The Journal of Craniofacial Surgery

& Volume 22, Number 6, November 2011

Brief Clinical Studies

facial atrophy of our patient might relate with the neurovascular compression of trigeminal nerve as her HFSs pathogenesis. We found an offending vessel at the REZ of facial nerve and settled with the compression. Nevertheless, we got no intraoperative evidence in operative exploration of ipsilateral trigeminal nerve, and the finding of the upper right of Figure 2 was false-positive. The finding of the operation demonstrated that the HFSs pathogenesis of our patient was neurovascular compression, but progressive facial hemiatrophy was not, although we found no evidence of the association between HFS and progressive facial hemiatrophy in our patient. Infection with Borrelia burgdoferi in patients with progressive facial hemiatrophy was mentioned in several reports.3,4 Eliminating the possibility of neurovascular compression of progressive facial hemiatrophys pathogenesis in our patient, we review our patients history. Retrospectively, her right facial hemiatrophy followed the tinea on the ipsilateral face. It was possible that an unknown infection of the dermis would be the cause of our report, and the cutaneous neurosis might be the pathogenesis of our patients facial atrophy. Unfortunately, we could not collect enough evidence of our hypothesis on our patient more than 40 years after the period of onset. In conclusion, we present a case of right-sided progressive hemifacial atrophy with HFS. The neurovascular compression of trigeminal nerve could be ruled out, which was proved by operative exploration. Progressive hemifacial atrophy combined with HFS in our patient might be a chance event. Its pathogenesis that might relate with dermis infection needs further study.

13. Kaufman MD. Masticatory spasm in facial hemiatrophy. Ann Neurol 1980;7:585Y587 14. Pensler JM, Murphy GF, Mulliken JB. Clinical and ultrastructural studies of Rombergs hemifacial atrophy. Plast Reconstr Surg 1990;85:669Y674 15. Ebersbach G, Kabus C, Schelosky L, et al. Hemimasticatory spasm in hemifacial atrophy: diagnostic and therapeutic aspects in two patients. Mov Disord 1995;10:504Y507 16. Pensler JM, Murphy GF, Mulliken JB. Hemimasticatory spasm associated with localized scleroderma and facial hemiatrophy. Arch Neurol 2000;57:576Y580 17. Menascu S, Padeh S, Hoffman C, et al. Parry-Romberg syndrome presenting as status migrainosus. Pediatr Neurol 2009;40:321Y323 18. Hamasaki K, Kubo K, Kanda H. Progressive facial hemiatrophy complicated by sclerodactyly, Raynauds phenomenon, anti-ribonucleoprotein antibody, and trigeminal nerve disturbance. Mod Rheumatol 2005;15:198Y200 19. Tebloev IK, Kalashnikov I. Pathogenesis of symptomatic facial hemiatrophy. Zh Nevropatol Psikhiatr Im S S Korsakova 1979;79:413Y416 20. Koziarski A, Mandat T, Warczynska A, et al. Surgical treatment of trigeminal neurinoma. Neurol Neurochir Pol 2002;36:587Y595

Primary Pediatric Osteosarcoma of the Skull

Emily J. Kirby, MD,* Hong Holly Zhou, MD, Louis Morales, Jr, MD* Abstract: A 16-year-old adolescent boy presented with osteosarcoma of the left parieto-occipital bone with no history of radiation, Paget disease, or retinoblastoma. The tumor presented as a painless lump, which he attributed to minor trauma 5 months before presentation. Biopsy of the lesion was inconclusive. Complete surgical resection was attained. Pathology revealed a diagnosis of osteosarcoma, grade 1 to 2/3. Adjuvant chemotherapy commenced after tumor resection. Primary pediatric osteosarcoma of the skull is an exceedingly rare malignancy, with only a handful of cases ever reported. Complete surgical resection with negative margins is the key to optimizing disease-free survival. Adjuvant chemotherapy is recommended for high-grade tumors and in cases of incomplete resection. Key Words: Osteosarcoma, osteogenic sarcoma, pediatric, adolescent, surgery, tumors of the skull arcomas account for 0.5% of all malignancies in humans.1,2 Only 20% of sarcomas are of the bony or cartilaginous type, compared to 80% arising from soft tissue.2 Among bony tumors, however,

REFERENCES
1. Lonchampt P, Emile J, Pelier-Cady MC, et al. Central sympathetic dysregulation and immunological abnormalities in a case of progressive facial hemiatrophy (Parry-Romberg disease). Clin Auton Res 1995;5:199Y204 2. Longo D, Paonessa A, Specchio N, et al. Parry-Romberg syndrome and rasmussen encephalitis: possible association. Clinical and neuroimaging features. J Neuroimag 2011;21:188Y193 3. Sahin MT, Baris S, Karaman A. Parry-Romberg syndrome: a possible association with borreliosis. J Eur Acad Dermatol Venereol 2004;18:204Y207 4. Baskan EB, Kacar SD, Turan A, et al. Parry-Romberg syndrome associated with borreliosis: could photochemotherapy halt the progression of the disease? Photodermatol Photoimmunol Photomed 2006;22:259Y261 5. Buonaccorsi S, Leonardi A, Covelli E, et al. Parry-Romberg syndrome. J Craniofac Surg 2005;16:1132Y1135 6. Carreno M, Donaire A, Barcelo MI, et al. Parry Romberg syndrome and linear scleroderma in coup de sabre mimicking rasmussen encephalitis. Neurology 2007;68:1308Y1310 7. Dupont S, Catala M, Hasboun D, et al. Progressive facial hemiatrophy and epilepsy: a common underlying dysgenetic mechanism. Neurology 1997;48:1013Y1018 8. Stone J. Parry-Romberg syndrome: a global survey of 205 patients using the Internet. Neurology 2003;61:674Y676 9. Sommer A, Gambichler T, Bacharach-Buhles M, et al. Clinical and serological characteristics of progressive facial hemiatrophy: a case series of 12 patients. J Am Acad Dermatol 2006;54:227Y233 10. Drummond PD, Hassard S, Finch PM. Trigeminal neuralgia, migraine and sympathetic hyperactivity in a patient with Parry-Romberg syndrome. Cephalalgia 2006;26:1146Y1149 11. Karim A, Laghmari M, Ibrahimy W, et al. Neuroretinitis, Parry-Romberg syndrome, and scleroderma. J Fr Ophtalmol 2005;28:866Y870 12. Brito JC, Holanda MM, Holanda G, et al. Progressive facial hemiatrophy (Parry-Romberg disease). Report of two cases associated with trigeminal neuralgia and cramps. Arq Neuropsiquiatr 1997;55:472Y477

From the *Craniofacial Foundation of Utah; Primary Childrens Medical Center, Department of Pathology; and Plastic Surgery Associates, Salt Lake City, Utah. Received May 9, 2011. Accepted for publication July 10, 2011. Address correspondence and reprint requests to Emily J. Kirby, MD, Craniofacial Foundation of Utah, 5089 S 900 E, Suite 100, Salt Lake City, UT 84117; E-mail: ekirbymd@gmail.com The authors report no conflicts of interest. Copyright * 2011 by Mutaz B. Habal, MD ISSN: 1049-2275 DOI: 10.1097/SCS.0b013e318231fe9b

* 2011 Mutaz B. Habal, MD

2399

Copyright 2011 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.

Brief Clinical Studies

The Journal of Craniofacial Surgery

& Volume 22, Number 6, November 2011

FIGURE 1. Initial presentation of painless mass on CT imaging. osteosarcoma is the most common primary malignant tumor across all ages, including children and adolescents,3Y5 comprising 20% to 60% of malignant bone tumors.3,5Y7 Osteosarcoma most commonly manifests in the long bones of adolescents (80%),4,7 specifically in the femur (42%), followed by the tibia (19%) and humerus (10%).3 Craniofacial osteosarcoma accounts for only 6% of all osteosarcomas.5Y12 Most craniofacial osteosarcomas arise in the maxilla, mandible, and zygoma.6,7,11,12 Skull osteosarcomas represent only 11% of craniofacial osteosarcomas and 1.6% of all osteosarcomas.7,12Y15 In the pediatric population, osteosarcoma is the eighth most common childhood cancer (2.4%), after leukemia (30%), brain/ central nervous system (22.33%), neuroblastoma (7.3%), Wilms tumor (5.6%), nonYHodgkin lymphoma (4.5%), rhabdomyosarcoma (3.1%), and retinoblastoma (2.8%).3 Only 2% to 5% of pediatric osteosarcomas arise in the head and neck.4,16 Most patients who develop these rare skull tumors have welldefined risk factors for developing osteosarcoma.13 Whereas the cause of osteosarcoma remains unknown, predisposing factors include a history of Paget disease, previous radiation, retinoblastoma, and fibrous dysplasia.4,5,7,11,12,17Y19 Retinoblastoma patients have an increased risk of developing osteosarcoma, independent of radiation exposure.19,20 During the course of a lifetime, osteosarcoma develops in 0.01% to 0.03% of all irradiated patients12 and in 1% of fibrous dysplasia patients.20 Less often, genetic predilection, multiple osteochondromatosis, Li-Fraumeni syndrome, RothmundThomson syndrome, chronic osteomyelitis, myositis ossificans, trauma, and exposure to thorium oxide (a radioactive scanning agent) have been linked to the development of osteosarcoma.1,4,7,11,20Y22

FIGURE 3. A and B, Computed tomographic scans showing left parietal mass with aggressive features. to his primary care physician with complaints of a painless mass on the left side of his head. It had been present since a fall from a bunk bed 5 months prior. The lesion was thought by the physician to be a lymph node or a cyst, and a computed tomographic (CT) scan was obtained, which suggested the lesion was a maturing subperiosteal hematoma (Fig. 1). Three months later, MP returned to his primary care physician, worried that the lesion had enlarged. He was sent to a craniofacial surgeon for consultation. The patient was taken to the operating room 1 month later by the craniofacial surgeon for excision of a left parieto-occipital cephalhematoma. The mass was noted to be 4.5 cm in diameter, over the left parieto-occipital area. Comment was made at the time of the procedure that the lesion appeared to violate the outer table into the diploic space in some areas, but mainly remained above the outer table. The pathologist determined that the lesion was likely fasciitis ossificans (a rare form of heterotopic bone formation) or an ossifying cephalhematoma. One year after removal of the lesion, MP returned to the plastic surgeon with complaints of recurrence of the cephalhematoma and pain in the area. Otherwise, he had no symptoms. Plans were made for removal of the lesion in June, after the patients high school graduation. The patient and his family sought a second opinion, at which time he presented to our craniofacial clinic (Fig. 2). A CT scan at this time showed a left parietal mass with intracranial extension and aggressive features (Fig. 3). He was seen by a neurosurgeon, who reviewed systems and performed a neurological examination, the results of which were benign except for the left parietal mass. Magnetic resonance (MR) imaging confirmed the suspicious tumor (Fig. 4), and a biopsy was performed the following day by the neurosurgeon. Laboratory tests showed a total calcium level of 9.40 mg/dL (reference range, 9Y10.5 mg/dL) and alkaline phosphatase level of 139.00 U/L (reference range, 20Y70 U/L in adults and 20Y150 U/L in children). A CT scan of the chest and nuclear medicine bone scans revealed no tumor metastasis. An MR image showed prominent suboccipital subcutaneous lymph nodes. Pathology evaluation

CLINICAL REPORT
MP is a 16-year-old white adolescent boy with history of asthma, attention-deficit/hyperactivity disorder and obesity who presented

FIGURE 2. View of left parietal skull mass after biopsy, before definitive resection.

FIGURE 4. Magnetic resonance image of coronal (A) and axial (B) views confirming suspicious mass. * 2011 Mutaz B. Habal, MD

2400

Copyright 2011 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.

The Journal of Craniofacial Surgery

& Volume 22, Number 6, November 2011

Brief Clinical Studies

FIGURE 5. Preoperative markings incorporating previous biopsy scar. of the tumor tissue showed a myxoid spindle cell neoplasm. Differential diagnosis included myoepithelial soft tissue tumor, fibromyxoid tumor, and extraskeletal chondrosarcoma. Case workup with immunohistochemistry failed to render a specific diagnosis. Definitive resection of the tumor by both neurosurgery and craniofacial plastic surgery with immediate reconstruction by the craniofacial surgery team was performed 1 month after the biopsy. Care was taken to include the previous biopsy scar within the tumor specimen. This was done by marking the skin preoperatively with a W-plasty pattern, including a 1-cm margin surrounding the previous scar (Fig. 5). The tumor was resected en bloc via craniectomy using stereotactic MR guidance (Fig. 6). A separate layer of dura representing the deep margin was sent to the pathology department, as well as a prominent left neck lymph node more than 1 cm in size, both of which were negative for tumor. All surgical margins were negative for tumor. Immediate reconstruction of the 10 14-cm defect was performed using a combination of split rib graft, split cranial bone graft, and demineralized bone (Fig. 7). Grafts were secured using titanium plates and stainless steel wires. The reconstruction site was then covered with fibrin sealant (Tisseel; Baxter BioSurgery, Deerfield, IL). The scalp was mobilized in every direction and closed tensionfree using the planned W-plasty closure. MP was admitted overnight to the pediatric intensive care unit before transferring to a floor bed on postoperative day 1. He had an uneventful recovery, with discharge from the hospital on postoperative day 4. Three months after tumor resection, the patient began chemotherapy with doxorubicin, cisplatin, and high-dose methotrexate. His course was complicated by multiple episodes of seizure activity followed by a violent reaction to the antiseizure medication, Keppra. He developed a diffuse left upper extremity deep vein thrombosis involving the internal jugular, innominate, subclavian, and axillary veins, for which he was given anticoagulants. After starting high-dose methotrexate, the patient was admitted again to the hospital for recurrent oral and anal mucositis. One month later, the patient was admitted again for a neutropenic fever. In the hos-

FIGURE 7. Split rib graft harvested for skull reconstruction.

pital, he was noted to be bradycardic and hypotensive, with septic shock. He was transferred to the pediatric intensive care unit, where he was found to have supraventricular tachycardia, which was successfully converted using adenosine. He required a blood transfusion during this time. Stool cultures were positive for Clostridium difficile colitis, which was treated using metronidazole. Postoperative photos were obtained at this time, 5 months after surgical resection and reconstruction (Fig. 8). Multiple postoperative imaging studies have demonstrated a stable cranial reconstruction with no evidence of local recurrence (Fig. 9).

PATHOLOGY
On gross examination, the left temporal parietal bone mass specimen measured 12.7 cm anterior-to-posterior and 10 cm superior-to-inferior. There was a 10 8-cm tumor mass arising from bone with extension to overlying soft tissues and to the epidural region. Histologically, the tumor was composed of spindle cells forming cords and a reticular network in myxoid matrix. There were areas of eosinophilic extracellular material consistent with osteoid/ bone matrix (Fig. 10). The bone, soft tissue, and deep dura margins were negative for tumor. Whereas many tumor cells appeared to be bland, there was focally increased cellularity and mitotic activity. Immunohistochemistry revealed the tumor cells were positive for vimentin and negative for A1/A3 cytokeratin, epithelial membrane antigen, S100 protein, and neuron-specific enolase. The INI-1 nuclear stain was retained. Ewing sarcoma breakpoint region and fluorescence in situ hybridization analysis showed no evidence of Ewing sarcoma breakpoint region rearrangement. Cytogenetic chromosomal study of the tumor tissue showed no abnormality. Pathology slides from this case were reviewed in consultation by outside bone

FIGURE 6. En bloc tumor resection: extracranial (A) and intracranial (B) views of specimen. * 2011 Mutaz B. Habal, MD

FIGURE 8. Postoperative photos: anterior-posterior (A) and lateral views (B) 5 months after surgery.

2401

Copyright 2011 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.

Brief Clinical Studies

The Journal of Craniofacial Surgery

& Volume 22, Number 6, November 2011

FIGURE 9. Postoperative imaging 5 months after resection: axial (A) and coronal (B) MR image views showing stable skull reconstruction and no evidence of local recurrence. pathology experts, and the final diagnostic interpretation was grade 1 to 2/3 osteosarcoma with a fibromyxoid pattern.

DISCUSSION
Osteosarcoma is defined by the production of osteoid, an unmineralized osseous matrix, by atypical, neoplastic osteoblasts that have unique biologic activity and morphologic patterns.4,5,9,11,13,15 Primary pediatric osteosarcoma of the head and neck is extremely rare, with most of these tumors found in the jaws.8 Since the first report of osteogenic sarcoma of the skull in 1945,23 only a handful of primary skull osteosarcomas have been reported, very few of which were in pediatric patients.7,8,15,24 Shinoda et al17 reviewed 98 previously reported cases of primary osteosarcoma of the skull and added their case to the tally, totaling 99 reported cases. In 2003, however, Kanazawa et al11 reported that 53 cases of osteosarcoma of the skull had been reported to date. It is unclear exactly how many of these previously reported cases were in pediatric patients because the demographic reporting is not consistent among the studies. Nora et al13 surveyed the Mayo Clinic files of 7000 bone tumors to find only 21 patients with skull osteosarcoma, 6 of which were primary tumors. None of these were in pediatric patients. Thompson et al25 examined 25 years of patient records to find 14 skull osteosarcomas, 2 of which were primary, none of which were pediatric. Similarly, Oda et al6 surveyed 15 years of patient records to find 2 skull osteosarcomas, 1 of which was primary. None of these were pediatric patients. Mark et al20 looked at 32 years of patient records and found 2 osteosarcomas in the skull, none of which were primary pediatric patients. Perhaps the largest report of primary pediatric skull osteosarcoma was reported in 1985 by Huvos et al.15 They reviewed 60 years of patient records to identify 19 craniofacial osteosarcomas, 10 of which arose in the skull, 3 of which were in pediatric patients. Two additional case reports7,24 described 1 pediatric patient each with primary skull osteosarcoma. Our search of the literature indicates there may be as few as 5 to 8 primary pediatric skull osteosarcomas ever reported.4Y10,13Y17,20,24Y30 Rowbotham31 insightfully postulated in 1957, Osteogenic sarcomata, such as occurs in the long bones of the limbs, must be exceedingly rare in the skull: it is doubtful whether they occur. Patients with primary craniofacial osteosarcoma are usually in their third or fourth decade of life.4,6,8,11,16 Secondary osteosarcoma arises later in life, often in the fifth or sixth decade.15,26 Patients typically present 5 to 6 months after symptoms begin,2,4,15,26 similar to MP. Equal sex distribution is found in craniofacial osteosarcomas, compared to long bone osteosarcoma, which shows clear male predominance.5,6,8,9,15,17 Osteosarcoma in any location affects blacks and Hispanics more often than whites.3

Symptoms of craniofacial osteosarcoma are dependent on tumor location,4,8,13,26 with no one symptom complex differentiating it from benign tumors or nonneoplastic conditions.10 Headaches, painless (or painful) lumps, cranial nerve deficiency, or cranial hypertension may be seen, with many signs and symptoms appearing after local invasion is already present.4,6,8,10,13Y15,26 The tumors are often asymptomatic and present late as large tumors that carry a poor prognosis.13Y15 In 1 study,15 6 of 19 skull osteosarcomas were initially diagnosed as benign tumors. Diagnosis by radiographic imaging may also pose a challenge in the patient with craniofacial osteosarcoma. Only 41% of osteosarcomas were correctly diagnosed based on radiographs.9 Radiographic findings of osteosarcoma vary depending on the relative amount of osteolytic and osteoblastic activity.10,11 Some tumors are radiolucent with little or no neoplastic bone visible, whereas others appear sclerotic or may be associated with a soft tissue mass.10 Tumor outlines may appear irregular and may penetrate the bone or exhibit periosteal new bone formation.13 The typical sun ray and sun burst patterns are seen in 25% to 31% of patients.9,11 Huvos et al15 report radiographic findings in 2 patients that mimicked benign entities. Another study noted no bone abnormalities in 2 patients.10 Even surgical biopsy confirmed craniofacial osteosarcoma in only 13 of 28 patients.8 The diagnosis of MPs tumor was difficult as well, with initial imaging and pathology reports supporting a benign lesion. Some laboratory values may be elevated in osteosarcoma patients. Serum alkaline phosphatase was found to be elevated in 6 of 22 patients. Serum calcium level was elevated in 6 of 26 patients.10 At 18 years old, our patient showed a serum calcium level that was within the reference range, but his alkaline phosphatase level was elevated at 139.00 U/L (reference range, 20Y70 U/L). Although these laboratory values are not diagnostic, they may help to point toward a diagnosis of osteosarcoma with other features suggestive of the disease. Pathologic diagnosis of craniofacial osteosarcoma can be challenging. In contrast to typically hypercellular, anaplastic tumor cells seen in conventional osteosarcoma arising in appendicular bone, osteosarcomas in the craniofacial region tend to be less anaplastic (ie, lower grade).5 They are often spindle cell tumors that are associated with irregular bone production, low cellularity, few mitoses, and a general absence of atypia.11 Many previous studies reviewing craniofacial osteosarcomas found that chondrogenic osteosarcoma is the most commonly encountered subtype,4,5,9,32,33 although the Armed Forces Institute of Pathology series found osteoblastic osteosarcoma to be more common.34 In their study, Vege et al9 reported that chondrogenic osteosarcomas of craniofacial bones often contain large myxoid and chondroid areas that may lead to

FIGURE 10. Pathology slides of tumor tissue. A, Osteosarcoma with neoplastic bone formation and abundant myxoid stroma (hematoxylin and eosin; original magnification, 200). B, Tumor cells grow in cords, forming a reticular network. There are focal areas of extracellular eosinophilic material representing neoplastic bone matrix (hematoxylin and eosin; original magnification, 200). * 2011 Mutaz B. Habal, MD

2402

Copyright 2011 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.

The Journal of Craniofacial Surgery

& Volume 22, Number 6, November 2011

Brief Clinical Studies

underdiagnosis of osteosarcomas as myxomas, chondromas, myositis ossificans, salivary glands neoplasms, or benign fibro-osseous lesions. Goepfert et al5 also observed that chondroblastic osteosarcomas of the head and neck contain unusual myxoid, periosteal osteosarcoma-like, clear cell, and hyaline forms, although they found that these subtypes did not seem to have direct impact on clinical outcome.5 In this case, initial biopsy findings of a highly myxoid lesion composed of deceptive banal cells without convincing malignant osteoid matrix led to consideration of a reactive lesion and other myxoid tumors of soft tissue. Osteosarcoma was not favored until the definitive resected tumor was evaluated. This case highlights that the pathologic features of craniofacial osteosarcoma may differ considerably from those in the osteosarcoma of appendicular bone. Wide surgical resection of tumor with adjacent soft tissue and bony margins is the treatment of choice for craniofacial osteosarcoma.4Y6,11,14,19,35,36 Although adequate bony margins are crucial, there is no consensus on what constitutes an adequate margin. Determination of such margins intraoperatively is difficult, if not impossible.5,14,19 Regardless, the patients prognosis depends on the ability of the surgeon to achieve complete surgical resection.12 One study found the presence of osteosarcoma at final resection margins to be uniformly fatal.5 At the time of tumor resection, there should be direct inspection for intracranial involvement of the tumor.12 If the periosteum is tightly adherent to the tumor, it should be included en bloc with the resection to serve as a barrier to tumor seeding during removal.12 MPs pathology specimen was suspicious for tumor cells near the internal margin, but a separate layer of dura that was resected confirmed negative margins. Although multiple studies describe the importance of negative margins for survival, current studies report up to 67% positive margins in patients with skull osteosarcoma.14,35 Complex anatomy of the craniomaxillofacial region likely contributes to the fear of damaging important structures that are difficult to reconstruct in this high percentage of positive margins.8,14 Close proximity to the orbit and intracranial cavities makes these tumors more difficult to resect.8 Fear of damaging vital structures is warranted because wound dehiscence, dura exposure, brain exposure, cerebrospinal fluid leak, and meningitis can be fatal.14 There may also be a risk of microdissemination or local shedding of tumor cells during resection.17 In recent years, however, improved knowledge of anatomy, advanced imaging studies, improved surgical technique, advances in reconstruction materials, and new adjuvant therapies make complete tumor resection more possible than ever before.12 Other than complete surgical resection, no other treatment has been firmly established for skull osteosarcoma.2 Several studies support adjuvant or neoadjuvant chemotherapy for high-grade tumors and for tumors anticipated to be difficult to resect (eg, resection likely to leave positive margins or a poor functional result).5,6,8,26,35 In many studies, however, the extreme rarity of skull osteosarcoma leads to an insufficient number of cases in each treatment group to determine a statistically significant result.4,13 Smeele et al36 showed that in 60 craniofacial osteosarcoma patients, those treated with a combination of surgery and chemotherapy had improved disease-free and overall 5-year survival rates compared to surgery alone. Chemotherapy has also been shown to improve outcomes of 201 craniofacial osteosarcoma patients who had complete or incomplete surgical resection of the tumor.2,8,12,36 It was previously thought that since craniofacial osteosarcoma is associated with less risk of distant metastasis, the role of chemotherapy was unclear.2,9 Local control of the tumor, however, remains a significant problem.9 Salvati et al26 showed that the use of chemotherapy in treatment of skull osteosarcoma reduces the risk of local relapse, reduces the number of metastases, and lengthens the time needed for their formation and clinical onset. * 2011 Mutaz B. Habal, MD

Less convincing data are available for the role of radiation therapy in the treatment of skull osteosarcoma. Radiation has been used for high-grade tumors, large lesions, close or positive surgical margins, and inoperable tumors.2 Radiation therapy, however, does not seem to significantly improve overall survival in head and neck osteosarcomas.2 This is likely because osteosarcomas are relatively radioresistant.6 Some studies have suggested that radiation therapy be reserved only for treatment of local recurrence or in metastatic disease for palliation.5 Kassir et al19 reported no survival benefit in head and neck osteosarcoma with the addition of chemotherapy or radiation therapy. Metastatic disease is uncommon in head and neck osteosarcomas.6 Only 7% to 17% of head and neck osteosarcomas lead to distant disease, most commonly within the first 2 years of treatment, occurring in the lungs and brain.2,4,8,11,35 A more prevalent problem in craniofacial osteosarcoma is local tumor recurrence. This affects the skull more than the maxilla or mandible. In the skull, recurrence rates of 59% to 81% have been reported,6Y8,11 compared to the maxilla (13%Y53%) and mandible (39%Y70%).6 Recurrence of skull osteosarcoma usually occurs within the first year, is extremely difficult to manage, and negatively impacts the quality of life.6,11 Approximately 75% of patients with recurrent disease died within 1 year of initial treatment.9 Progressive local tumor recurrence, with extension to the brain, is the most common cause of death in patients with craniofacial osteosarcoma.11,17,20 Several prognostic indicators have been identified in skull osteosarcoma. The most crucial factor affecting survival is complete surgical resection of the tumor.3Y6,11,14,18,35 Patients with positive surgical margins have significantly higher rates of local recurrence and poorer recurrence-free, disease-specific, and overall survival rates.18,35,36 Adjuvant chemotherapy or radiation therapy is only likely to improve outcomes in patients who have undergone complete surgical resection.19 Anatomic location of the tumor is another important factor in determining the prognosis of osteosarcoma. Among all head and neck osteosarcomas, skull tumors have the worst prognosis.3,5,6,8,13,14,19,20,35 Event-free and overall survival is worse in patients with extragnathic tumors.8,10 This may be because of incomplete resection for anatomic or functional reasons.19 The size of the tumor and histologic response of the tumor to neoadjuvant therapy also impact osteosarcoma prognosis.3,35,37 Patients with larger tumors show poor histologic response.37 The initial size of the tumor is also an independent prognostic factor.37 Tumors less than 4 cm in size show a favorable response to neoadjuvant chemotherapy and have improved recurrence-free survival.35 Other factors found to affect prognosis include age (advancing age correlates with poorer prognosis, unless metastases are present), tumor staging, histologic grade, presence and location of metastases, status of local recurrence, and degree of intracranial involvement.2,3,6,17,35 Shinoda et al17 found that 10 of 13 patients without intracranial involvement of skull osteosarcoma survived more than 1 year and 5 patients survived more than 2 years. In 13 patients with intracranial involvement, however, only 4 of 13 survived more than 1 year, and none survived more than 2 years.17 It is clear that surgical resection with negative margins is the treatment of choice, giving patients the optimal chance of survival.4Y6,11,14,35,36 Two-year survival rates for head and neck osteosarcoma are reported at 48% to 66%.2,5,12,19 Five-year survival rates for osteosarcoma of all areas is 68%,3 but for osteosarcomas of the head and neck, it is only 23% to 63%.2,4,5,12,14,19 The rates plummet for skull osteosarcoma, which has reported 5-year survival rates of 9% to 11%.7,10,12,19 This depends on the extent of intracranial involvement at the time of diagnosis.7,12 Caron et al10 reported an average duration of life after diagnosis of skull osteosarcoma of 20.4 months (median, 5 mo).

2403

Copyright 2011 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.

Brief Clinical Studies

The Journal of Craniofacial Surgery

& Volume 22, Number 6, November 2011

CONCLUSIONS
Primary pediatric osteosarcoma of the skull is a rare entity, with few cases ever reported. The natural course differs enough from other head and neck osteosarcomas that it must be evaluated and described separately. The rarity and severity of the disease, however, preclude analysis in a randomized, prospective, or multi-institutional approach. We must rely on case reports and small studies for insight to the disease process and outcomes of treatment. Diagnosis by CT, MR imaging, or biopsy followed by complete surgical resection with wide margins is the mainstay of treatment. Neoadjuvant chemotherapy and radiation may be effective in high-grade or inoperable tumors or those in which the tumor remains at the margins. Adjuvant therapies have not shown improved survival for skull osteosarcoma, however, which remains extremely poor. To optimize outcomes in craniofacial osteosarcoma patients, we recommend the following caveats: 1. Diagnosis: be aggressive in the quest for definitive diagnosis of firm bony lesions. Even if the patient does not conform to a demographic group typical of the disease process, pursue final diagnosis of the lesion. Assumption of benign lesions can delay appropriate treatment. 2. Imaging: be aware of the wide range of radiographic findings associated with osteosarcoma. If there is any doubt to the benign nature of a lesion, further imaging or biopsy should be performed for definitive diagnosis. 3. Biopsy: although invasive, biopsy of a lesion is ideal for diagnosis of bony malignancies. Care should be taken in the planning of the incision so that it may be incorporated within further tumor resection or reconstruction options. 4. Ossified cephalhematoma: Calcified cephalhematoma is not uncommon in infants after birth trauma; however, in adults, it is rare. Truly ossified cephalhematoma is rare, even in infants, and it has not been reported in adults.38 5. Surgical resection: complete surgical resection with wide margins is critical in optimizing prognosis. a. Resection en bloc should be undertaken for any involvement of the skin, including localized edema. b. Direct inspection for intracranial extension and involvement of dura should be performed at the time of tumor resection. i. If any abnormality in the dura is perceived, dural resection should be performed, followed by dural reconstruction using acellular human dermis or dural graft matrix. 6. Pathology: features of craniofacial osteosarcoma may differ considerably from those in osteosarcoma of appendicular bone. Osteosarcoma must be considered in tumors with unusual histological features in the craniofacial region to avoid potential diagnostic pitfalls.

REFERENCES
1. Ries LAG, Smith MA, Gurney JG, Linet M, Tamra T, Young JL, Bunin GR, eds. Cancer Incidence and Survival Among Children and Adolescents: United States SEER Program 1975Y1995. Bethesda, MD: National Cancer Institute, SEER Program, 1999. NIH Pub. No. 99-4649. Available at: http://www-seer.ims.nci.nih.gov. Accessed January 5, 2011 2. Sturgis EM, Potter BO. Sarcomas of the head and neck region. Curr Opin Oncol 2003;15:239Y252 3. Ottaviani G, Jaffe N. The epidemiology of osteosarcoma. In: Jaffe N, Bruland OS, Bielack S, eds. Pediatric and Adolescent Osteosarcoma. New York, NY: Springer, 2010:3Y13 4. Gadwal SR, Gannon FH, Fanburg-Smith JC, et al. Primary osteosarcoma of the head and neck in pediatric patients. A clinicopathologic study of 22 cases with a review of the literature. Cancer 2001;91:598Y605

5. Goepfert H, Raymond AK, Spires JR, et al. Osteosarcoma of the head and neck. Cancer Bull 1990;42:347Y354 6. Oda D, Bavisotto LM, Schmidt RA, et al. Head and neck osteosarcoma at the University of Washington. Head Neck 1997;19:513Y523 7. Chander B, Ralte AM, Dahiya S, et al. Primary osteosarcoma of the skull. J Neurosurg Sci 2003;47:177Y181 8. Jasnau S, Meyer U, Potratz J, et al. Craniofacial osteosarcoma experience of the cooperative German-Austrian-Swiss osteosarcoma study group. Oral Oncol 2008;44:286Y294 9. Vege DS, Borges AM, Aggrawal K, et al. Osteosarcoma of the craniofacial bones. A clinic-pathological study. J Craniomaxillofac Surg 1991;19:90Y93 10. Caron AS, Hajdu SI, Strong EW. Osteogenic sarcoma of the facial and cranial bones. Am J Surg 1971;122:719Y725 11. Kanazawa R, Yoshida D, Takahashi H, et al. Osteosarcoma arising from the skullVcase report. Neurol Med Chir (Tokyo) 2003;43:88Y91 12. Chou EK, Chang CS, Chen PKT, et al. Long-term management of craniofacial osteosarcoma. J Craniofacial Surg 2009;20:406Y409 13. Nora FE, Unni KK, Pritchard DJ, et al. Osteosarcoma of extragnathic craniofacial bones. Mayo Clin Proc 1983;58:268Y272 14. Chang CS, Bergeron L, Liao CC, et al. Craniofacial reconstruction of primary osteogenic sarcoma of the skull. J Plast Reconstr Aesthet Surg 2010;63:1265Y1268 15. Huvos AG, Sundaresan N, Bretsky SS, et al. Osteogenic sarcoma of the skull: a clinicopathological study of 19 patients. Cancer 1985;56:1214Y1221 16. Daw NC, Mahmoud HH, Meyer WH, et al. Bone sarcomas of the head and neck in children. Cancer 2000;88:2172Y2180 17. Shinoda J, Kimura T, Funakoshi T, et al. Primary osteosarcoma of the skull. J Neurooncol 1993;17:81Y88 18. Ha PK, Eisele DW, Frassica FJ, et al. Osteosarcoma of the head and neck: a review of the Johns Hopkins experience. Laryngoscope 1999;109:964Y969 19. Kassir RR, Rassekh CH, Kinsella JB, et al. Osteosarcoma of the head and neck: meta-analysis of nonrandomized studies. Laryngoscope 1997;107:56Y61 20. Mark RJ, Sercarz JA, Tran L, et al. Osteogenic sarcoma of the head and neck. The UCLA experience. Arch Otolaryngol Head Neck Surg 1991;117:761Y766 21. Leonard A, Craft AW, Moss C, et al. Osteogenic sarcoma in the Rothmund-Thomson syndrome. Med Pediatr Oncol 1996;26: 249Y253 22. Operskalski EA, Preston-Martin S, Henderson BE, et al. A case-control study of osteosarcoma in young persons. Am J Epidemiol 1987;126:118Y126 23. Garland LH. Osteogenic sarcoma of the skull. Radiology 1945;45: 45Y48 24. Goodman MA, McMaster JH. Primary osteosarcoma of the skull. Clin Orthop Relat Res 1976;120:110Y114 25. Thompson JB, Patterson RH, Parsons H. Sarcomas of the calvaria. Surgical experience with 14 patients. J Neurosurg 1970;32:534Y538 26. Salvati M, Ciappetta P, Raco A. Osteosarcomas of the skull. Cancer 1993;71:2210Y2216 27. Dahlin DC, Coventry MB. Osteogenic sarcoma. A study of 600 cases. J Bone Joint Surg 1967;49:101Y110 28. McKenna RJ, Schwinn CP, Higinbotham NL. Osteogenic sarcoma in children. C A Cancer J Clin 1966;16:26Y28 29. Weinfeld MS, Dudley HR. Osteogenic sarcoma: a follow-up study of the ninety-four cases observed at the Massachusetts General Hospital from 1920 to 1960. J Bone Joint Surg 1962;44:269Y276 30. Kosary IZ, Braham J, Bubis JJ. Primary osteogenic sarcoma of the skull. Case report. J Neurosurg 1966;25:87Y89 31. Rowbotham GF. Neoplasms that grow from the bone-forming elements of the skull; a survey of 20 cases. Br J Surg 1957;45:123Y134 32. Dahlin DC, Unni KK. Osteosarcoma of bone and its important recognizable varieties. Am J Surg Pathol 1977;1:61Y72 33. Clark JL, Unni KK, Dahlin DC, et al. Osteosarcoma of the jaw. Cancer 1983;51:2311Y2316

2404

* 2011 Mutaz B. Habal, MD

Copyright 2011 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.

The Journal of Craniofacial Surgery

& Volume 22, Number 6, November 2011

Brief Clinical Studies

34. Hoffman S, Jacoway JR, Krolls SO. Malignant odontogenic tumours of the jaws. In: Intraosseous and Paraosteal Tumors of the Jaws. Atlas of Tumor Pathology. 2nd series. Fascicle 24. Washington, DC: Armed Forces Institute of Pathology, 1987:170Y180 35. Patel SG, Meyers P, Huvos AG, et al. Improved outcomes in patients with osteogenic sarcoma of the head and neck. Cancer 2002;95: 1495Y1503 36. Smeele LE, Kostense PJ, van der Waal I, et al. Effect of chemotherapy on survival of craniofacial osteosarcoma: a systematic review of 201 patients. J Clin Oncol 1997;15:363Y367 37. Kim MS, Lee SY, Cho WH, et al. Initial tumor size predicts histologic response and survival in localized osteosarcoma patients. J Surg Oncol 2008;97:456Y461 38. Gupta PK, Mathew GS, Malik AK, et al. Ossified cephalhematoma. Pediatr Neurosurg 2007;43:492Y497

normal adipose tissue (P = 0.2). Expressions of VEGF-A, hypoxiainducible factor 1>, stromal cellYderived factor 1>, angiopoietin 1 and 2, MMP-2 and -9, VEGF receptors 1 and 2, neuropilin receptors 1 and 2, and CD133 messenger RNA were not elevated compared to control fat (P = 0.1). Endothelial progenitor cells were not present in specimens of infiltrating lipomatosis. Infiltrating lipomatosis does not exhibit elevated angiogenic or vasculogenic factors compared to normal fat; the vasculature is stable. Neovascularization does not seem to play a role in the pathogenesis of this condition. Key Words: Angiogenesis, endothelial, endothelial progenitor cells, facial, infiltrating, lipomatosis, vasculogenesis

Facial Infiltrating Lipomatosis: Expression of Angiogenic and Vasculogenic Factors

Rafael Alejandro Couto, BA, John B. Mulliken, MD, Bonnie L. Padwa, DMD, MD, Aladdin H. Hassanein, MD, MMSc, Gary F. Rogers, MD, JD, MBA, MPH, Ann M. Kulungowski, MD, Arin K. Greene, MD, MMSc Abstract: Facial infiltrating lipomatosis causes diffuse overgrowth of subcutaneous fat, muscle, and bone. Because adipose tissue mass is angiogenesis dependent, the purpose of this study was to determine whether neovascularization is upregulated in this disease. Infiltrating lipomatosis tissue was collected prospectively from the preauricular cheek of 5 patients; neovascularization was compared to normal postauricular adipose. Specimens were analyzed using immunofluorescence for CD31 (microvascular density), >smooth muscle actin (pericyte marker), CD31/Ki67 (proliferating endothelial cells), and CD34/CD133 (endothelial progenitor cells). Quantitative reverse transcriptionYpolymerase chain reaction was used to determine messenger RNA expression of progenitor cells (CD133) and factors that recruit them: vascular endothelial growth factor (VEGF-A), hypoxia-inducible factor 1>, matrix metalloproteinase 9 (MMP-9), and stromal cellYderived factor 1>. Angiopoietin 1 and 2, MMP-2, VEGF receptors, and neuropilin receptors were quantified using quantitative reverse transcription polymerase chain reaction. There was no difference in microvascular density, pericytic density, or endothelial proliferation between infiltrating lipomatosis and
From the Department of Plastic and Oral Surgery, Vascular Anomalies Center, Childrens Hospital Boston, Harvard Medical School, Boston, Massachusetts. Received May 5, 2011. Accepted for publication July 12, 2011. Address correspondence and reprint requests to Arin K. Greene, MD, MMSc, Department of Plastic Surgery, Vascular Anomalies Center, Childrens Hospital Boston, 300 Longwood Ave, Boston, MA 02115; E-mail: arin.greene@childrens.harvard.edu The authors report no conflicts of interest. Copyright * 2011 by Mutaz B. Habal, MD ISSN: 1049-2275 DOI: 10.1097/SCS.0b013e318231fe6c

acial infiltrating lipomatosis is a rare congenital disorder characterized by hemifacial overgrowth and fatty infiltration of adjacent structures.1,2 Males and females are affected equally; there is no predilection between the right and left sides of the face.2 Phenotypic features include osseous and soft tissue hypertrophy, macrodontia, macroglossia, premature tooth eruption, and mucosal neuromas.1,2 Histopathologic examination shows benign, nonencapsulated, mature adipocytes with increased fibrous tissue.1 Previous studies have suggested that the growth of adipose tissue is dependent on vascularization.3,4 Adipocytes are surrounded by an extensive capillary network, and their mass can be regulated by manipulating angiogenesis.3,5 Angiogenic growth factors also stimulate adipocytic differentiation.6,7 The purpose of this study was to determine whether angiogenesis (growth of new blood vessels from preexisting vasculature) or vasculogenesis (de novo formation of new vessels) are upregulated in infiltrating lipomatosis.

METHODS Tissue Specimens

After approval by the Committee on Clinical Investigation at Childrens Hospital Boston, specimens of infiltrating lipomatosis were collected prospectively from patients undergoing a clinically indicated procedure. Subcutaneous tissues were obtained from the cheek through a preauricular incision from an area that had not previously undergone resection. Neovascularization was compared to normal postauricular subcutaneous fat obtained from other patients during otoplasty.

Immunofluorescence
Specimens were fixed in OCT (Sakura Finetek, Torrance, CA), and 5-Km slides were cut. Frozen sections were fixed in methanol for 20 minutes at j20-C and incubated with 5% normal goat serum (Vector Labs, Burlingame, CA) for 1 hour at room temperature. Primary antibodies were applied at 4-C overnight and incubated with Alexa 488Y or Alexa 594Ylabeled secondary antibodies (Invitrogen, Carlsbad, CA) for 45 minutes at room temperature. To define angiogenic or vasculogenic factors in the specimens, the following primary antibodies were used: CD31, an endothelial marker (DAKO, Carpinteria, CA); >-smooth muscle actin, present in pericytes (DAKO); Ki67, a proliferation antigen (DAKO); CD34, located in immature endothelial cells (DAKO); and CD133, a hematopoietic progenitor cell marker (Cell Signaling, Danvers, MA). Quantification of immunofluorescence was determined using image analysis. Pictures were obtained with a Nikon Eclipse E800

* 2011 Mutaz B. Habal, MD

2405

Copyright 2011 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.