Osteosarcoma of the jaws: A 30-year retrospective review

Vol. 90 No. 3 September 2000

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY ORAL AND MAXILLOFACIAL PATHOLOGY

Editor: Alan R. Gould

Osteosarcoma of the jaws: A 30-year retrospective review J. H. Bennett,a G. Thomas,a A. W. Evans,b and P. M. Speight,a London, United Kingdom UNIVERSITY COLLEGE LONDON

Objective. This article reviews osteosarcomas of the jaws referred to the Department of Oral Pathology, Eastman Dental Institute, for histologic diagnosis during the 30 years from 1968 to 1998, to compare the clinical behavior of the tumors, to assess how they differ from the reported characteristics of tumors from other sites, and to report on observations of clinical and diagnostic significance. Study design. The clinical, radiographic, and histopathologic records of 25 patients were obtained for retrospective review. Supportive and follow-up clinical and histopathologic material was obtained from the referring clinicians. Results. The mean age of presentation of the primary lesions was 36.9 years (range, 10-87 years) with a slight female predilection. The most common presenting features were swelling, pain, ulceration, or a neurologic deficit. The radiographic appearance of the lesions was mixed with areas of radiopacity and radiolucency. Histologically, the lesions ranged from those dominated by immature bony trabeculae separated by a cytologically bland stroma to high-grade, cellular lesions with regions of marked atypia and mitotic activity. Most had areas of chondroid formation, in addition to neoplastic osteoid. The main complication was local recurrence. Metastasis was rare and occurred as a solitary process or at a late stage in the disease progression. This was in contrast to lesions metastatic to the jaws, which were higher grade in appearance and had metastasized widely, early in the disease process. Primary osteosarcoma occurring in patients with a history of radiotherapy was typically more aggressive. Conclusion. Primary osteosarcomas of the jaws are a group of lesions distinct from those occurring in the long bones. Osteosarcomas of the jaw arising in a former radiation field may be a discrete group of lesions with a more aggressive behavior pattern. (Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2000;90:323-33)

The term osteosarcoma refers to a heterogeneous group of primary malignant neoplasms affecting boneforming or mesenchymal tissue that have histopathologic evidence of osteogenic differentiation. The World Health Organization (WHO) recognizes several variants that differ in location, clinical behavior, and degree of cytologic atypia.1,2 The most frequently occurring is the conventional or classical osteosarcoma, which arises centrally within the bone and can be subdivided into osteoblastic and chondroblastic variants, depending on the predominant cell type.3 Conventional osteosarcomas have to be distinguished from the centrally occurring low-grade medullary osteosarcoma and parosteal osteosarcoma, which arise subperiosteally. Classical osteosarcoma accounts for approxaDepartment of Oral Pathology, Eastman Dental Institute, University College London. bDepartment of Maxillofacial Surgery, Eastman Dental Institute, University College London. Received for publication Oct 18, 1999; accepted for publication Mar 28, 2000. Copyright © 2000 by Mosby, Inc. 1079-2104/2000/$12.00 + 0 7/14/108274 doi:10.1067/moe.2000.108274

Fig 1. Histogram showing age on presentation for each case included in this study.

imately 20% of primary bone tumors.3,4 It has a bimodal age distribution, with a major peak in the second decade and a somewhat smaller peak after the age of 50 years.5 The most common sites are the long 323

324 Bennett et al

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY September 2000

Table I. Summary of clinical details from 25 patients included in this study No.

Age/sex

Site/presenting oral features

1 2 3 4 5 6 7 8 9 10

20/M 80/F 16/M 17/F 36/F 37/F 35/F 26/F 34/F 74/M

Buccal swelling left maxilla canine/premolar region for 3 mo Body right mandible; neurologic deficit inferior alveolar/mental nerves for 4 mo Swelling right mandible Angle left mandible bony swelling in lesion previously diagnosed as an ossifying fibroma Swelling right mandible for 6 mo Swelling body left mandible; buccal and lingual expansion, mucosal ulceration for <1 y Right mandible; condyle, rumus, and body for <1 y Bony swelling right maxilla, buccally, canine region Recurrent swelling body left mandible Swelling body left mandible; meurologic deficit mential nerve for 2 mo

11 12 13 14

26/F 12/F 63/F 14/F

Swelling left mandible Swelling body right mandible Left mandible Swelling right maxilla for 6 mo previously diagnosed as fibrous dysplasia

15 16

32/M 32/F

Rapidly growing swelling, body left mandible, premolar, and molar region for 9 mo Radiolucency left mandible adjacent to mesial root of the 3rd molar; incidental finding

17 18

10/F 45/M

Swelling maxillary premolar region Painless firm swelling right mandibular premolar region for 2 mo

19 20 21 22 23 24 25

19/M 41/M 87/F 30/F 45/M 49/M 43/M

Gingival swelling left mandible premolar region for 6 wk Gingival swelling left mandibular incisors Swelling right maxillary tuberosity Bony swelling right mandibular molar region, neurologic deficit mental nerve for 3 mo Left maxillary swelling for 2 mo Swelling right mandibular premolar region 2 mo; parasthesia right mental nerve for 1 y Periapical radiolucency of lower 2nd permolar and 1st molar region, delayed healing 1 wk after extraction

M, Male; F, female.

bones, particularly in adolescent patients, with the majority occurring in the distal femoral metaphysis, proximal tibia, and humeral metaphysis. In older patients, this predilection for the appendicular skeleton is less marked, and lesions in the axial skeleton and membranous bones account for up to 40% of cases.4 Lesions are slightly more common in men. In an extensive series of 1649 patients, 133 were over the age of 50 years. In this group, the tumor developed against a background of a preexisting medical condition in 51 cases, most frequently Paget’s disease of bone or a history of radiotherapy.3 Some regard the latter as a distinct group of tumors; indeed, high-grade osteosarcomas developing in a radiation field are specifically mentioned by the WHO. Other rare variants of osteosarcoma listed by the WHO include telangiectatic, round cell, periosteal, and high-grade surface osteosarcomas. Lesions of the jaws are uncommon, and most are regarded as being variants of classical osteosarcoma, representing about 6% of the total.3,6 However, their bimodal distribution differs from those in the appendicular skeleton, with the first peak occurring somewhat later, in the third decade.7

Tumors of the appendicular skeleton occurring in young patients (and those arising within radiation fields) are aggressive lesions that may metastasize widely at an early stage8-11 with a 5-year survival rate of about 41%. However, combined chemotherapeutic and surgical management of lesions that have not metastasized has improved this to 70%.4 In contrast, osteosarcoma of the jaws, with the exception of those tumors occurring in radiation fields, often shows little cellular atypia and late metastasis. Control of local spread remains the main clinical challenge. Furthermore, the absence of cytologic features of malignancy may lead to diagnostic difficulty in distinguishing osteosarcoma from benign or reactive bony lesions. Low-grade medullary osteosarcomas and parosteal (juxtacortical) osteosarcomas are low-grade lesions showing mild atypia and a relatively indolent growth pattern. The peak incidence is in the third or fourth decades, and in this respect, they resemble osteosarcomas of the jaws rather than osteosarcomas occurring in the long bones. Examples might rarely occur in the jaws. If recognized and treated early, these respond well to surgery, but if untreated or treated inappropri-

Bennett et al 325

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 90, Number 3

Predisposing factors None Previous radiotherapy for carcinoma of the tongue Widespread metastasis 1 degree osteosarcoma left humerus None None None History of previous radiotherapy None Previous giant cell granuloma treated with radiotherapy None

Diagnosis

Widespread metastasis 1 degree osteosarcome left femur Rhabdomyosarcome a 2 y right cheek, treated with radiotherapy Not recorded None

Osteoblastic osteosarcome Osteoblastic osteosarcome Metastatic osteoblastic osteosarcome Osteoblastic osteosarcome Chondroblastic osteosarcome Chondroblastic osteosarcome Osteoblastic osteosarcome Low-grade central osteosarcome Osteoblastic osteosarcome Osteoblastic osteosarcome (high grade) in later biopsy specimens Osteoclast-rich osteoblastic sarcoma Osteoblastic osteosarcome Osteoblastic osteosarcome Osteoblastic osteosarcome

None None

Osteoblastic osteosarcome Osteoblastic osteosarcome

Widespread metastasis, primary osteosarcoma left femur None None – History of oral squamous cell carcinoma treated with radiotherapy None None None

Osteoblastic osteosarcome Low-grade osteoarcoma/periosteal osteosarcoma Chondroblastic osteosarcoma Chondroblastic osteosarcoma Osteoblastic osteosarcome Osteoblastic osteosarcome Osteoblastic osteosarcome Osteoblastic osteosarcome

None

Osteoblastic osteosarcoma

ately, they may undergo transformation to a highgrade lesion. The aim of this article is to review 25 cases of osteosarcoma of the jaws that were referred to the Department of Oral Pathology for histologic diagnosis during the 30-year period from 1968 to 1998, to compare their clinical behavior, to assess how their characteristics differ from the reported characteristics of tumors from other sites, and to report on observations of clinical and diagnostic significance.

PATIENTS AND METHODS The clinical, radiographic, and histopathologic records of patients referred to the Department of Oral Pathology, Eastman Dental Institute, over the 30-year period from 1968 to 1998 were obtained for retrospective review. Supportive and follow-up clinical and histologic material was obtained whenever possible from the referring clinicians. Formalin-fixed, paraffin-embedded 6 µm sections were stained with hematoxylin and eosin or periodic acid-Schiff and compared with the original pathology reports. If the archival slides were of insufficient quality, the original blocks were retrieved and new

Follow-up Alive 2 y Local recurrence; died within 2 y Died with 2 y Alive 5 y Lost to follow-up Lost to follow-up Died after 1 y Alive 10 y Local recurrence; alive 10 y Local recurrence; late-stage pulmonary metasasis; died within 2 y Died within 2 y Died within 2 y Lost to follow-up Solitary pulmonary metastasis after 2 y; alive and well after 10 y Lost to follow-up Local recurrence; late-stage pulmonary metasasis, died within 2 y Died within 2 y Lost to follow-up Local recurrence; died within 2 y Died within 5 y Died within 2 y Lost to follow-up Lost to follow-up Lost to follow-up Lost to follow-up

sections prepared. Many sections were of decalcified specimens. All sections were reviewed by an experienced oral pathologist, and in most cases a second opinion had been sought from a specialist bone pathologist (see acknowledgments).

RESULTS In the 30-year period between 1968 and 1998, a total of 28 cases were entered into the files of the Department of Oral Pathology, Eastman Dental Institute. Of these, 3 were inadequately documented and, therefore, not included in this study. Of the 25 remaining cases, 22 were primary osteosarcomas of the jaws and 3 were metastatic lesions. Age and sex The mean age at first presentation of the primary oral tumors was 39.6 years. The age distribution is summarized in Fig 1. A slight female predominance (13 cases) was evident. The mean age of presentation of maxillary tumors was 39.1 years (range, 10 to 87 years), whereas that of mandibular tumors was 39.8 years (range, 12 to 80 years). The 3 lesions metastatic to the jaws occurred in patients aged 10, 16, and 26 years.

326 Bennett et al

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY September 2000

A

B Fig 2. Case 4: Frontal photograph showing prominent left sided facial swelling (A). Pan-oral radiograph demonstrates radiopaque tumor of left mandibular angle (B). (Figure 2 continued on next page.)

Site and presenting features Eight tumors presented in the maxilla and 17 in the mandible. Of the maxillary tumors, 3 patients were men and 5 women. Of the mandibular tumors, 8 occurred in men and 9 in women. In the maxilla there was a predilection for the canine-premolar area, whereas in the mandible, the body or ascending ramus was the most common site. For the primary jaw tumors, the most common presenting feature, affecting 19 of 22 cases, was swelling with expansion of the buccal or lingual plates (Table I; Fig 2). In case 20, a marked gingival swelling occurred with periodontal radiolucency (Fig 3). The mean duration of the swelling from onset to diagnosis was 4.6 months (range, 0.5-12 months), with most

patients presenting after 2 to 6 months. In 2 cases (cases 5 and 10), there was a long history of treatment for preexisting jaw lesions (ossifying fibroma and giant cell granuloma, respectively) and it was not possible to separate the approximate date of diagnosis of the osteosarcoma from the date of diagnosis of these lesions. In a further case (case 15), the tumor was initially diagnosed as a fibrous dysplasia. A neurologic deficit was recorded in the distribution of the mental, inferior alveolar, or the infraorbital nerves in 5 of 22 patients. In cases 16 and 25, the only evidence of tumor was radiolucency around a premolar or molar tooth and, in the latter, this led to delayed healing after a dental extraction. Perioral or cervical lymphadenopathy was only recorded in 2 cases,

Bennett et al 327

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 90, Number 3

D

C

E Fig 2, continued. Case 4: Computer-aided axial tomography shows extension of lesion buccally and medially. (C). Histologically, the tumor was composed of areas of immature bone and osteoid separated by cytologically bland fibrous stroma (D and E). In several areas, tumor was locally invasive. Bar: D and E-80 µm.

and it was not clear from the notes whether this was reactive or whether it represented metastatic tumor. The 3 metastatic lesions presented as swellings, 2 in the mandible (cases 3 and 11) and 1 in the maxilla (case 17). These patients had widely metastatic osteosarcoma of the long bones and the jaw involvement was noted within a year of diagnosis of the primary lesions.

Predisposing factors Of the 22 patients with primary osteosarcoma of the jaws, 5 had a history of radiotherapy, 2 for squamous cell carcinoma of the tongue (cases 2 and 21), 1 for rhabdomyosarcoma of cheek (case 12), 1 for a giant cell granuloma of the jaws (case 9), and 1 for treatment

of the primary osteosarcoma at another center (case 7). Additionally, in 1 of these cases, there was an unconfirmed history of Paget’s disease of bone (case 21). Fifteen patients had no history of any predisposing condition, and for 2 others, details of the medical history were not available. None of the patients had a clear history of trauma.

Management and follow-up Treatment and follow-up information was available for 13 of the 22 patients with primary jaw lesions, although in several cases, these data were incomplete. Surgery, aimed at excision of the tumor, was the treatment of choice in 18 of the primary jaw lesions. One of the remainder was a woman of 87 for whom extensive

328 Bennett et al

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY September 2000

A

B Fig 3. Case 20: Intraoral photograph showing labial gingival swelling in lower mandibular incisor region (A). Intraoral radiographs show bone loss around left lower incisors and canine teeth with evidence of root resorption (B). (Figure 3 continued on next page.)

radical surgery would have been inappropriate. Of those treated surgically, 7 died within 2 years of diagnosis, 2 survived between 2 and 5 years, and 4 were alive after 5 years. The predominant complication in the 9 patients who survived for less than 5 years was local spread. In only 2 cases (cases 10 and 16), pulmonary metastasis was a late-stage complication. With regard to predisposing factors, 4 of the 5 patients who had a history of radiotherapy survived for less than 5 years. Of the 4 patients surviving for more than 5 years, 3 were alive more than 10 years after diagnosis, whereas the fourth is still alive at 7 years. One patient, case 8,

was diagnosed as having a low-grade lesion, possibly a parosteal osteosarcoma. Case 16 developed a solitary pulmonary metastasis within 2 years of diagnosis, treated by metastectomy with no recurrence. Three patients remain tumor-free. However, the fourth (case 9) had a lesion within a radiotherapy field with extensive local recurrence. She was never completely tumor-free and underwent repeated surgical intervention, leading to considerable disability and a requirement for full-time nursing care. No follow-up information was available for 9 patients. The 3 patients with metastases to the jaws underwent

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 90, Number 3

Bennett et al 329

D

C

E Fig 3, continued. Case 20: After resection, macroscopic examination showed locally invasive lesion confined to alveolar bone (C). This was composed of areas of neoplastic bone, osteoid, and chondroid separated by a cellular stroma with cellular atypia and pleomorphism (D and E). Bar: C-5 mm; D-80 µm; E-30 µm.

surgical debulking procedures, but none survived longer than 2 years.

Histology and radiology The 22 primary lesions showed considerable tissue and cytologic heterogeneity. Nevertheless, 20 had been reported as classical osteosarcoma, a diagnosis confirmed in most instances by the United Kingdom Bone Tumor Panel or a specialist bone pathologist. These tumors showed varying degrees of cellularity associated with different proportions of extracellular

matrix. In all tumors there was evidence of neoplastic osteoid formation, although in several, the newly formed bone was well differentiated and not readily distinguishable from reactive bone. The majority of tumors showed some evidence of chondroid formation, but only in 4 cases was this sufficient to justify the designation chondroblastic osteosarcoma. Some tumors were well-differentiated and showed little cellular atypia, with plump, cuboidal osteoblasts adjacent to newly formed osteoid (Fig 2). In places, this was poorly mineralized, and in others it was organized

330 Bennett et al

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY September 2000

A

C

B

Fig 4. Photomicrographs of case 7 with areas of chondroid and osteoid within stroma showing some cellular atypia (A). In others, cytologically cellular stroma predominated with occasional mitotic figures and bizarre giant cells (B). Regions were noted with cells densely packed, fibroblastic, and arranged in a ‘herringbone’ pattern (C). Bar: A-80 µm; B-80 µm; C-125 µm.

into parallel or randomly arranged trabecular patterns. Others (cases 2, 7, 9, 12, 14, 20, and 21) had features of a high-grade lesion with regions of dysplastic osteoid or chondroid (Fig 3), marked cellularity, and pleomorphism (Fig 4). Areas of densely packed fibroblastic cells arranged in a herringbone pattern were occasionally seen (Fig 4). An occasional feature was the presence of bizarre multinucleate cells (Fig 3). Interestingly, this latter group included all the patients with a history of radiotherapy. One further case (case 10) presented

with a cytologically low-grade tumor that developed over a period of 2 years into an aggressive, high-grade lesion. Of the classical osteosarcomas, 16 were classified as osteoblastic osteosarcoma. Two cases were well-differentiated, low-grade lesions. One was of medullary origin, whereas the other was possibly a parosteal osteosarcoma. However, it was not possible to exclude an endosteal origin to confirm the diagnosis. The 3 metastatic lesions to the jaws were all cellular,

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 90, Number 3

Bennett et al 331

high-grade lesions with many irregular mitoses, which arose from primary tumors in the long bones and were part of widespread metastatic dissemination of the primary tumor. Conversely, the few metastases from primary jaw lesions, typified by case 14, were solitary lesions. In cases 10 and 16, pulmonary metastases were observed on radiographic examination as very late events in the disease process when comprehensive investigation was not appropriate. Histology was, therefore, not available.

These observations have suggested that an oxygen or nutrient gradient may help to determine whether a cell will become osteoblastic or chondroblastic. Factors such as local oxygen tension and angiogenesis at sites within a tumor may contribute to localized areas of chondroid differentiation. Osteosarcomas of the jaws are generally regarded as variants of classical osteosarcoma. Nevertheless, in common with the earlier studies, several of the lesions reported here formed a group that showed little cytologic atypia and, unlike long bone lesions, a lower rate of metastasis. Furthermore, these tumors occurred between the third and fifth decades, somewhat later than classical osteosarcoma. Several patients survived for over 5 years. The major complication was local relapse, with metastasis occurring as a solitary or a late stage event. Tumors metastatic to the jaws were, by comparison, widely metastatic high-grade lesions showing marked atypia and an aggressive behavior pattern. These observations support 2 earlier studies in which the mean time interval from treatment of a primary osteosarcoma of the jaws to recognition of a metastatic lesion was 20 and 29 months, respectively, as opposed to 6 months for lesions in the long bones.14,15 In terms of behavior, therefore, the primary jaw lesions have a characteristic behavior pattern and we endorse the view that they are a distinct group of primary osteosarcomas.1,3 However, it is important to note that among the cases presented here, one tumor, that presented as a low-grade lesion, underwent dedifferentiation to a high-grade, aggressive lesion. A further group of lesions highlighted by this study were high-grade primary osteosarcomas that were associated with a much poorer survival rate. These showed considerable cytologic atypia, a marked increase in mitotic rate and a tendency to occur over a wider age range. An important risk factor for this group was a history of radiotherapy that included the 4 patients in whom tumors developed in tissue irradiated up to 10 years previously. Several authors have commented on the role of radiotherapy in the etiology of osteosarcoma20; indeed, the WHO lists it as a category of osteosarcoma in its own right.2 Our observations endorse this view. Because many osteosarcomas of the jaws appear cytologically unremarkable, care must be taken to separate them from benign or reactive lesions, such as fibrous dysplasia21,22 or osteoblastoma.23,24 Two of our cases were diagnosed as fibrous dysplasia and ossifying fibroma before the true nature of the condition became apparent. Conversely, in poorly differentiated lesions, the tumor may be cellular, and malignant osteoid can be difficult to spot. If a spindled appearance dominates, the lesion will have to be

DISCUSSION Because osteosarcomas of the jaws are rare, small retrospective studies represent the sole opportunity for collation and comparison of several cases. Those reported here follow a trend established in earlier studies12-15 and highlight several important points. Lesions occurred over a wide age range, with a peak in the fourth decade, somewhat later than those previously reported. In common with an earlier study, we found a mandibular predominance,14 whereas others have reported a more even distribution between mandibular and maxillary lesions. Although diagnosis was based on the recognition of osteoid production by malignant cells, individual lesions were heterogeneous with fibroblastic, osteoblastic, and chondroblastic elements, and could be arbitrarily designated as osteoblastic or chondroblastic on the basis of the type of extracellular matrix present. Areas of well formed chondroid were found in most of the tumors reported here, and in common with other authors, patches of chondroid differentiation were regarded as consistent with the designation osteoblastic osteosarcoma.3,12 Of the lesions reported in this study, 18% were designated chondroblastic osteosarcoma. Other studies have reported a much higher percentage,14 which may reflect the lack of a clear consensus when defining the osteoblastic and chondroblastic variants. Designation as osteoblastic or chondroblastic osteosarcoma may, however, be of clinical significance because the latter reportedly has a marginally better prognosis.5,16 From a biologic perspective, the cells of origin of an osteosarcoma are by no means well defined, and it is likely that they are not osteoblasts, but undifferentiated mesenchymal precursors with osteogenic potential. Some cellular heterogeneity within tumor clones, therefore, would not be surprising because fibroblasts, chondroblasts, and osteoblasts are believed to share a common lineage.17 Furthermore, both genetic and locally active epigenetic factors may influence whether a particular cell progresses down an osteogenic or chondrogenic pathway. When fibroblastic cells derived from the marrow stroma were implanted in vivo in diffusion chambers, a bone-like tissue formed peripherally, with chondroid or fibroblastic areas centrally.18,19

332 Bennett et al

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY September 2000

distinguished from other fibroblastic lesions, or if there is a predominance of small round cells, tumors such as Ewing’s sarcoma will need to be excluded. Lesions are frequently extensive, with different histopathologic patterns occurring at various sites. Therefore, all lesions must be sampled extensively and the histopathologic appearance considered together with the radiographs in excluding an invasive behavior pattern. To date, there are few special investigations that can be used on histopathologic material to assist in the identification of malignancy. It has been suggested that bone matrix proteins may assist in the recognition of malignant osteoid in the differential diagnosis of osteosarcoma. Osteocalcin, for example, is a bonespecific protein that may be useful in distinguishing osteosarcomas from malignant fibrous histiocytoma.25 However, it is a late marker of osteoblastic differentiation,26 is expressed with difficulty in some osteosarcoma-derived cell lines,27,28 and might therefore be of limited use to detect poorly differentiated cells of the osteoblast lineage. Both collagen type I and osteonectin have been recognized in tumor osteoid,29,30 but it has been difficult to use these data because these proteins are not specific to malignant tissue. Recently cbfa 1,31-33 a gene encoding an intranuclear osteocalcin promoter, has been identified that appears to be specific to cells of the osteoblastic lineage and might have a role in the differential diagnosis of osteosarcoma. However, its use in the differential diagnosis of malignant tumors of bone has yet to be fully evaluated. A recurrent problem in the use of special investigations in the diagnosis of bone tumors is the requirement for decalcification of mineralized tissue. This may be overcome to some extent by the use of cytologic imprint preparations. The detection of alkaline phosphatase activity in imprint preparations obtained from the cut surface of osteosarcomas before fixation is regarded as diagnostic of osteosarcoma if used in combination with radiographs.34 However, a negative result might not exclude the diagnosis. The morphologic and behavioral heterogeneity observed in osteosarcomas and emphasized in this study are consequent to molecular changes within the tumor cells. These are beginning to be characterized but, to date, studies have been restricted to lesions of the appendicular skeleton, and no specific data are available for osteosarcoma of the jaws. Nevertheless, they are likely to highlight points that may reflect or contrast with events occurring in oral lesions and thus aid understanding of the biologic basis of the heterogeneity of this group of lesions. Cytogenetic abnormalities, such as ring chromosomes, have been reported in

parosteal osteosarcoma,35 but the search for specific deletions or gene rearrangements has not, so far, been fruitful, only showing changes that are indicative of the malignant process in general rather than specific to osteosarcoma. Chromosomal alterations in the p53 and retinoblastoma genes, localized to 17p13 and 13q14 respectively, are common, and patients with LiFraumeni syndrome or hereditary retinoblastoma have an increased risk of developing osteosarcoma.36,37 Although not restricted to osteosarcoma, disturbances of the expression pattern of these proteins may have some prognostic value as they may be indicative of a higher grade, dedifferentiated lesion.4 Similarly, lowgrade and high-grade lesions may differ in expression of other proteins, such the transmembrane glycoprotein C-erb-B2.38 Transgenic mice overexpressing c-fos have a tendency to develop osteosarcomas,38 and disturbances of several other oncogenes, including Cfos, c-myc, and N-myc, but not the ras genes, have been reported.39 Anomalous expression of members of the cadherin family is seen in several different tumors, including those of the oral cavity.40 A recent report has highlighted differences in the expression pattern of cadherin-11 (Ob-cadherin) in normal tissue and highly malignant primary or metastatic osteosarcomas arising in the long bones.41 Little is known about their expression in lesions of the jaws. Nevertheless, identification of variations in the expression pattern of cadherin-11 may help to explain the differences between jaw lesions and those of the appendicular skeleton. In summary, osteosarcomas of the jaws are less aggressive than those of the long bones. They metastasize later and present in a slightly older age group. We can conclude, therefore, that there is sufficient evidence to consider primary osteosarcoma of the jaws, in the absence of any predisposing history, a distinct entity, separate from osteosarcoma in the long bones. We thank Dr Jeremy Whelan, Department of Medical Oncology, University College London, Dr N. A. Ratcliffe, Department of Histopathology, St Peters Hospital, Chertsey, Surrey, and Professor Malcolm Harris, Department of Oral and Maxillofacial Surgery, the Eastman Dental Institute. We also acknowledge the contribution made by Professor Ivor Kramer and Dr Khee Wee Lee of the Department of Oral Pathology, Eastman Dental Institute, Dr Jean Pringle of the Department of Morbid Anatomy, Royal National Orthopaedic Hospital, Stanmore, Middlesex, UK, and Professor Peter Revell of the Department of Histopathology, the Royal Free Hospital, Pond Street, London NW3 1DU, UK, who, over the years, have reviewed and reported on the histopathology of the lesions reported here. REFERENCES 1. Schajowicz F. Histological typing of tumors of bone; World

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 90, Number 3

Bennett et al 333

Health Organization international histological classification of tumors. Berlin: Springer-Verlag; 1993. Schajowicz F, Sissons HA, Sobin LH. (1995) The World Health Organization’s histologic classification of bone tumors: a commentary on the second edition. Cancer 1995:75;1208-14. Unni KK. Dahlin’s bone tumors, general aspects and data on 11,087 cases. 5th ed. Philadelphia: Lippincot-Raven; 1996. Dorfman HD, Czerniak B. Bone tumors. St Louis: Mosby; 1998. Dorfman HD, Czerniak B. Bone cancers. Cancer 1995;75:203-10. Coley B. Neoplasms of bone. New York: Paul B. Hoeber Inc; 1960. Huvos AG. Bone tumors: diagnosis, treatment and prognosis. Philadelphia: WB Saunders; 1979. Bentzen SM, Poulsen HS, Kaae S, Jensen OM, Johansen H, Mouridsen HT, et al. Prognostic factors in osteosarcomas: a regression analysis. Cancer 1993;62:707-16. Brostrom LA, Strander H, Nilsonne U. Survival in osteosarcoma in relation to tumor size and location. Clin Orthop 1982;167:2504. Glasser DB, Lane JM, Huvos AG, Marcove RC, Rosen G. Survival, prognosis and therapeutic response in osteogenic sarcoma. Cancer 1992;69:698-708. Taylor WF, Ivins JC, Unni KK, Beabout JW, Golenzer HJ, Black LE. Prognostic variables in osteosarcoma: a multi-institutional study. J Natl Cancer Inst 1989:81;21-30. Clarke JL, Unni KK, Dahlin DC, Devine KD. Osteosarcoma of the jaw. Cancer 1983;51:2311-6. Caron AS, Hajdu SI, Strong EW. Osteogenic sarcoma of the facial and cranial bones. Am J Surg 1971;122:719-25. Garrington GE, Scofield HH, Cornyn J, Hooker SP. Osteosarcoma of the jaws. Cancer 1966;20:377-90. Slootweg PJ, Muller H. Osteosarcoma of the jaw bones. J Max Fac Surg 1985:13;158-66. Saito K, Unni KK, Wallan PC, Lund BA. Chondrosarcoma of the jaws and facial bones. Cancer 1995:76;1550-8. Owen ME. Lineage of osteogenic cells and their relationship to the stromal system. In: Peck WA, editor. Bone and mineral research. Volume III. Amsterdam: Elsevier; 1985. p. 1-25. Ashton BA, Allen TD, Howlett CR, Eagesom CC, Hattori A, Owen M. Formation of bone and cartilage by marrow stromal cells in diffusion chambers in vivo. Clin Orthop 1980;151:294-307. Gundle R, Joyner CJ, Triffitt JT. Human bone tissue formation in diffusion chamber culture in vivo by bone derived cells and marrow stromal fibroblastic cells. Bone 1995;16:597-601. Sim FH, Cupps RE, Dahlin DC, Ivins JC. Postirradiation sarcoma of bone. J Bone and Joint Surg 1972:54-A;1479-89. Present D, Bertoni F, Enneking WF. Osteosarcoma of the mandible arising in fibrous dysplasia. Clin Orthop 1986:204;238-44. Franceschina MJ, Hankin RC, Irwin RB. Low-grade central osteosarcoma resembling fibrous dysplasia. A report of two cases. Am J Orthop 1997;26:432-40. Lucas DR, Unni K, McLeod RA, O’Connor MI, Sim FH. Osteoblastoma: clinicopathologic study of 306 cases. Human Pathology 1994:25;117-34. Bertoni F, Unni KK, McLeod RA, Dahlin DC. Osteosarcoma resembling osteoblastoma. Cancer 1985;55:416-26. Ushigome S, Shimoda T, Fukunaga M, Takakuwa T, Nakajima H. Immunocytochemical aspects of the differential diagnosis of osteosarcoma and malignant fibrous histiocytoma. Surg Pathol 1988:1;347-57.

26. Stein GS, Lian JB. Molecular mechanisms mediating developmental expression of genes in osteoblasts. In: Noda M, editor. Cellular and molecular pathology of bone. San Diego: Academic Press; 1993. p. 48-97. 27. Clover J, Gowen M. Are MG 63 and HOS TE85 human osteosarcoma cell lines representative models of the osteoblastic phenotype? Bone 1994;15:585-91. 28. Lajeunesse D, Frondoza C, Schoffield B, Sactor B. Osteocalcin secretion by the human osteosarcoma cell line MG-63. J Bone Min Res 1990;5:915-22. 29. Ueda Y, Nakanishi I. Immunohistochmical and biochemical studies on collagenous proteins of human osteosarcoma. Virchows Arch (B) 1989:58;79-88. 30. Schultz A, Jundt G, Berghauser KH, Gehron-Robey P, Termine JD. Immunohistochemical study of osteonectin in various types of osteosarcoma. Am J Pathol 1988:132;233-8. 31. Ducy P, Zhang R, Geoffroy V, Ridall A, Karsenty G. Osf2/Cbfa1: a transcriptional activator of osteoblast differentiation. Cell 1997;89:747-54. 32. Mundlos S, Otto F, Mundlos C, Mulliken JB, Aylsworth AS, Albright S, et al. Mutations involving the transcription factor CBFA1 cause cleidocranial dysplasia. Cell 1997:89;773-9. 33. Otto F, Thornell AP, Crompton T, Denzel A, Gilmour KC, Rosewell IR, et al. Cbfa1, a candidate gene for cleidocranial dysplasia syndrome is essential for osteoblast differentiation and bone development. Cell 1997:89;756-71. 34. Pringle J. Osteosarcoma: the experiences of a specialist unit. Cur Diag Pathol 1996:3;127-36. 35. Bridge JA, Nelson M, McComb E, McGuire MH, Rosenthal H, Vergara G, et al. Cytogenetic findings in 73 osteosarcoma specimens and a review of the literature. Cancer Genet Cytogenet 1997;95:74-87. 36. Malkin D, Li FP, Strong LC, Fraumeni JF, Nelson CE, Kim DH, et al. Germ line p53 mutations in a familial syndrome of breast cancer, sarcomas and other neoplasms. Science 1990:250;1233-8. 37. Onda M, Matsuda S, Higaki S, Iijima T, Fukashima J-I, Yokohara T, et al. ErbB-2 expression is correlated with poor prognosis for patients with osteosarcoma. Cancer 1996:77;71-8. 38. Grigoriadis AE, Schellander K, Wang Z-Q, Wagner EF. Osteoblasts are target cells for transformation in c-fos transgenic mice. J Cell Biol 1993:122;685-701. 39. Pompetti F, Rizzo P, Simon RM, Freidilin B, Mew DJ, Pass HI, et al. Oncogene alterations in primary, recurrent and metastatic human bone tumors. J Cell Biochem 1996:63;37-50. 40. Downer CS, Speight PM. E-cadherin expression in normal, hyperplastic and malignant oral epithelium. Oral Ankle Eur J Cancer 1993;29B:303-5. 41. Kashima T, Kawaguchi J, Takeshita S, Kuroda M, Takanashi M, Horiuchi H, et al. Anomalous cadherin expression in osteosarcoma: possible relationships to metastasis and morphogenesis. Am J Pathol 1999;155:1549-55.

2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

Reprint requests: Dr Jon Bennett Department of Oral Pathology Eastman Dental Institute University College London