five of the six methods discussed in this article, computer-aided diagnosis of melanoma has progressed beyond the feasibility study stage. Some progress has been made toward commercial use in the areas of teledermatology, mole mapping, and computer-aided dermoscopes. In this authors opinion, the two methods having the highest potential for significant improvement in early detection and treatment of melanoma are multispectral computer vision and near-infrared CSLM. The CSLM images can be understood by analogy to the images of histopathologic slides. The statistical features of multispectral images that play a role in machine vision do not provide as ready a parallel to current practice. The CSLM technique enables visualization of the lesion at subcellular resolution, down to the superficial dermis. At present, the challenge is to correlate the CSLM images with known histopathology. The data content of CSLM images is much higher than that of dermoscopic images. One must first learn how to extract reliably the information that is central to the differential diagnosis, however, and then decide whether this proves to be excessively tedious to be cost effective in the practical environment. It may become necessary to use automated machine vision analysis if CSLM is to provide practical assistance to the physician. Multispectral computer vision, by contrast, provides a second opinion at the click of a button. Both of these techniques would be major market innovations, and as such must overcome a number of barriers before becoming practical parts of the physicians arsenal for early detection of melanoma. Barriers to the market include the need for FDA clearance, and to secure reimbursement for use by the Centers for Medicare and Medicaid Services (which serves 75 million Americans) or private thirdparty payers. The processes to overcome these barriers are time-consuming and expensive. Unless the medical community involved in melanoma detection demonstrates a perception of the need for progress in computer-aided diagnostic methods, such methods will remain subjects of research, without yielding practical benefits. The first step needed in bringing computer-aided diagnostic methods to medical practice is to establish the current (state-of-
the-art) diagnostic benchmarks of the accuracy of the dermatologists and PCPs who perform melanoma detection based on clinical and dermoscopic examination. Regrettably, there are no reliable benchmarks in the form of multicenter sensitivity and specificity analyses (eg, using receiver-operator curve methodology [8,37]). Most of the research papers published on this subject come from individual institutions, and the protocols vary between institutions. Recently, progress toward establishing reliable benchmarks was made through the 2000 Consensus Meeting by the Internet [16]. The reported results on sensitivity and specificity for detection of melanoma by the 40 leading practitioners of dermoscopy in that study set the currently highest bar against which to judge the performance of computer-aided diagnostic methods. Reliable benchmarks serve to quantify the potential for improvements. The Centers for Medicare and Medicaid Services will likely provide reimbursement for a computer-aided diagnostic method only if there is a demonstrable increase in sensitivity (without compromise in specificity) relative to such benchmarks. When evaluations by the regulatory agencies authenticate the device-makers claims, medical practice will start embracing the technology.
Acknowledgments The following clinical collaborators and their colleagues participated in the collection of the lesion image data reported here: Alfred W. Kopf, MD, and David Polsky, MD, PhD (NYU School of Medicine, New York, NY); Harold S. Rabinovitz, MD (Skin and Cancer Associates, Plantation, FL); Armand B. Cognetta, Jr, MD (Dermatology Associates, Tallahassee, FL); Gary L. Peck, MD (Washington Cancer Institute, Washington, DC); Richard G.B. Langley, MD, and Arthur J. Sober, MD (Harvard Medical School, Boston, MA). The study histopathologists were Hideko Kamino, MD (NYU School of Medicine) and Martin C. Mihm, Jr, MD (Harvard Medical School). The author also acknowledges contributions by his colleagues at Electro-Optical Sciences, Inc.: Alexandru Bogdan, ScD, Dina Gutkowicz-Krusin, PhD, Sunguk Keem, PhD, Michael Greenebaum, PhD, Adam Jacobs, Nick Kabelev, and Joanna Melman.
References
[1] Breslow A. Tumor thickness, level of invasion, and node dissection in stage 1 cutaneous melanoma. Ann Surg 1975;182:572 5.
746
M. Elbaum / Dermatol Clin 20 (2002) 735747 surface microscopy. J Am Acad Dermatol 1990;23 (6 pt 1):1159 62. Menzies SW, Ingvar C, Crotty KA, et al. Frequency and morphologic characteristics of invasive melanomas lacking specific surface microscopic features. Arch Dermatol 1996;132:1178 82. Argenziano G, Soyer HP, De Giorgi V, et al. Interactive atlas of dermoscopy. Milan: Edra Medical Publishing and New Media; 2000. Schiffner R, Schiffner-Rohe J, Vogt T, et al. Improvement of early recognition of lentigo maligna using dermatoscopy. J Am Acad Dermatol 2000;42(1 pt 1): 25 32. Saida T, Oguchi S, Ishihara Y. In vivo observation of magnified features of pigmented lesions on volar skin using video macroscope: usefulness of epiluminescence techniques in clinical diagnosis. Arch Dermatol 1995;131:298 304. Oguchi S, Saida T, Koganehira Y, et al. Characteristic epiluminescent microscopic features of early malignant melanoma on glabrous skin: a videomicroscopic analysis. Arch Dermatol 1998;134:563 8. Stolz W, Schiffner R, Pillet L, et al. Improvement of monitoring of melanocytic skin lesions with the use of a computerized acquisition and surveillance unit with a skin surface microscopic television camera. J Am Acad Dermatol 1996;35(2 pt 1):202 7. Rhodes AR, Weinstock MA, Fitzpatrick TB, et al. Risk factors for cutaneous melanoma: a practical method of recognizing predisposed individuals. JAMA 1987; 258:3146 54. Kittler H, Pehamberger H, Wolff K, et al. Follow-up of melanocytic skin lesions with digital epiluminescence microscopy: patterns of modifications observed in early melanoma, atypical nevi, and common nevi. J Am Acad Dermatol 2000;43:467 76. Fleming MG. Digital dermoscopy. Dermatol Clin 2001;19:359 67. Menzies SW. Automated epiluminescence microscopy: human vs machine in the diagnosis of melanoma. Arch Dermatol 1999;135:1538 40. Gutkowicz-Krusin D, Elbaum M, Szwajkowski P, Kopf AW. Can early malignant melanoma be differentiated from atypical melanocytic nevi by in-vivo techniques? Part II: Automatic machine vision classification. Skin Res Tech 1997;3:15 22. National Institutes of Health Consensus Conference. Diagnosis and treatment of early melanoma. JAMA 1992;268:1313 9. Solomon AR, Ellis CN, Headington JT. An evaluation of vertical growth in thin superficial spreading melanomas by sequential serial microscopic sections. Cancer 1983;52:2338 41. ODonnell BF, Marsden JR, ODonnell CA, et al. Does palpability of primary cutaneous melanoma predict dermal invasion? J Am Acad Dermatol 1996;34: 632 7. Partsch B, Binder M, Puspok-Schwarz M, et al. Limitations of high frequency ultrasound in determining
[2] Friedman RJ, Rigel DS, Kopf AW. Early detection of malignant melanoma: the role of physician examination and self-examination of the skin. Cancer J Clin 1985;35:130 51. [3] Pehamberger H, Binder M, Steiner A, et al. In vivo epiluminescence microscopy: improvement of early diagnosis of melanoma. J Invest Dermatol 1993;100: 356S 62S. [4] Stolz W, Riemann A, Cognetta AB, et al. The ABCD rule of dermoscopy: a new practical method for early recognition of malignant melanoma. Eur J Dermatol 1994;7:521 8. [5] Argenziano G, Fabbrocini G, Carli P, et al. Epiluminescence microscopy for the diagnosis of doubtful melanocytic skin lesions: comparison of the ABCD rule of dermatoscopy and a new 7-point checklist based on pattern analysis. Arch Dermatol 1998;134):1563 70. [6] Binder M, Schwarz M, Winkler A, et al. Epiluminescence microscopy: a useful tool for the diagnosis of pigmented skin lesions for formally trained dermatologists. Arch Dermatol 1995;131:286 91. [7] Kopf AW, Elbaum M, Provost N. The use of dermoscopy and digital imaging in the diagnosis of cutaneous malignant melanoma. Skin Res Tech 1997;3: 1 7. [8] Elbaum M, Kopf AW, Rabinovitz HS, et al. Automatic differentiation of melanoma from melanocytic nevi with multispectral digital dermoscopy: a feasibility study. J Am Acad Dermatol 2001;44:207 18. [9] Tiersten AD, Grin CM, Kopf AW, et al. Prospective follow-up for malignant melanoma in patients with atypical-mole (dysplastic-nevus) syndrome. J Dermatol Surg Oncol 1991;17:44 8. [10] Kittler H, Binder M. Risks and benefits of sequential imaging of melanocytic skin lesions in patients with multiple atypical nevi. Arch Dermatol 2001;137: 1590 5. [11] Gutkowicz-Krusin D, Elbaum M, Jacobs A, et al. Precision of automatic measurements of pigmented skin lesion parameters with a MelaFind(TM) multispectral digital dermoscope. Melanoma Res 2000;10:563 70. [12] Moncrieff M, Cotton S, Claridge E, et al. Spectrophotometric intracutaneous analysis: a new technique for imaging pigmented skin lesions. Br J Dermatol 2002;146:1 11. [13] Selkin B, Rajadhyaksha M, Gonzalez S, et al. In vivo confocal microscopy in dermatology. Dermatol Clin 2001;19:369 77. [14] Provost N, Kopf AW, Rabinovitz HS, et al. Comparison of conventional photographs and telephonically transmitted compressed digitized images of melanomas and dysplastic nevi. Dermatology 1998;196: 299 304. [15] Eedy DJ, Wootton R. Teledermatology: a review. Br J Dermatol 2001;144:696 707. [16] Argenziano G, Soyer HP, Chimenti S, Talamini R. et al. Dermoscopy of pigmented skin lesions: results of a consensus meeting via the Internet. [17] Bahmer FA, Fritsch P, Kreusch J, et al. Terminology in
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26] [27]
[28]
[29]
[30]
[31]
[32]
M. Elbaum / Dermatol Clin 20 (2002) 735747 the invasiveness of cutaneous malignant melanoma. Melanoma Res 1996;6:395 8. [33] Tacke J, Haagen G, Hornstein OP, et al. Clinical relevance of sonometry-derived tumour thickness in malignant melanoma: a statistical analysis. Br J Dermatol 1995;132:209 14. [34] Argenziano G, Fabbrocini G, Carli P, et al. Clinical and dermatoscopic criteria for the preoperative evaluation of cutaneous melanoma thickness. J Am Acad Dermatol 1999;40:61 8.
747
[35] Busam KJ, Hester K, Charles C, et al. Detection of clinically amelanotic malignant melanoma and assessment of its margins by in vivo confocal scanning laser microscopy. Arch Dermatol 2001;137:923 9. [36] Langley RG, Rajadhyaksha M, Dwyer PJ, et al. Confocal scanning laser microscopy of benign and malignant melanocytic skin lesions in vivo. J Am Acad Dermatol 2001;45:365 76. [37] Hagen D. Test characteristics: how good is that test? Prim Care 1995;22:213 33.
Dermatol Clin 20 (2002) 749 761
Update on botulinum toxin for facial aesthetics

Jennifer Clay Cather, MDa,b,*, J. Christian Cather, MDa, Alan Menter, MDa,b
a
Texas Dermatology Research Institute, 5310 Harvest Hill Road, Suite 260, Dallas, TX 75230, USA b Baylor Medical University, 3500 Gaston Avenue, Dallas, TX 75246, USA
Clostridium botulinum was implicated over 100 years ago as the cause of muscle paralysis secondary to food poisoning [1]. This gram-positive anaerobic bacterium produces the most potent neurotoxins known to mankind. Seven distinct antigenic botulinum toxins (A, B, C1, D, E, F, and G), produced by different strains of C. botulinum, have been described. Human disease is caused by five of these serotypes (A, B, E, F, and G) [2]. Type A is the strongest, followed by types B and F, which are potentially of value in patients who develop antibodies to type A [3,4]. Botulinum toxin types E and F are short acting [5]. Botulinum toxin is synthesized as a single-chain protein, which is inactive until it is cleaved by bacterially produced proteases into its active form. All active botulinum toxins are comprised of two chains: one heavy chain joined to one light chain by a disulfide bond. Clostridium botulinum has been in therapeutic use since the 1970s. Scott [6] is largely responsible for the initial clinical use of botulinum toxin in the treatment of strabismus where botulism toxin injected into extraocular muscles results in selective muscle paralysis and improved ocular alignment. Since that time, numerous other conditions including dystonias (ie, blepharospasm and torticollis), involuntary muscle hyperactivity (ie, hemifacial spasm, tremors, and tics), and spasticity (as in multiple sclerosis and
cerebral palsy), have been treated successfully with botulinum toxin [7]. Facial rejuvenation has been revolutionized by the use of botulinum toxin over the past 15 years with safe use by clinicians using botulism toxin type A producing excellent cosmetic results [5,8,9].
The toxins: Botox, Dysport, and Myobloc There are three sources of botulinum toxin commercially available. Type A botulinum toxins include Botox (Allergan, Inc., Irvine, CA) and Dysport (Ipsen Limited, Maidenhead, Berkshire, UK). One type B botulinum toxin is currently available: Myo(neuro)bloc (Neurobloc in the United States) (Elan Pharmaceuticals, South San Francisco, CA). Currently, in the United States, only Myo(neuro)bloc and Botox are available. Botox was approved by the Food and Drug Administration (FDA) for ophthalmic indications in 1989 and for cervical dystonia in December 2000. The FDA approved Botox for the treatment of glabellar wrinkles on April 15, 2002. Myo(neuro)bloc was approved by the FDA for the treatment of cervical dystonia in December 2000. Mechanism of action Botulinum toxins act by a three-step process: (1) binding, (2) internalization by receptor-mediated endocytosis, and (3) enzymatic activation (Fig. 1). Different serotypes bind to different sites on the motor nerve terminal and within the motor neuron. The heavy chain mediates the selective saturable binding to the presynaptic cholinergic neuromuscular end plate. The light chain acts inside the cell to block
* Corresponding author. Texas Dermatology Research Institute, 5310 Harvest Hill Road, Suite 260, Dallas, TX 75230. E-mail address: amresearch@texasderm.com (J.C. Cather).
750
J.C. Cather et al. / Dermatol Clin 20 (2002) 749761
Fig. 1. Botulinum toxin mechanism of action at motor end plate.
the release of the vesicle-bound neurotransmitter acetylcholine from autonomic nerve endings [10]. Botulinum toxins have specific light-chain intracellular binding sites and different sites of action on different SNARE (synaptosomal associated protein receptor [SNAP]) proteins, which is a neuronal exocytoxic apparatus that regulates the membrane docking and fusion of synaptic vesicles and the release of acetylcholine. The SNARE proteins, which include vesicle-associated membrane protein (VAMP), syntaxin, and SNAP-25, are intimately involved in releasing acetylcholine at presynaptic terminals [11]. Botulinum toxins type A, C1, and E cleave SNAP-25 [12]. Botulinum toxins type B, D, F, and G all cleave synaptobrevin (VAMP) each at different sites [13]. Synaptobrevin is an intracellular protein necessary for the fusion of the synaptic vesicle to the cell membrane leading to the eventual release of acetylcholine into the neuromuscular junction [14]. Myo(neuro)bloc acts on synaptobrevin at Gin76-Phe77. Restoration of muscle activity usually commences between 3 and 4 months after injection probably because of the regeneration of new end plate units [15]. The toxicity of all botulinum toxins is expressed in units. One unit is the amount of toxin required to kill 50% of a group of female Swiss-Webster mice weighing 18 to 20 g after intraperitoneal injection [16]. It is essential to recognize that units vary greatly between the different botulinum toxin preparations. In the literature the approximate equivalencies are as follows: 1 U Botox = 3 to 5 U Dysport = 50 to 100 U Myo(neuro)bloc [11,15,17,18]. The commercial product name is mandatory for medical communica-
tions addressing specific applications to avoid complications related to dosing. The lethal dose of Botox for humans is estimated to be between 2500 and 3000 U for a 70-kg person [7,19,20]. The administration of the entire contents of one vial of Botox (ie, 100 U) is way below the lethal dose for a human and because most cosmetic applications usually require less than 30 U per session, the safety margin is great [7,19,20]. Botox is a protein that is capable of inducing the formation of IgG-neutralizing antibodies [21]. These antibodies are more likely to form with larger protein loads per session. Neutralizing antibodies have only been documented in the neurologic literature in cases involving more than 100 units or more of Botox per injection session [22]. The incidence of antibody production is estimated to be between 3% and 5% in neurology cases; however, no antibody resistance has been reported to date in dermatologic cases using Botox [23]. It is wise for clinicians using Botox for cosmetic purposes to limit the total amount of Botox injected to less than 100 units per session [22]. The antigenic potential of Botox was further decreased when a new batch was released in 1997 with a lower protein load per dose [24]. In patients with neutralizing antibodies to Botox and no clinical improvement after injection, the serologically distinct Myo (neuro)bloc has been found effective in neurologic cases [13,18,25]. Neutralizing antibodies to type A toxins do not cross-neutralize the activity of type B [26]. The antigenic potential of Myo(neuro)bloc is unknown, although 50- to 100-fold higher doses are required resulting in a 10 to 20 times larger protein load per dose [5].
751
Preparation of toxins Preparation of Botox Botox is available in vials containing 100 U vacuum-dried type A toxin, albumin, and sodium chloride, which must be kept frozen at 5C or colder until reconstituted. Most clinicians use between 1 and 3 mL of saline to reconstitute Botox [27]. To produce a preparation of 5 U per 0.1 mL, 2 mL of preservative-free saline is mixed carefully and slowly into a 100-U vial of Botox. Acceptable results have also been reported by diluting one vial in up to 10 mL preservative-free saline (1 U/0.1 mL) [28]. If vigorously mixed, inactivation of toxin may result in decreased potency [2]. Once diluted, the mixture should be stored in a refrigerator at 2 to 8C and used within 4 hours [15,29], although some authors have found the potency to decrease slowly over a 1-week period, with minimal to no activity remaining at 2 weeks [15]. Although the solution is rarely used after 1 week given the lack of preservative in the dilutant, effectiveness has been noted 1 month after reconstitution with preserved saline and refrigeration [30]. The dilutant volume used can be adjusted to give two different outcomes. A smaller dilutant volume translates into a higher dose delivered in a smaller volume and a more localized effect [27,31,32]. A larger dilutant volume translates into a smaller dose per volume injected and a more diffuse effect [27]. Larger injection volumes, however, are possibly associated with more discomfort [33]. Preparation of Myo(neuro)bloc Myo(neuro)bloc is available in vials containing 2500, 5000, and 10,000 U. The aqueous solution (5000 U per milliliter) does not require constitution and is ready to use at pH 5.6. It can be diluted to desired concentration. Because of the acidic pH it may possibly produce more stinging than type A. The solution is composed of albumin, sodium succinate, and sodium chloride, and is stable in sealed unopened vials for up to 21 months if refrigerated; however, once diluted with preservative-free normal saline it should be injected within 4 hours because of lack of preservative. Preparation of Dysport Similar to Botox, this product must be reconstituted. Usually a 500-U vial, which may be stored at room temperature, is diluted with 2.5 mL of
preservative-free saline per vial to obtain a concentration of 200 U per milliliter [15,34,35].
Clinical uses Skin changes associated with the aging face are categorized as either wrinkles or lines [36]. Wrinkling consists of many multidirectional, superficial indentations of the skin, which result from thinning of the dermis with age [33,36]. Lines are single distinct depressions of the skin and are further classified into superficial creases (extending to dermis) or deeper furrows (extending to subcutaneous tissue); regardless, these generally result from underlying hyperfunctional muscles [36,37]. The term rhytide is derived from the Greek word rhytis, which means wrinkle. Facial rhytides have been categorized based on whether they are static or dynamic and whether they are caused by tissue redundancy or tissue loss (Fig. 2). Dynamic rhytides are common over the upper one third of the face, especially the glabella and forehead. Static rhytides are caused by exogenous sources, such as gravity, and environmental toxins, such as smoke and ultraviolet radiation. Static and dynamic rhytides are seen together around the eyes (crows feet); forehead; and cheeks. Treatments for facial rhytides include peels; laser resurfacing; fillers (collagen, fascia, or hyaluronic acid); fat injections; and surgical procedures, such as rhytidectomy, eyebrow lift, and blepharoplasty [36]. Surgical procedures correct gravity-induced changes and can remove excess tissue; however, these procedures may lead to significant downtime, nerve damage, and infection [36]. Fillers replace volume lost through aging or disease; they correct the static changes within the skin. Additionally, fillers improve facial contours and augment facial features, such as lips, cheekbones, and chin, but in most instances only have temporary effect. Most facial rejuvenation techniques do not target the cause of hyperfunctional lines: the muscle. The chemical paresis induced by botulism toxin is a relatively noninvasive and safe means of eliminating muscle contraction, which leads to dynamic wrinkles and furrows. With aging, ridges and wrinkles appear perpendicular to the causative muscle fibers [38]. There are multiple reports attesting to the excellent cosmetic results obtained with Botox injections [7,8,15,39 43]. In one study involving 162 patients, statistically significant reduction in hyperfunctional facial lines were found at 2 to 3 weeks after Botox injection; 95% of the patients treated had cosmetic
752
Fig. 2. Hyperfunctional lines of the face.
improvement with the best results noted with forehead lines, followed by glabellar furrows and crows feet [39]. More than 80% of these patients were extremely pleased with their results and returned for further treatment at 4 to 5 months when the effects wore off [39]. Horizontal forehead lines, glabellar forehead furrows (frown lines), and lateral canthal lines (crows feet) are the most common dynamic, hyperkinetic facial lines treated with Botox [44,45]. Complications are rare and mostly temporary (see later) [3,46,47]. A site-specific review follows. Upper one third of the face Glabellar furrows Glabellar furrows are formed by the actions of the procerus, corrugator supercilii, and medial fibers of the orbicularis oculi (Fig. 3) [48]. Numerous injection techniques have been described with early injections incorporating electromyographic guidance to locate the muscle accurately [44,49,50]. In a placebocontrolled study of 263 patients with glabellar lines of at least moderate severity [41], Botox decreased glabellar lines in 80% to 90% as assessed by both
physician and patient measures and photographic methods (Fig. 4). Injection of Botox into the corrugator supercilii muscles gives temporary chemical denervation muscle paralysis for up to 6 months followed by full recovery [7,8,37,39,40,50,51]. A double-blind placebo-controlled study with an
Fig. 3. Important facial musculature (not drawn to scale).
753
Fig. 4. Glabellar injection sites. Solid circles = Allergan FDA study injection technique.
electromyographic injection technique involving 30 patients found Botox to be effective and safe for the reduction of glabellar frown lines [50]. Another pilot study involving 57 patients using Botox injected under electromyographic guidance found that 81% (46 of 57) of patients injected with 10 U into each corrugator achieved full corrugator muscle paralysis and 67% (38 of 57) were satisfied with their improvement and desired continued injections [51]. Four patients who were pleased with their improvement but wanted complete disappearance of glabellar wrinkles were found to have minimal activity of their corrugators on electromyographic analysis; however, their lateral brow musculature seemed to contribute to their glabellar wrinkling. After 5 U Botox was injected into these accessory muscles two of the four patients had disappearance of their glabellar wrinkles. In experienced hands, electromyographic guidance is seldom used except in rare instances where clinical response was not obtained with prior injections. Superior results are more difficult to obtain in patients with thick sebaceous skin and deep dermal scarring [51], especially those with a single midline deep glabellar furrow. The glabellar spread test (spreading the glabellar wrinkles apart with the thumb and index finger) is helpful in estimating the maximal results likely from Botox [51,52]. A Botox dose-ranging study on glabellar creases by Hankins et al [33] involving 46 patients using the five standard injection site technique (one midline glabellar injection 4 mm below the brow line, and two on each side, one just above and one just below the medial brow in line with the medial canthus of the eye) determined that the minimum effective dose was 2.5 to 4 U per injection site. The founders of Botox for facial cosmetic purposes, Allistair (dermatologist) and Jean (ophthalmologist) Carruthers, currently use seven injection sites with a total of 20 to 25 U for a
female brow and 35 U for a male brow (Fig. 5) [5,46]. If Dysport is used, the doses reported in the literature range from 16 to 80 U for the glabellar region [15,34,53]. Transient ptosis is the most significant complication of injection in the glabellar area and can occur in up to 5% of patients [8,41]. The risk of ptosis is minimized by the correct injection volume and site (ie, staying above the orbital ridge), possibly by the patient staying vertical for 4 hours, and the injection site not being manipulated [52]. Lid ptosis results when Botox affects the levator muscle, which normally elevates the eyelid and may persist for 2 to 4 weeks. A placebo-controlled trial involving electromyographic-guided injections of 10 U Botox into each corrugator revealed increased incidence of ptosis when injections were given in a downward and medial direction [54]. In an excellent review by Werschler and Baumann [55], the following clinical pearls were given to minimize complications. First, placing a fingertip against the medial to superior orbital rim during injection decreases the risk of toxin migration [3,55]. Second, all Botox injections should be intramuscular with the needle bevel down to direct the toxin directly to the targeted muscle and limit diffusion elsewhere unless specified otherwise. Finally, if ptosis is severe, one to three drops of aproclonidine 0.5% eye drops (Iopidine) can be given three times a day to the affected side to give 1 to 2 mm of elevation [3,46,52]. Horizontal forehead lines The frontalis (see Fig. 3), a vertically oriented muscle, is responsible for horizontal forehead lines [48]. To soften or eliminate forehead lines and to open the eyes, injections are given at 4 to 8 sites, 2 to 3 cm above the orbital rim with several injection techniques and doses reported in the literature (Fig. 5). Guerrissi and Sarkissian [56] injected 17 patients with 14 to 20 U Botox (25 U per milliliter) and achieved satisfactory results. Despite avoiding injecting below 2.5 cm above the brow, 2 out of the 17 developed brow ptosis (lasted 55 to 70 days). Another technique used involves asking patients to raise their eyebrows and then injecting Botox 10U per milliliter into the ridges between the lines, again avoiding injecting below two finger-breaths above the eyebrow [57]. Some authors believe the elevations between the lines contain the most frontalis muscle and are the optimal injection sites [52,55,57]. Carruthers and Carruthers [5,46] use 10 to 20 U Botox for women distributed over four to five injection sites (the lateral two at the midpupillary line and then two further injections spaced equally
754
(ie, injecting too much of the frontalis with no skip areas allowing movement or treating within one fingerbreadth or 1 to 2 cm of the eyebrow) [3,48]. Also, treating the lateral frontalis past its temporal insertion (ie, lateral to the midpupillary line) can lead to drooping of the lateral eyebrow and eyelid complex, resulting in a tired appearance [46,48,55]. Additionally, patients who recruit the frontalis to elevate the eyebrows above the superior orbital rim may benefit from surgical correction of the lid and brow [48,55]. Brow lift A temporary chemical brow lift can be obtained by chemoparalysis of the brow depressors, which include the procerus; medial fibers of orbicularis oculi (depressor supercilii); and corrugator supercilii (see Fig. 3) [48]. The elevation results from an unopposed action of the frontalis muscle, which raises the brow [58 61]. Huang et al [61] reported the largest brow lifts (2 to 3 mm at central brow) after 5 U Botox (5 U/0.1 mL) was injected into the glabellar region of each brow along with an additional 10 U given in four injections along the orbital rim starting at the midpupillary line and extending to the lateral brow on each side (Fig. 7). To avoid eyelid ptosis, the author advised slowly injecting Botox with the needle aimed at an upward and horizontal direction above the orbital rim to decrease the incidence of toxin dispersion to the upper eyelid levator muscle [61]. It is worthwhile noting that injections performed below the orbital rim in the medial and midorbital region are associated with a greater risk of ptosis.
Fig. 5. Carruthers injection sites.
between) across the midbrow (an imaginary horizontal line halfway between the eyebrows and the hairline). Care must be taken to ensure that injections are given at least 2 cm above the eyebrow. Additionally, for individuals with lower brows or any degree of brow ptosis, injecting up to 2.5 U in the lateral aspect of the eyebrow (see later) along with glabellar injections lifts the lateral brow, thereby avoiding brow ptosis, improving the treatment of forehead lines [46,58]. The result is a greater opening of the orbit and a more relaxed appearance. For men with male pattern balding, injections are typically given higher up on the lateral forehead (Fig. 6) [55]. The amount of Dysport reported in the literature for this site ranges between 40 and 120 U [15,53]. Complications in this area, primarily brow ptosis, may result from overtreating the frontalis muscle
Fig. 6. Forehead injection sites. Solid circles = frontalis injection sites; solid squares = additional injection sites to produce mild brow lift.
Fig. 7. Brow lift. Solid circles = Huang injection technique (each site received 0.05 mL or 2.5 U); solid squares = Carruthers injection technique (glabellar area, 7 to 10 U; brow, 0 to 2.5 U; dilution or 1 U/0.01 mL).
755
Another technique involves one injection of 7 to 10 U Botox in the midline glabellar area immediately below the line joining the eyebrows, followed by one injection on each side into the supralateral eyebrow outside the bony orbital rim (see Fig. 7) [58]. Obtaining a brow lift in men is more difficult, because the muscle mass is larger and 35 U or more are frequently required [46]. Male brows are typically horizontal, and often they have lateral depressors. Botox brow lift injections should be given in the medial and lateral brow (2 to 3 U injected directly above the lateral canthus just above the orbital rim) [46,55]. Medial elevation of the brow occurs with injection into the midpoint of the root of the nose at the level usually equal to the medial canthi. Injections at this site weaken the medial orbicularis oculi muscle fibers (depressor supercilli), which interdigitate with the most medial aspect of the insertion of the corrugator complex [48]. These fibers have a tendency to pull the brow in a downward direction; weakening leads to slight median brow elevation and at times a widening of the glabella because of the unopposed action of the frontalis [7,42,48]. Massaging the procerus muscle after injection may help enhance this effect [7,46]. Women have two options: accentuation of the midbrow arch or a flare brow [55]. To obtain a midbrow arch, a single injection of 2 to 4 U at the midpupillary line into the muscles of the depressor of the brow, staying 1 cm above the orbital rim, results in unopposed elevation of the midbrow by the frontalis [46,55]. A flare brow (ie, greater elevation laterally) is obtained by doing the midpupillary injection above along with another injection at the lateral aspect of the brow at the lateral canthal line and at the midpoint between the medial and lateral injections [55]. For greatest effect, the lateral orbicularis muscle fibers also should be injected, as described later. Periocular crows feet Contraction of the orbicularis oculi along with the risorius and zygomaticus (mouth corner elevators) gives rise to crows feet [48]. The lateral orbicularis is injected in one to four sites with total doses ranging from 5 to 15 U Botox followed by gentle massage for a more even result (Fig. 8) [46,62]. In one study involving 15 women, 2 U of Botox injected subdermally 3 mm inferior to the lid margin at the midpupillary line improved lower eyelid wrinkles [63]. It is important to avoid injecting too deeply at this site because the toxin may migrate into the orbit and paralyze the extraocular muscles. When this 2-U injection is given in combination with 12 U Botox
Fig. 8. Injection sites for crows feet.
at the lateral orbital area, further improvement was seen [63]. Three lateral injection sites were chosen, all at 1.5-cm radius from the lateral canthus, which is 1 cm outside orbital rim. The superior injection point was 3 mm above the horizontal, the middle injection site located 1 cm below the first, and the inferior injection site just lateral to the vertical 1.5 cm below the lateral canthus. A significant increase in palpebral aperture was seen when both the lower lid and the lateral orbital area were treated; lateral rounding of the lower eye was found to be inviting or attractive according to the study patients. This lateral rounding is especially evident in Asian eyes. Areas injected are the lateral canthal or orbital orbicularis and the malar area near the inferior aspect of the orbicularis oculi muscle. One case of herniation of the orbital fat has been reported 1 week after routine periocular injection. This presented as a deformity of the inferior orbital rim that persisted as chronic edema for 5 months after the procedure. To avoid this complication the authors recommend avoiding the inferior aspect of the orbicularis of the lower eyelids [64]. In addition, to reduce purpura, the injection sites must be chosen carefully (possibly under magnification), making every effort to avoid the numerous vessels around the lateral eye region, and pressure should be applied immediately after injection. Injections over the proximal aspect of the zygomatigus major muscle may soften lateral canthal rhytides associated with smiling; however, an asymmetric smile may result if injecting below the inferior margin of the zygoma [45]. Younger patients with minimal skin laxity do better than older patients with redundant skin [62]. The amount of Dysport reported in the literature for this region is 20 to 30 U per side [15].
756
It is important to individualize treatment and see where the crows feet are located when patients squint and smile. For many patients they are both above and below the lateral canthal line and all dynamic sites should be treated for the best results [55]. Injections can be done every 1 cm where there is muscle movement up to the midpupillary point [55]. Additionally, the lower eyelid can be treated by asking the patient to hold their head back and look up, retracting the lid inferiorly, injecting 2 to 3 U subcutaneously at the midpupillary line and then distributing the Botox by rolling a cotton swab in the area (beware of bruising). For a more rounded appearance one can inject 2 U at the midpupillary line and another 2 U halfway between the midpupillary line and the lateral canthus; again, especially useful in Asians. It is important to remember that injections below the orbital rim are associated with an increased risk of diplopia because of diffusion of Botox into the extraocular muscles [48,55]. Injections within 1 cm of the lateral canthus may also allow toxin migration to the lateral rectus, which may result in diplopia. Also, Werschler and Baumann [55] note that lateral eyebrow elevation usually occurs to some extent when the lateral orbicularis oculi muscles are weakened because of the unopposed elevation of the lateral brow by the lateral frontalis. If too much elevation occurs, a few injections into the frontalis reposition the brow in a downward position.
Mouth frowns The downward turn of the lateral corners of the mouth (marionette lines, drool groove, or melomental folds) may be improved by chemodenervation of depressor anguli oris, thereby allowing elevation of the mouth corners by the unopposed zygomaticus major and minor; additionally, platysmal injections may also be helpful because the lateral bands of platysma help pull the mouth down [46,66,67]. Depressor anguli oris muscle injections may produce a good lift of the corners of the mouth but may produce transient smiling and elocution problems; it is not recommended for singers, musicians, or patients who use their perioral muscles with intensity [65]. The depressor anguli oris can be located inferior to a point 1 cm lateral to the commissure while pulling down on the corners of the mouth [46]. Three to 5 U of Botox are injected 10 mm lateral to the angle of the lips. This technique may be useful in conjunction with soft tissue filling agents in this area because it prevents molding and contortion of the filler [65,67]. Mentalis A prominent mental fold and a pebbly or cobblestone peau dorange chin may be improved by injection of 5 to 10 U Botox into the point of the chin; this may also flatten the mental fold [44,46,67]. Injection here can also augment the use of fillers. If the injection is done into the mental fold area, however, an incompetent mouth may result [46]. Gentle massage after injection may be helpful [3]. Nasolabial folds Injection to weaken the zygomaticus major has not given dramatic results and may not be uniformly useful [46]. The upper lip may lengthen or droop, which simulates aging [46]. If a short lip is present, Botox may be useful because it may flatten the nasolabial fold, lengthen the lip, have a prolonged effect of up to 6 months, and augment the use of fillers. Injection of 2 to 3 U into the levator labii superioris alaeque nasi, on either side of the nose, may smooth out the superomedial nasolabial fold [3,46,67]. Neck The first report of platysmal rejuvenation using Botox was by Blitzer et al [68]. Both hypertrophic platysmal bands and horizontal neck lines have improved after Botox [44,46,67,69]. The central platysmal bands thicken and contract with time and contribute to the formation of jowls and loss of definition of the neck [66,70]. The platysma originates in the superior fascia of the upper chest (the pectoralis and deltoid fascia) and extends over the full length of
Middle and lower face Lips Perioral rhytides, accentuated by pursing the lips, may be softened when small amounts of Botox are injected symmetrically. Vertical upper lip wrinkles (smokers lines) are caused by contraction of the orbicularis musculature. One to 2 U under each wrinkle at the vermilion border may soften the lines, but no more than 1 to 2 U should be injected per lip quadrant (ie, on either side of Cupids bow) and only two to three wrinkles should be treated at a time [46,65]. Alternatively, this may be performed using injections 5 mm above the vermilion, 1 to 2 cm apart, or by injecting from the buccal surface of the upper lip, outward, using lidocaine gel anesthesia (Nick Lowe, personal communication, 2001). Microparesis may be a concern in this area interfering with whistling, suction, and elocution. Lip orbicularis muscle injections must be done using small volumes with a high dilution, and produce only partial and short-term results.
757
the neck up past the mandible. Most fibers (two of three) interlace in the midline [70]. Some of the fibers insert into the mandible and a few insert at the lateral oral commissure, and result in a downward pulling on the corners of the mouth (see Fig. 3) [70]. A variety of different injection techniques have been reported in the literature [3,46,66]. Patients are requested to contract the platysma, the hypertrophic band is grasped, and injections given at regular intervals from 1 to 3 cm from the jaw line to the lower neck for a total dose of 15 to 21 U per band [3,44,46,66]. Usually two to four bands are treated in one session using 50 to 100 U per session [3]; however, up to 200 U has been required in one session [66]. Immediate relaxation of the platysma is seen with onset of weakness occurring in 3 to 5 days. Repeat injections every 4 to 6 months are required with prolonged effects seen with each subsequent injection [66]. Horizontal neck lines are injected in the deep intradermal plane with a total of 10 to 30 U given at intervals of 2 to 3 cm [46,67]. Complications seen in this region include transient edema, ecchymoses, muscle soreness, neck flexor weakness, and headaches [66]. Rarely, laryngeal muscle weakness, hoarseness, and dysphagia may occur 3 to 4 days after injection, especially with doses around 75 to 100 U [3,46]. Meloclopramide hydrochloride (Reglan) may improve swallowing because it stimulates the upper gastrointestinal tract [3]. Hyperhidrosis Hyperhidrosis most commonly involves the axilla, palms, or soles. For all these sites Botox has effectively decreased sweating [71 76]. Hyperhidrosis may also involve the face and neck. Gustatory sweating (Freys syndrome and auriculotemporal syndrome) may occur after parotid gland surgery or after the preauricular region has been traumatized. Denervated sweat glands may become inappropriately reinnervated by parasympathetic nerve fibers that previously had the salivary gland as their target. As a result, sweating of the facial skin occurs during meals. Intracutaneous injections of Botox, 2.5 U per square centimeter in doses up to 21 to 38 U, depending on affected area, using 7 to 25 sites, lead to satisfactory responses lasting an average of 17.3 months [77 81]. Complications reported include weakness of the mimic muscles; no masticatory muscle weakness has been reported.
geons have come to appreciate the effect of Botox when used before the surgical resection of the corrugator musculature for the permanent correction of glabellar creases [82,83]. Botox and brow lift Botox-assisted brow lift has stable, long-lasting, aesthetic results with minimal morbidity [59,69]. Brow lifts may have an unpredictable cosmetic outcome based on postsurgical healing. Stabilization of brow musculature is important for a predictable final brow position. Studies have suggested that after surgical elevation, periosteal refixation requires 12 weeks to become adherent. The hypokinesis provided by Botox injections provides more predictable healing during this interval [59]. Some surgeons believe Botox-assisted brow lifts may help improve the cosmetic outcome. Botox and laser resurfacing In general, lasers address static wrinkles and stimulate new collagen, whereas Botox prevents the recurrence of dynamic wrinkles [52,65,84]. Botox injection 2 to 3 weeks before laser resurfacing may produce superior results [65]. In one study involving 53 patients, the group that received Botox before and after laser resurfacing (N = 37) had an aesthetic outcome that was rated 21% greater after 6 months than those who underwent laser resurfacing alone [85]. By using Botox along with skin resurfacing, the recurrent muscle activity that can disturb the lamellar organization of the newly rejuvenated skin is prevented, and the cosmetic improvement achieved by the laser resurfacing is better maintained giving an improved and longer-lasting outcome [65]. Botox prevents muscle contractions from shaping new collagen into wrinkles. Botox and peels Botox should be done 2 to 3 days before chemical peels and not immediately before or after peels because soft tissue swelling may result in increased migration of the toxin [55]. Botox and fillers
Botox as adjunct to surgical procedures When used in conjunction with surgical procedures, Botox is frequently helpful [65]. Plastic sur-
Botox may decrease the amount of fillers required for lines. Giving Botox 2 to 4 weeks before fillers softens the muscle contractions and gives a more accurate idea of how much filler is required [55].
758
Injecting Botox and tissue fillers on the same day may possibly lead to overcorrection.
Clinical response There is no clinical standardization on the use of botulinum toxin type A. Known factors affecting clinical response are commercial preparations, reconstitution, anatomy, dose and response relationship, storage, immunogenicity, and operator skill [2]. When feasible, injections should be performed intramuscularly. Intradermal injections are associated with less efficacy because Botox can only reach the muscle fiber by dispersion. Injection technique varies greatly in the literature and is site specific as previously described (see individual sites). Some authors advocate tangential injections instead of perpendicular injections [28,52,84]. The toxin diffuses an average of 1 cm in all directions from injection sites [46]. The greater the dilutant, the greater is the diffusion. Site manipulation (rubbing or massaging) also increases diffusion. For botulinum toxin type B, the extent of diffusion may be increased because its molecular weight is less than the type A toxins [11]. Many authors advocate overuse of muscles for 2 to 4 hours after injection to encourage toxin uptake and obtain optimal responses [84]. Effects usually appear in 1 to 3 days [39] and at-rest appearance effects last longer with over 85% of patients responding in 7 days of treatment and peak response occurring 30 days after injection. Forty percent of patients are likely to report moderate or better improvement 4 months after treatment [41].
minimizes bruising and deep muscle hematomas, which may be painful for days after the procedure [55,84]. Headaches may be the result of hitting the periosteum or deep muscle hematomas [55]. Interestingly, relief of tension headaches and migraine headaches is more commonly reported [46,87]. Pain may be more often severe for women when they are menstruating [55]. Unbuffered lidocaine has been added to reconstituted Botox; however, the injections are more painful with this acidic mixture and it is not recommended [55]. Clinicians have reported less injection site pain with Botox that had been reconstituted using benzyl-alcohol-preserved saline [88].
Contraindications There are no reported cases of teratogenicity from Botox injection. One study involving nine patients who were pregnant during injection (dose unspecified) had one premature delivery, which was not thought to be related to the toxin [16]. Despite the lack of teratogenicity, most of the scientific community refrains from injecting pregnant or lactating women. A history of a neuromuscular disease or a known history of sensitivity to Botox or human albumin should exclude patients from Botox injection. Aminoglycosides can interfere with neuromuscular transmission and potentiate the effects of Botox injections [89]. Areas with active infection should not be injected.
Summary The use of botulinum toxin has revolutionized the treatment of facial lines with an incomparable safety record over the past 14 years. The most common used injection sites are shown in Fig. 9. With the recent FDA approval for Botox in the treatment of glabellar lines, its use will likely increase dramatically. It is essential that practitioners have a detailed and specific knowledge of the facial and neck musculature to be injected to minimize untoward side effects, especially in the early days of new users learning curve. The specifics of the dilutions and units per amount used for the various different commercial forms of botulinum toxin types A and B need to be understood fully and standardized together with the potential for antigenicity with the higher protein load of type B. In addition, specific indications for the use of botulinum toxin as adjunctive therapy for specific facial surgical procedures (ie, blepharoplasty, surgical brow lift, and laser resurfacing) will become better understood.
Adverse effects Botox has been associated with side effects that are minor and transient [3,7]. No reports of allergic or urticarial reactions have been reported with facial aesthetic procedures. A psoriasiform eruption, which resolved spontaneously after 5 months, has been reported in one patient who received 15 U Botox for a lateral rectus palsy [86]. Transient bruising at injection site (especially common in patients taking aspirin, nonsteroidal anti-inflammatory drugs, warfarin, vitamin E, and in crows feet injections), transient headaches (13.3%), and pain (2.2%), have been reported [41]. Bruising can be minimized by using a 30-gauge needle and changing it every three to seven injections, preinjection application of ice or topical anesthetics, and avoidance of blood thinners [28,46,55]. Application of pressure after injection
759
Fig. 9. Injection summary.
Finally, even though the anatomy of the facial musculature is well described, individual differences in men and women, in different population groups, and in tissue qualities, such as turgor and elasticity [87], are important factors to be considered before undertaking botulinum toxin injections. It is likely that the use of specific measuring devices, such as digital imaging, will further help define the use of botulinum toxin for different muscle groups and facial aesthetic indications.
References
[1] Scott AB, Rosenbaum A, Collins CC. Pharmacologic weakening of extraocular muscles. Invest Ophthalmol 1973;12:924 7. [2] Huang W, Foster JA, Rogachefsky AS. Pharmacology of botulinum toxin. J Am Acad Dermatol 2000;43: 249 59. [3] Klein AW. Complications and adverse reactions with the use of botulinum toxin. Semin Cutan Med Surg 2001;20:109 20. [4] Shimizu T, Sakaguchi G. Production and properties of type F toxin. In: Jankovic J, Hallet M, editors. Therapy with botulinum toxin. New York: Marcel Dekker; 1994. p. 87 92. [5] Carruthers A, Carruthers J. Botulinum toxin type A: history and current cosmetic use in the upper face. Semin Cutan Med Surg 2001;20:71 84.
[6] Scott AB. Clostridial toxins as therapeutic agents. In: Simpson LL, editor. Botulinum neurotoxin and tetanus toxin. New York: Academic Press; 1989. p. 399 412. [7] Carruthers A, Kiene K, Carruthers J. Botulinum A exotoxin use in clinical dermatology. J Am Acad Dermatol 1996;34:788 97. [8] Carruthers A, Carruthers J. Botulinum toxin in the treatment of glabellar frown lines and other facial wrinkles. In: Jankovic J, Hallet M, editors. Neurological disease and therapy: therapy with botulinum toxin. New York: Marcel Dekker; 1994. p. 577 95. [9] Carruthers JDA, Carruthers A. Botulinum A exotoxin in clinical ophthalmology. Can J Ophthalmol 1996; 31:389 400. [10] Simpson LL. The origin, structure and pharmacological activity of botulinum toxin. Pharmacol Rev 1981;33:155 88. [11] Carruthers A, Carruthers J. Toxins 99. new information about the botulinum neurotoxins. Dermatol Surg 2000; 26:174. [12] Binz T, Blasi J, Yamasaki S, et al. Proteolysis of SNAP-25 by types E and A botulinum neurotoxins. J Biol Chem 1994;269:1617 20. [13] Callaway JE, Arezzo JC, Grethlein AJ. Botulinum toxin type B: an overview of its biochemistry and preclinical pharmacology. Semin Cutan Med Surg 2001; 20:127 36. [14] Schiavo G, Benfenati F, Poulain B, et al. Tetanus and botulinum-B neurotoxins block transmitter release by proteolytic cleavage of synaptobrevin. Nature 1992; 359:832. [15] Lowe NJ. Botulinum toxin type A for facial rejuvenation. Dermatol Surg 1998;24:1216 8. [16] Schantz EJ, Johnson EA. Dose standardization of botulinum toxin. Lancet 1990;335:421. [17] Brashear A, Lew MF, Dystra DD, et al. Safety and efficacy of NeuroBloc (botulinum type B) in type A-responsive cervical dystonia. Neurology 1999;53: 1439 46. [18] Brin MF, Lew MF, Adler CH, et al. Safety and efficacy of NeuroBloc (botulinum toxin type B) in type A-resistant cervical dystonia. Neurology 1999; 53:1431 8. [19] Lamanna C, Hillowalla RA, Alling CC. Buccal exposure to botulinal toxin. J Infect Dis 1967;117:327. [20] Scott AB, Suzuki D. Systemic toxicity of botulinum toxin by intramuscular injection in the monkey. Mov Disord 1988;3:333 5. [21] Goschel H, Wohlfarth K, Frevert J, Dengler R, Bigalke H. Botulinum A toxin therapy: neutralizing and nonneutralizing antibodies therapeutic consequences. Exp Neurol 1997;147:96 102. [22] Matarasso SL. Update on the aesthetic uses of botulinum-A neurotoxin in facial rejuvination. Curr Probl Dermatol 2001;13:46 53. [23] Zuber M, Sebald M, Bathien N, et al. Botulinum antibodies in dystonic patients treated with type A botulinum toxin: frequency and significance. Neurology 1993;43:1715 8.
760
J.C. Cather et al. / Dermatol Clin 20 (2002) 749761 [42] Foster JA, Barnhorst D, Papay F, Oh PM, Wulc AE. The use of botulinum A toxin to ameliorate facial kinetic frown lines. Ophthalmology 1996;103:618 22. [43] Guyuron B, Huddleston SW. Aesthetic indications for botulinum toxin injection. Plast Reconstr Surg 1994; 93:913 8. [44] Binder WJ, Blitzer A, Brin MF. Treatment of hyperfunctional lines of the face with botulinum toxin A. Dermatol Surg 1998;24:1198 205. [45] Fagien S. Botox for the treatment of dynamic and hyperkinetic facial lines and furrows: adjunctive use in facial aesthetic surgery. Plast Reconstr Surg 1999; 103:701 13. [46] Carruthers A, Carruthers J. Clinical indications and injection technique for the cosmetic use of botulinum A exotoxin. Dermatol Surg 1998;24:1189 94. [47] Matarasso SL. Complications of botulinum A exotoxin for hyperfunctional lines. Dermatol Surg 1998;24: 1249 54. [48] Wieder JM, Moy RL. Understanding botulinum toxin. Dermatol Surg 1998;24:1172 4. [49] Klein AW, Mantell A. Electromyographic guidance in injecting botulinum toxin. Dermatol Surg 1998; 24:1184 6. [50] Lowe NJ, Maxwell A, Harper H. Botulinum A exotoxin for glabellar folds: a double-blind, placebo-controlled study with an electromyographic injection technique. J Am Acad Dermatol 1996;35:569 72. [51] Pribitkin EA, Greco TM, Goode RL, Keane WM. Patient selection in the treatment of glabellar wrinkles with botulinum toxin type A injection. Arch Otolaryngol Head Neck Surg 1997;123:321 6. [52] Edelstein C, Shorr N, Jacobs J, Balch K, Goldberg R. Oculoplastic experience with the cosmetic use of botulinum A exotoxin. Dermatol Surg 1998;24:1208 12. [53] Tambasco N, Lalli F, Manicini ML, Rossi A. Botulinum toxin in the treatment of facial expression wrinkles (ITA). Riv Ital Chir Plast 1997;29:197 200. [54] Lowe NJ, Wieder JM. Botulinum toxin for hyperkinetic facial lines: a placebo-controlled study. Cosmetic Dermatology 1995;8:46 7. [55] Werschler P, Baumann L. Everything you need to know about Botox injections. Skin & Aging 2001;9:36 42. [56] Guerrissi J, Sarkissian P. Local injection into mimetic muscles of botulinum toxin A for the treatment of facial lines. Ann Plast Surg 1997;39:447 53. [57] Goodman G. Botulinum toxin for the correction of hyperkinetic facial lines. Australas J Dermatol 1998; 39:158 63. [58] Huilgol SC, Carruthers A, Carruthers JDA. Raising eyebrows with botulinum toxin. Dermatol Surg 2000; 25:373 6. [59] Dyer WK, Yung RT. Botulinum toxin-assisted brow lift. Facial Plast Surg 2000;8:343 54. [60] Fagien S. Temporal brow lift using botulinum toxin A: discussion. Plast Reconstr Surg 2000;105:1136 8. [61] Huang W, Rogachefsky AS, Foster JA. Browlift with botulinum toxin. Dermatol Surg 2000;26:55 60. [62] Keen M, Kopelman JE, Aviv JE, Binder W, Brin M,
[24] Aoki R, Merlino G, Spanoyannis A, et al. Botox (botulinum toxin type A) purified neurotoxin complex prepared from the new bulk toxin retains the same preclinical efficacy as the original but with reduced immunogenicity. Neurology 1999;52:A521 5. [25] Bowmer EJ. Preparation and assay of the international standards for Clostridium botulinum types A,B,C,D, and E antitoxins. Bull World Health Organ 1963; 29:219 60. [26] Arron SS, Schechter R, Inglesby TV, et al. Botulinum toxin as a biologic weapon: mental and public health management. JAMA 2001;285:1059 70. [27] Klein AW. Dilution and storage of botulinum toxin. Dermatol Surg 1998;24:1179 80. [28] Fulton JE. Utilizing botulinum toxin in your cosmetic surgery practice. Cosmetic Dermatology 1997; 10:41 6. [29] Alcon Pharmaceuticals. Botox package insert. Irvine, CA: Alcon Pharmaceuticals; 1995. [30] Garcia A, Fulton JE. Cosmetic denervation of the muscles of facial expression with botulinum toxin. Dermatol Surg 1996;22:39 43. [31] Borodic GE, Ferrante R, Pearce B, et al. Histologic assessment of dose-related diffusion and muscle fiber response after therapeutic botulinum toxin A injection. Mov Disord 1994;9:31 9. [32] Shaari CM, Sanders I. Quantifying how location and dose of botulinum toxin injection affect muscle paralysis. Muscle Nerve 1993;16:964 9. [33] Hankins CL, Strimling R, Rogers GS. Botulinum A toxin for glabellar wrinkles. Dermatol Surg 1998; 24:1181 3. [34] Ascher B, Klap P, Marion M-H, et al. Le toxine botulique dans le treatment de rides fronto-glabellaires et de la region orbitaire. Ann Chir Plast Esthet 1995; 40:67 76. [35] Le Louarn C. A new injection procedure. Toxine botulique et rides faciales: Une nouvelle procedure dinjection (FRE). Ann Chir Plast Esthet 1998;43:526 33. [36] Carter SR, Seiff SR. Cosmetic botulinum toxin injections. Int Ophthalmol Clin 1997;37:69 79. [37] Keen M, Blitzer A, Aviv J, Binder W, Prystowsky J, Smith H, et al. Botulinum toxin A for hyperkinetic facial lines: results of a double-blind, placebo controlled study. Plast Reconstr Surg 1994;94:94 9. [38] Moore KL. Clinically oriented anatomy. Baltimore: Williams and Wilkins; 1992. [39] Blitzer A, Binder WJ, Aviv JE, Keen MS, Brin MF. The management of hyperfunctional facial lines with botulinum toxin. Arch Otolaryngol Head Neck Surg 1997;123:389 92. [40] Carruthers J, Carruthers A. Treatment of glabellar frown lines with C. botulinum-A exotoxin. J Dermatol Surg Oncol 1992;18:17 21. [41] Carruthers JA, Lowe NJ, Menter MA, Gibson J, Nordquist M, Mordaunt J, et al. A multicenter, double blind, placebo-controlled study of the efficacy and safety of botulinum toxin type A in the treatment of glabellar lines. J Am Acad Dermatol 2002;46:840 9.
J.C. Cather et al. / Dermatol Clin 20 (2002) 749761 Glasgow LA. Botulinum toxin A: a novel method to remove periorbital wrinkles. Facial Plast Surg 1994; 10:141 6. Flynn TC, Carruthers JA, Carruthers JA. Botulinum-A toxin treatment of the lower eyelid improves infraorbital rhytides and widens the eye. Dermatol Surg 2001; 27:703 8. Paloma V, Samper A. A complication with the aesthetic use of Botox: herniation of the orbital fat. Plast Reconstr Surg 2001;107:1315. Carruthers J, Carruthers A. The adjunctive usage of botulinum toxin. Dermatol Surg 1998;24:1244 7. Brandt FS, Bellman B. Cosmetic use of botulinum A exotoxin for the aging neck. Dermatol Surg 1998;24: 1232 4. Carruthers J, Carruthers A. Botox use in the mid and lower face and neck. Semin Cutan Med Surg 2001; 20:85 92. Blitzer A, Brin M, Keen MS, Aviv JE. Botulinum toxin for the treatment of hyperfunctional lines of the face. Arch Otolaryngol Head Neck Surg 1993;119: 1018 22. Ascher B. Botulinum toxin: anatomic basis, classical and new aesthetic indications. Mov Disord 2000;15 (suppl 2):27 8. Hoefflin SM. Anatomy of the platysma and lip depressor muscles. Dermatol Surg 1998;24:1225 31. Glogau RG. Botulinum A neurotoxin for axillary hyperhidrosis. Dermatol Surg 1998;24:817 9. Naumann M, Hofmann U, Bergmann I, Hamm H, Toyka KV, Reiners K. Focal hyperhidrosis. Arch Dermatol 1998;134:301 4. Odderson IR. Axillary hyperhidrosis: treatment with botulinum toxin A. Arch Phys Med Rehabil 1998;79: 350 2. Odderson IR. Hyperhidrosis treated by botulinum A exotoxin. Dermatol Surg 1998;1237 41. Schnider P, Moraru E, Kittler H, Voller B, Auff E. High-dose botulinum toxin type A for axillary hyperhidrosis. Arch Dermatol 2000;136:1567.
761
[63]
[64]
[65] [66]
[67]
[68]
[69]
[70] [71] [72]
[73]
[74] [75]
[76] Shelley WB, Talanin NY, Shelley ED. Botulinum toxin therapy for palmar hyperhidrosis. J Am Acad Dermatol 1998;38:227 9. [77] Bjerkhoel A, Trobbe O. Freys syndrome treatment with botulinum toxin. J Laryngol Otol 1997;3:839 44. [78] Drobik C, Laskawi R. Freys syndrome: treatment with botulinum toxin. Acta Otolaryngol (Stockh) 1995; 115:459 61. [79] Laskawi R, Drobik C, Schonebeck C. Up to date report of botulinum toxin type A treatment in patients with gustatory sweating (Freys sweating). Laryngoscope 1998;108:381 4. [80] Nauman M, Zellner M, Toyka KV, Reiners K. Treatment of gustatory sweating with botulinum toxin. Ann Neurol 1997;42:973 5. [81] Schulze-Bonhage A, Schroder M, Ferbert A. Botulinum toxin in the therapy of gustatory sweating. J Neurol 1996;213:143 6. [82] Hamas RS. Reducing the subconscious frown by endoscopic resection of the corrugator muscles. Aesthetic Plast Surg 1995;19:21 5. [83] Hamas RS. An endoscopic brow lift that does not raise the hairline. Aesthetic Surgery Journal 1997;17:127 9. [84] Fulton JE. Botulinum toxin. Dermatol Surg 1998;24: 1219 24. [85] Worcester S. Use Botox before and after laser facial resurfacing. Skin Allergy News 2000;31:6. [86] Bowden JB, Rapini RP. Psoriasiform eruption from intramuscular botulinum A toxin. Cutis 1992;50:415 6. [87] Heckmann M, Schon-Hupka G. Quantification of the efficacy of botulinum toxin type A by digital image analysis. J Am Acad Dermatol 2001;45:508 14. [88] Alam M, Dover JS, Arndt KA. Pain associated with injection of botulinum toxin A exotoxin reconstituted using isotonic sodium chloride with and without preservative. Arch Dermatol 2002;138:510 4. [89] Santos JI, Swenson P, Glasgow LA. Potentiation of clostridium botulinum toxin aminoglycoside antibiotics: clinical and laboratory observations. Pediatrics 1981;68:50 4.

2002, Vol.20, Issues 4, Melanoma

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

2002, Vol.20, Issues 4, Melanoma

Hochgeladen von

Copyright:

Verfügbare Formate

Dermatol Clin 20 (2002) xiii xiv

A special thanks to Alfred W. Kopf, MD

Dedication / Dermatol Clin 20 (2002) xiiixiv

Dermatol Clin 20 (2002) xv

Melanoma and pigmented lesions

Darrell S. Rigel, MD Guest Editor

Dermatol Clin 20 (2002) 589 595

Melanoma incidence trends

C. Bevona, A.J. Sober / Dermatol Clin 20 (2002) 589595

C. Bevona, A.J. Sober / Dermatol Clin 20 (2002) 589595

C. Bevona, A.J. Sober / Dermatol Clin 20 (2002) 589595

C. Bevona, A.J. Sober / Dermatol Clin 20 (2002) 589595

[8] [9] [10]

[36] [37] [38]

[17] [18] [19]

[46] [47] [48]

Dermatol Clin 20 (2002) 597 600

Risk factors for the development of primary cutaneous malignant melanoma

R.M. MacKie / Dermatol Clin 20 (2002) 597600

R.M. MacKie / Dermatol Clin 20 (2002) 597600

Dermatol Clin 20 (2002) 601 606

The effect of sunscreen on melanoma risk

D.S. Rigel / Dermatol Clin 20 (2002) 601606

Salicylates and anthranilates

D.S. Rigel / Dermatol Clin 20 (2002) 601606

D.S. Rigel / Dermatol Clin 20 (2002) 601606

D.S. Rigel / Dermatol Clin 20 (2002) 601606

Dermatol Clin 20 (2002) 607 616

Congenital melanocytic nevi Evaluation and management

E-mail address: marghooa@mskcc.org (A.A. Marghoob).

A.A. Marghoob / Dermatol Clin 20 (2002) 607616

A.A. Marghoob / Dermatol Clin 20 (2002) 607616

10. What complications other than melanoma can develop?

11. What laboratory tests or consultations should be considered?

A.A. Marghoob / Dermatol Clin 20 (2002) 607616

A.A. Marghoob / Dermatol Clin 20 (2002) 607616

A.A. Marghoob / Dermatol Clin 20 (2002) 607616

A.A. Marghoob / Dermatol Clin 20 (2002) 607616

A.A. Marghoob / Dermatol Clin 20 (2002) 607616

Dermatol Clin 20 (2002) 617 628

The dilemma of the dysplastic nevus

E-mail address: tsalopek@ualberta.ca (T.G. Salopek).

T.G. Salopek / Dermatol Clin 20 (2002) 617628

Chromosome 1p36 9p21 12q14 9p21

Gene product Unknown CDKN2A (p16) CDK4 Possibly p19

T.G. Salopek / Dermatol Clin 20 (2002) 617628

T.G. Salopek / Dermatol Clin 20 (2002) 617628

T.G. Salopek / Dermatol Clin 20 (2002) 617628

T.G. Salopek / Dermatol Clin 20 (2002) 617628

T.G. Salopek / Dermatol Clin 20 (2002) 617628

T.G. Salopek / Dermatol Clin 20 (2002) 617628

Dermatol Clin 20 (2002) 629 640

Screening for melanoma

A.C. Geller / Dermatol Clin 20 (2002) 629640

A.C. Geller / Dermatol Clin 20 (2002) 629640

A.C. Geller / Dermatol Clin 20 (2002) 629640

A.C. Geller / Dermatol Clin 20 (2002) 629640

A.C. Geller / Dermatol Clin 20 (2002) 629640

A.C. Geller / Dermatol Clin 20 (2002) 629640

A.C. Geller / Dermatol Clin 20 (2002) 629640

A.C. Geller / Dermatol Clin 20 (2002) 629640

Dermatol Clin 20 (2002) 641 646

* Corresponding author. E-mail address: grin@nso1.uchc.edu (C.M. Grin).

C.M. Grin et al. / Dermatol Clin 20 (2002) 641646