Treatment of Orbital Rhabdomyosarcoma: Survival and Late Effects of TreatmentResults of an International Workshop

By Odile Oberlin, Annie Rey, James Anderson, Modesto Carli, R. Beverley Raney, Joern Treuner, and Michael C.G. Stevens for the International Society of Paediatric Oncology Sarcoma Committee, the Intergroup Rhabdomyosarcoma Study Group, the Italian Cooperative Soft Tissue Sarcoma Group, and the German Collaborative Soft Tissue Sarcoma Group
Purpose: Orbital rhabdomyosarcoma (RMS) historically has been associated with an excellent survival rate. The majority of patients are cured with the use of both chemotherapy and radiation therapy, but a signicant number experience important late sequelae of treatment. In an attempt to determine optimal therapy in relation both to cure and to sequelae, the experience of the four international collaborative groups (Intergroup Rhabdomyosarcoma Study Group [IRSG], International Society of Paediatric Oncology [SIOP] Sarcoma Committee, German Collaborative Soft Tissue Sarcoma Group [CWS], and Italian Cooperative Soft Tissue Sarcoma Group [ICG] studies) was shared at an international workshop. Patients and Methods: A total of 306 eligible patients were identied from group records (186 from IRS, 43 from SIOP MMT, 40 from CWS, and 37 from ICG). Median age was 6.8 years, and median follow-up was 6.5 years. Eighty percent of patients received radiation therapy (RT) as part of primary therapy, but there were signicant differences in the use of RT between the individual groups (93% in IRSG,
76% in ICG, and 70% in CWS, but only 37% in the SIOP MMT group). Results: At 10 years, event-free and overall survival for the whole cohort were 77% (range, 71% to 81%) and 87% (range, 82% to 92%), respectively. There was no difference in overall survival between the collaborative groups regardless of differences in the use of initial RT. In total, 34 (12%) of 273 survivors had not received RT, although this varied between the different groups (41% in the SIOP MMT group, 20% in CWS, 7% in ICG, and 6% in IRSG). There was no difference in overall survival for the whole cohort regardless of whether radiotherapy was used as part of initial therapy (86% at 10 years for both). Conclusion: These data suggest that a subset of patients with orbital RMS can be cured without systematic local therapy, although the total burden of treatment (primary therapy and treatment for relapse) must be taken into account when assessing the implications for late sequelae. J Clin Oncol 19:197-204. 2001 by American Society of Clinical Oncology.

HABDOMYOSARCOMA (RMS), the most common of the pediatric soft tissue sarcomas, arises in a multiplicity of sites. Forty percent, however, originate from the head and neck region, and 10% develop within the orbit. In the majority of published series, the orbit is the most favorable site, with a 5-year survival rate greater than 85%.1-3 The approach to treatment has undergone radical change over the last 20 years and has evolved from primary exenteration to a more conservative multidisciplinary approach combining systemic chemotherapy and local radiation therapy. The excellent survival of this group of children has allowed clinical investigators to follow survivors for many years and, unfortunately, to observe the development of important late effects resulting from local treatment.4 Problems include both functional and structural changes: cataract and changes in the cornea and the retina are amongst the most common problems, but bony hypoplasia of the orbit and facial asymmetry are also frequently described. Knowledge of such late effects has led to attempts to avoid systematic local therapy (almost exclusively given as

radiotherapy) in those patients in whom complete remission is achieved with primary chemotherapy. This has largely been explored by the studies promoted by the International Society of Pediatric Oncology (SIOP) group.5 The challenge must always be to maintain excellent survival while demonstrating a reduction in late effects of therapy. Studies from other collaborative groups, including the North American Intergroup Rhabdomyosarcoma Study Group (IRSG),

From the Institut Gustave-Roussy, Villejuif, France; University of Nebraska Medical Center, Omaha, Nebraska; University Hospital of Padova, Padova, Italy; M.D. Anderson Cancer Center, Houston, TX; Olga Hospital, Stuttgart, Germany; and The Childrens Hospital, Birmingham, United Kingdom. Submitted February 17, 2000; accepted August 2, 2000. Supported in part by Association de Recherche Contre le Cancer, Villejuif, France (grant no. 1381). Address reprint requests to O. Oberlin, MD, Department of Paediatric Oncology, Institut Gustave-Roussy, 94805 Villejuif Cedex, France; email oberlin@igr.fr. 2001 by American Society of Clinical Oncology. 0732-183X/01/1901-197

Journal of Clinical Oncology, Vol 19, No 1 (January 1), 2001: pp 197-204

Downloaded from jco.ascopubs.org on January 30, 2014. For personal use only. No other uses without permission. Copyright 2001 American Society of Clinical Oncology. All rights reserved.

197

198
the German Collaborative Soft Tissue Sarcoma Group (CWS), and the Italian Cooperative Soft Tissue Sarcoma Group (ICG), have been based on protocols that have generally included chemotherapy and orbital radiation therapy for all patients. Because of such signicant differences in treatment philosophy, and in an attempt to determine the optimal balance between a high cure rate and a low risk of signicant late effects from therapy, a workshop was held in 1997 to assess the treatment of orbital RMS and its outcome between the four international groups contributing data. Representatives from the IRSG, CWS, ICG, and SIOP all participated. Data from the different treatment protocols were pooled to undertake an analysis of prognostic factors in a large cohort of children with nonmetastatic orbital RMS. The primary purpose of the workshop was to explore the impact of radiotherapy on ultimate survival when given as part of primary treatment and to assess, where possible, the extent to which local treatments contributed to signicant late effects of therapy.
PATIENTS AND METHODS

OBERLIN ET AL before initiation of chemotherapy. Dactinomycin and doxorubicin were omitted during radiation. The total duration of chemotherapy was 2 years. In the IRS III protocol, all patients (n 102) received VA for 1 year with radiation therapy beginning at week 2. Patients with completely resected tumor were not given radiotherapy, and patients whose tumors were not removed completely were given radiation at doses dened as in IRS II. In the IRS IV study, patients (n 20) were randomized to receive either VAC; vincristine, dactinomycin, and ifosfamide (VAI); or vincristine, ifosfamide, and etoposide (VIE) for 1 year. Radiation therapy was given from week 9. The radiotherapy schedule was randomized between 50.4 Gy in a conventional daily schedule or 59.4 Gy in a hyperfractionated schedule. A single patient with a completely removed tumor received only VA chemotherapy without radiation therapy. CWS Studies. In the CWS 81 study (14 patients), chemotherapy was given as vincristine, dactinomycin, cyclophosphamide, and doxorubicin (VACA) for a 35-week period. Radiation dose was stratied depending on the outcome of second-look surgery: patients who did not undergo second-look surgery or who had macroscopic residue were given 50 Gy, and patients with microscopic residue received 40 Gy, all by conventional fractionation. In the CWS 86 study (28 patients), the chemotherapy combination was changed by replacing cyclophosphamide with ifosfamide (VAIA). The radiation dose was determined according to chemotherapy response derived by the degree of tumor volume reduction (32 Gy if regression was two thirds of the initial volume and 54.4 Gy in all other patients). ICG Studies. In ICG 79, patients (n 25) were randomized to receive either VAC or modied VAC (VACM).6 Radiation therapy was scheduled at weeks 13 to 14 and dose was dened according to chemotherapy response or the result of second-look surgery, if performed. The dose given was 40 to 45 Gy in patients with microscopic residual disease and 50 to 55 Gy in those with gross residual disease, depending on age and tumor size at diagnosis. Initially, patients with orbital tumors achieving complete clinical remission with primary chemotherapy were not intended to receive radiation therapy, but the protocol was amended after the occurrence of three local relapses in the rst four children treated. Treatment continued with alternate courses of VACM and cyclophosphamide, doxorubicin, and vincristine (CAV) for a total of 12 to 18 months. In the ICG 88 study, patients (n 16) received the VAIA regimen as initial chemotherapy and the VAI regimen as subsequent treatment for a total of 6 to 8 months. Radiation therapy was scheduled at week 11 and given in an accelerated hyperfractionated schedule at a dose of 40 to 54.4 Gy according to chemotherapy response. Radiation therapy was not given to patients with histologically conrmed complete remission. SIOP Studies. Patients in the SIOP 84 study (n 20) received chemotherapy consisting of ifosfamide, vincristine, and dactinomycin (IVA) followed by second-line chemotherapy with doxorubicin and cisplatin in cases of partial response or progressive disease. In the SIOP 89 study (n 24), the same initial chemotherapy combination was used but second-line chemotherapy consisted of vincristine, teniposide/ etoposide, and carboplatin (Vincaepi). The total duration of chemotherapy was 3 to 6 months according to stage. In neither study did patients who achieved complete remission with chemotherapy alone receive radiation therapy as part of rst-line therapy. Those who had residual macroscopic or histologically proven microscopic disease received radiation therapy at a dose of 45 Gy by conventional fractionation.

Patient Population
Analyses were performed on the data derived from nine studies from the four cooperative groups (IRS II, IRS III, IRS IV, CWS 81, CWS 86, ICG 79, ICG 88, SIOP MMT 84, and SIOP MMT 89). The overall study population consisted of 306 children with nonmetastatic RMS of the orbit treated between 1979 and 1992. Patients with parameningeal extension of their tumor, regional nodal involvement, or metastatic disease were all excluded from the analysis. All patients had received histologic conrmation of tumor and all were more than 2 years beyond the date of last treatment. Arrangements for central pathology review existed within each collaborative group structure and diagnoses were not specically rereviewed for this study. Furthermore, the span of years of diagnosis for the patients in this study implied that differences in diagnostic criteria and staging technology will have evolved both within and between the contributing groups. It is possible that these differences may contribute to effects otherwise attributed to differences in treatment strategy.

Treatment
All patients received multiagent chemotherapy, but major differences in philosophy and practice were observed with regard to the use of radiation therapy. IRSG Studies. In the IRS II study (84 patients), children with microscopic residual disease were randomized to receive either a two-drug regimen (vincristine and dactinomycin [VA]) or a three-drug regimen (vincristine, dactinomycin, and cyclophosphamide [VAC]); children with gross residual disease after surgery (usually after initial biopsy only) were randomized to receive either VAC chemotherapy or VAC plus doxorubicin. All patients were scheduled to start irradiation at week 6: the prescribed dose was 41.4 Gy for patients with microscopically positive margins and 45 to 55 Gy for those with gross residual disease, depending on age and on the size of the primary tumor

Downloaded from jco.ascopubs.org on January 30, 2014. For personal use only. No other uses without permission. Copyright 2001 American Society of Clinical Oncology. All rights reserved.

Table 1.
Total % No. of Patients % % % No. of Patients No. of Patients No. of Patients % IRSG CWS ICG SIOP

Patient Characteristics According to Collaborative Group

No. of Patients

No. of patients Period of accrual, year Study, no. of patients IRS II: 841 IRS III: 102 IRS IV: 20 82 25-204 6.8 1-17 16 30 85 71 90 0.94 89 70 NA 32 8 89 35 3 2 179 7 96 35 3 2 80 88 1.0 33 3 1 30 7 92 34 2 1 30 8 7 19 48 20 50 21 1.31 89 25 6 12 81 22 21 94 38 2 3 81 16 7 14 19 18 3 14 20 6.5 1-16 6.7 1-17 7.9 1-17 8 10 17 16 57 26 82 25-204 91 40-159 116 50-189 CWS 81: 14 CWS 86: 28 ICG 79: 25 ICG 88: 16 MMT 84: 20 MMT 89: 24

306 1978-1992

186 1978-1992

40 1981-1990

37 1980-1992

43 1984-1992

86 39-145 6.1 1-14 23

TREATMENT OF ORBITAL RHABDOMYOSARCOMA

50 130 126 51 1.05

157

60 1.53 58

271 19 15

84 36 91 112 14 60 172 14 3 16 14 153 92 32 4 4

51

219 21 66 87 13 10 3 12 12 73

95

Downloaded from jco.ascopubs.org on January 30, 2014. For personal use only. No other uses without permission. Copyright 2001 American Society of Clinical Oncology. All rights reserved.
80 10 10 33 6 4 82 3 8 8 21 52 2 8 10 17 46 1 6 5 31

Follow-up of survivors, months Median Range Age at diagnosis, years Median Range Age distribution 3 years 3-7 years 7 years Sex Male Male:female Primary tumor Orbit Eyelid Both Local extension T1 T2 Size 5 cm 5 cm Unknown Histology Embryonal RMS Embryonal sarcoma Alveolar RMS Initial surgery* Complete removal Microscopic residue Macroscopic residue Biopsy alone

267 10 29

77 14 9

9 38 37 222

73

Abbreviations: T1, Tumor localized to the organ or tissue of origin; T2, Tumor extending beyond the tissue of origin to involve one or more adjacent tissues; NA, not available. *No enucleation or exenteration was carried out as part of initial surgery.

199

200
Statistical Analysis
Statistical analyses were performed at the Institut Gustave Roussy in Villejuif, France, using a general database management system. Survival curves were calculated by the method of Kaplan-Meier.7 Survival was calculated from the date of the start of treatment to the time of the last follow-up or death. Event-free survival (EFS) was calculated from the date of the start of treatment to the date of the rst event, such as failure to achieve complete remission, relapse, or death from any cause. Local control was dened as disappearance of all clinical and radiologic signs of disease or as stable residual radiographic images for 6 months after completion of treatment. The date of analysis was May 1997, providing a minimum of 4 years from the last date of study entry. The statistical signicance of each variable was tested in a univariate analysis using the log-rank test.8

OBERLIN ET AL

Remission, Survival, and Relapse Complete remission was achieved in 294 patients (96%). Recurrence occurred in 51 patients (17%). Median time to relapse was 18 months, with only two relapses occurring beyond 5 years from diagnosis. Among the 51 patients who relapsed, 47 (92%) experienced a local relapse (including three patients with local plus distant relapse). Only four patients (8%) developed distant metastases (Table 3). Among the 47 local recurrences, 27 occurred in nonirradiated patients (27 of 61; 44%) and 20 occurred after radiotherapy (20 of 245; 8%). The median radiation dose received by the patients who relapse after radiation was not different from the median dose given to those who did not. The rate of local relapse varied between the IRSG and European studies (5% in IRS, 30% in CWS, 35% in ICG, and 36% in SIOP). At the time of analysis, the median follow-up of survivors was 7 years. The 10-year actuarial EFS was 77% (range, 71% to 81%) for all patients. Figure 1 shows EFS according to the different treatment groups. This was signicantly better for children in the IRS study than for those in the three other studies (P .001). The 10-year actuarial overall survival (OS) was 87% (range, 82% to 92%) for the whole cohort of patients. Figure 2 shows OS according to the different treatment groups. There was no difference in survival between groups. Figure 3 shows the EFS according to the use of radiotherapy as part of primary treatment and Fig 4 shows the similar analysis for OS. It can be seen that initial radiation therapy has impact on EFS but no impact on OS. Table 3 gives the status of patients at the time of last contact. A total of 273 patients were alive, of whom 238 were in rst remission and 26 were in subsequent remission after relapse; nine were alive with disease. In total, 33 patients had died, of whom seven never achieved complete remission, 21 died after relapse, and four died from treatment-related causes (two with sepsis and two from anthracycline-related cardiomyopathy). The cause of death was unknown in one patient. Burden of Therapy in Surviving Patients Within the IRS group studies, 91% of the surviving patients had received radiation therapy during initial treatment, but data relating to treatment received for relapse were not available. For the children included in the three European group studies, data were available both for initial treatment and for the treatment of any relapse (Table 4). Summation of local therapy given both as part of initial treatment and for the treatment of relapse indicated that the utilization of signicant local treatment was less in the

RESULTS

Patient Characteristics The initial characteristics of the 306 patients are presented in Table 1. The median age at diagnosis was 6.8 years, with a range of less than 1 year to 17 years; 50 patients (16%) were younger than 3 years (IRS 16%, CWS 18%, ICG 8%, and SIOP 23%). Male patients represented 51% of the population. Median follow-up of survivors was 82 months, with a range of 25 to 204 months. Details of stage, size, site, histology, and initial surgery are listed in Table 1. Primary Treatment Overall, 72% (n 222) of patients underwent initial biopsy only and had tumors that were considered unresectable at diagnosis, although there was an apparent difference in surgical practice between the different groups, with the percentage of patients starting therapy after biopsy only varying from 82% (IRSG) to 46% (ICG). A partial excision was achieved by initial surgery in 75 patients (25%), but only nine patients (3%) had a successful initial complete excision (Table 1). The details of radiotherapy are given in Table 2. Altogether, 245 patients (80%) were irradiated, but this proportion varied signicantly between the different groups. In those treated in the SIOP studies, only 16 (37%) were irradiated, all after failure of initial chemotherapy to achieve complete remission. Radiation therapy was withheld in the ICG 88 study only in those with histologically conrmed complete remission, and in all other studies initial irradiation was planned for every patient, although not all actually received it. Overall, 178 patients (93%) in the IRS studies, 28 patients (70%) in the CWS studies, and 28 patients (76%) in the ICG studies received radiation therapy. Among the irradiated cohort, 33 (13%) were younger than 3 years of age (IRS, 16%; CWS, 8%; ICG, 8%; SIOP, 13%).

Downloaded from jco.ascopubs.org on January 30, 2014. For personal use only. No other uses without permission. Copyright 2001 American Society of Clinical Oncology. All rights reserved.

TREATMENT OF ORBITAL RHABDOMYOSARCOMA

Table 2. Details of Radiation Treatment as Primary Treatment
Total IRSG CWS ICG SIOP

201
Prognostic Factors
43 16 37 45 39-64 0 2

No. of patients Patients irradiated No. % Total dose, Gy Median Range Missing data No. of patients irradiated younger than 3 years

306 245 80 45 6-64 29 33

186 173 93 45 6-60 27 27

40 28 70 32 25-72 2 2

37 28 76 46 40-70 0 2

European studies when analyzed collectively (74%) as compared with patients treated in the IRS studies, of whom 91% at least received radiation therapy. Among the European groups, 41% of the surviving SIOP patients had been cured without radiation therapy or radical surgery, compared with 7% for ICG patients and 20% for CWS patients.

Analysis of 10-year OS rates by prognostic variables are listed in Table 5 for the whole population. Survival differed by age and histologic subtype. Age less than 3 years was associated with a less favorable outcome (P .004). Histology was also correlated with survival, with an unfavorable outcome for patients with alveolar histology and embryonal sarcoma compared with patients with embryonal RMS (P .001). Sex, tumor size, T stage, clinical group, and era of therapy did not correlate with outcome. Age and histology seem to be correlated factors: the mean age at diagnosis of children with embryonal RMS was 6.7 years, which is signicantly older than patients with alveolar histology (mean age, 5.5 years). However, histology maintained its prognostic value after adjustment for age (P .001), and age remained signicant even after adjustment for histology (P .007). Relative risks of death were 1.57, 1.6, and 8.8 for patients younger than 3 years with favorable histology, patients older than 3 years with unfavorable

Table 3.
Total

Outcome
IRSG CWS ICG SIOP

No. of patients Patients achieving complete remission No. % Patients experiencing relapse No. % Local, n Local metastatic, n Metastatic,* n Local relapse rates in relation to use of radiation therapy No. of patients not irradiated Patients experiencing local relapse No. % No. of patients irradiated Patients experiencing local relapse No. % Status of patients at last follow-up Alive In rst CR In CR after relapse With disease Dead Without rst CR After relapse From therapy Unknown cause

306 294 96 51 17 44 3 4 61 27 44 245 20 8 273 238 26 9 33 7 21 4 1

186 13 94 13 7 9 0 4* 13 3 173 6

40 40 100 10 25 8 2 0 12 4 28 6

37 37 100 13 35 12 1 0 9 7 28 6

43 42 98 15 36 15 0 0 27 13 16 2

170 160 3 7 16 6 8 1 1

35 28 7 0 5 0 3 2 0

31 24 6 1 6 0 6 0 0

37 26 10 1 6 1 4 1 0

Abbreviation: CR, complete remission. *Patterns of metastatic relapse: one bone marrow; one bone marrow brain; one bone marrow bone; one bone marrow bone eye (all embryonal RMS).

Downloaded from jco.ascopubs.org on January 30, 2014. For personal use only. No other uses without permission. Copyright 2001 American Society of Clinical Oncology. All rights reserved.

202
Table 4. Local Therapy in European Patients Alive Without Disease (in rst or subsequent remission)
Total CWS ICG SIOP

OBERLIN ET AL

No. of patients alive in remission Treatment Chemotherapy only Chemotherapy conservative surgery Total treated without signicant local therapy No. % Treatment Radiotherapy Radiotherapy conservative surgery Exenteration Exenteration radiotherapy Total treated with signicant local therapy No. % Missing data: therapy after relapse unknown

101 12 12 24 24 48 19 3 5 75 74 2

35 1 6 7 20 15 7 1 4 27 77 1

30 2 0 2 7 18 8 1 1 28 93 0

36 9 6 15 41 15 4 1 1 21 59 1

Fig 2. OS according to treatment groups: IRS, 88% (range, 80% to 95%); CWS, 87% (range, 77% to 98%); ICG, 85% (range, 74% to 97%); and SIOP, 85% (range, 74% to 96%).

histology, and patients younger than 3 years with unfavorable histology, as compared with the group of patients older than 3 years with favorable histology. Importantly, there was no statistical difference in OS between the irradiated patients and the nonirradiated patients (87% v 86%), although EFS of irradiated patients was signicantly better (82% v 53%). This implies that a signicant number of nonirradiated patients could be cured after relapse by using radiation therapy at that time. Late Effects of Therapy An attempt was made to systematically collect detailed information about late effects of therapy amongst patients from the European studies. Questionnaires were completed by the treating clinician for 83 (80%) of 103 surviving

patients (CWS, 60%; ICG, 97%; and SIOP, 78%. Data were available for comparison from a recent publication of a different series of 94 patients treated in the IRS III study9 (Table 6). The European questionnaire evaluations were undertaken at a median age of 16 years (range, 6 to 28 years) and at a median follow-up from diagnosis of 8 years. The most frequent sequelae were dry eye (72%), impaired vision in the treated eye (54%), cataract (51%), ptosis (36%), and orbital hypoplasia (29%). All occurred almost exclusively in patients who had been irradiated. Similar results were noted from data provided from the IRSG patients, although orbital hypoplasia (59%) and cataract (82%) were noticeably more frequent in this group. Seven patients had evidence of growth impairment. Data on other endocrine functions and fertility were poorly described and could not be adequately assessed from this retrospective data collection.

Fig 1. EFS according to treatment groups: IRS, 86% (range, 78% to 89%); CWS, 70% (range, 58% to 88%); ICG, 64% (range, 47% to 80%); and SIOP, 58% (range, 43% to 76%) (P < .001).

Fig 3. EFS according to initial radiotherapy: radiotherapy, 82% (range, 76% to 87%); no radiotherapy, 53% (range, 76% to 87%) (P < .001).

Downloaded from jco.ascopubs.org on January 30, 2014. For personal use only. No other uses without permission. Copyright 2001 American Society of Clinical Oncology. All rights reserved.

TREATMENT OF ORBITAL RHABDOMYOSARCOMA

Table 6. Late Sequelae of Treatment
European Patients (pooled data) Type of Sequelae No. of Survivors Affected %

203

IRSG Patients* % Survivors Affected (n 94)

Fig 4. OS according to initial radiotherapy: radiotherapy, 86% (range, 82% to 93%); no radiotherapy, 86% (range, 77% to 95%).

Globe removed Cataract Reduced vision in affected eye Dry eye Painful eye Keratitis Corneal ulcer Retinal damage Ptosis Orbital hypoplasia Maxillary hypoplasia

9/83 36/71 37/68 20/72 10/72 10/72 4/72 8/72 26/72 27/72 7/72

11 51 54 28 14 14 6 11 36 29 10

14 82 70 23 14 18 4 3 28 59 11

DISCUSSION

*IRSG data based on data from patients treated in the IRS III study.9

Children with orbital RMS have an excellent prognosis for cure. The results of this analysis of 306 patients without

Table 5.

Overall Survival Rate by Prognostic Variables

Total No. Deaths OS at 10 Years (%) Relative Risk Signicance

Age 3 years 3 years Histology Embryonal sarcoma or alveolar RMS Embryonal RMS Sex Male Female Tumor size 5 cm 5 cm Unknown Clinical group Complete removal Microscopic residue Macroscopic residue or biopsy Radiotherapy Yes No Collaborative group IRS CWS ICG SIOP Era of therapy 1978-1983 1984-1988 1989

50 256 39 267 157 149 219 21 66 9 38 259

12 21 11 22 17 16 22 2 9 3 2 23

68 90 68 90 88 86 88 90

3.3 1 3.6 1

P .004

P .001

NS

NS

78 97 89

NS

245 61 186 40 37 43 85 125 96

25 8 16 5 6 6 11 14 8

87 86 88 87 86 85 88* 91* 91*

NS

NS

NS

Abbreviation: NS, not signicant. *5-year OS.

parameningeal extension, regional lymph node involvement, or distant metastatic disease show a 10-year OS rate of 87%, regardless of the initial approach to therapy. Nevertheless, the risk of late sequelae of treatment is signicant, particularly in relation to the consequences of local therapy. The most comprehensive published review of the late effects induced by radiotherapy derives from the report of experience from the IRSG I study.4 This study of 50 children surviving more than 6 years after treatment in the IRS I study and incorporating radiation to the orbit at doses between 50 and 60 Gy showed late effects that included reduced vision (80%), cataract (92%), orbital hypoplasia (52%), and facial asymmetry (41%). Thirteen patients (27%) had required surgery for cataract extraction, correction of ptosis, or dilatation of the lachrymal ducts, and four patients (8%) had had late enucleation for the complications of therapy. Sixty-one percent had evidence of growth impairment. These data, however, derive from a treatment strategy initiated 20 years ago, and subsequent improvement in radiation techniques may have reduced the incidence or lessened the severity of such sequelae. Nevertheless, total radiation doses of more than 45 Gy remain in current practice, and preliminary data from patients treated in IRS III from 1984 through 1991 suggest that changes in equipment, planning techniques, or fractionation schedules are unlikely to have signicantly diminished the problems encountered.9 Data from patients treated at higher doses (55 to 60 Gy) suggest higher rates of visual loss.10 The results of the analysis of the European patients in this study and the data recently published by the IRS group conrm that similar problems arise in patients treated in more recent eras (Table 6). The debate should focus on whether it is possible to cure patients with orbital RMS without local treatment and, if

Downloaded from jco.ascopubs.org on January 30, 2014. For personal use only. No other uses without permission. Copyright 2001 American Society of Clinical Oncology. All rights reserved.

204
so, at what cost in terms of any diminution of survival or in the expression of alternative late effects of treatment relating to chemotherapy. Data from this study do not permit any clear conclusions to be made about the efcacy, or otherwise, of reduced-dose radiotherapy. To date, the efcacy of doses less than 40 Gy is unproved and any signicant reduction in the adverse effects on facial growth and visual function would not be expected until the total dose administered was substantially below this level. Indeed, cataract formation is reported at radiation doses below 10 Gy.11 The concept of the total burden of therapy is important in considering the cost of survival, particularly in patients who survive after relapse and who have therefore been exposed to additional treatment. It is often the case that the disadvantages of radiotherapy are more obvious clinically than the sequelae resulting from chemotherapy; longer follow-up and careful and systematic investigation are required to determine the incidence and importance of some of the other possible consequences of treatment, eg, infertility after exposure to alkylating agents and cardiotoxicity after anthracycline exposure, as well as the risk of second malignancy induced by both chemotherapy and radiotherapy. Detailed data relating to the quality control of radiation therapy were not available for all patients in this study.

OBERLIN ET AL

Overall, the local failure rate for patients treated with radiotherapy was small (8%; Table 3), conrming the efcacy of such treatment, and although the incidence of relapse after radiotherapy was higher in the European studies, the results do not permit clear conclusions to be made to explain this in terms of the dose given. Other variables that may inuence the efcacy of local control achieved with radiotherapy include the criteria for selection of patients for such treatment (for example, in relation to chemotherapy response, histology, and age) and the quality of treatment given. Nevertheless, it is clear that important late effects are encountered by a signicant number of survivors, and the experience of the strategy promoted by the SIOP group illustrated in this study and reported previously by Rousseau et al3 suggests that up to 40% of patients with localized RMS of the orbit can be treated successfully without the use of radiotherapy and without disadvantage to the survival of the whole group. The challenge for the future must be to identify the characteristics shown by patients who can be safely treated in this manner without a signicant risk of relapse and to ensure that radiation therapy is delivered to the remaining patients in a manner that is both effective and offers the least chance of unacceptable late sequelae.

REFERENCES
1. Crist WM, Garnsey L, Beltangady M, et al: Prognosis in children with rhabdomyosarcoma: A report of the Intergroup Rhabdomyosarcoma Studies I and II. J Clin Oncol 8:443-452, 1990 2. Crist W, Gehan EA, Rajab AH, et al: The Third Intergroup Rhabdomyosarcoma Study. J Clin Oncol 13:610-630, 1995 3. Rousseau P, Flamant F, Quintana E, et al: Primary chemotherapy in rhabdomyosarcoma and other malignant mesenchymal tumours of the orbit: Results of the International Society of Paediatric Oncology MMT 84 Study. J Clin Oncol 12:516-521, 1994 4. Heyn R, Ragab A, Raney RB, et al: Late effects of therapy in orbital rhabdomyosarcoma in children: A report from the Intergroup Rhabdomyosarcoma Study. Cancer 57:1738-1743, 1986 5. Flamant F, Rodary C, Rey A, et al: Treatment of non metastatic rhabdomyosarcoma in childhood and adolescence: Results of the second study of the International Society of Paediatric Oncology; MMT 84. Eur J Cancer 34:1050-1062, 1998 6. Carli M, Pastore G, Perilongo G, et al: Tumor response and toxicity after single high-dose versus standard ve-day divided-dose dactinomycin in childhood rhabdomyosarcoma. J Clin Oncol 6:654658, 1988 7. Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 8. Mantel N: Evaluation of survival data and two new rank order situations arising from its consideration. Cancer Chemother Rep 50:163-170, 1966 9. Raney RB, Anderson J, Kollath J, et al: Late effects of therapy in 94 patients with localized rhabdomyosarcoma of the orbit: Report from the Intergroup Rhabdomyosarcoma Study (IRS)-III, 1984-1991. Med Pediatr Oncol 34:413-420, 2000 10. Abramson DH, Notis CM: Visual acuity after radiation for orbital rhabdomyosarcoma. Am J Ophthalmol 11:808-809, 1994 11. Merriam GR Jr, Focht EF: A clinical study of radiation cataracts and the relationship to dose. Am J Roentgenol Rad Ther Nuclear Med 77:759-789, 1957

Downloaded from jco.ascopubs.org on January 30, 2014. For personal use only. No other uses without permission. Copyright 2001 American Society of Clinical Oncology. All rights reserved.