Low-grade central osteosarcoma is a rare subtype of osteosarcoma that was first described by Unni et al. 1 in 1977. In contrast, fibrous dysplasia is not a true neoplasm, but rather a developmental abnormality. It represents one of the more commonly encountered benign lesions of bone. Despite these differences, the main feature that brings these two tumors together is their remarkably similar histologic appearance. In addition, there can be overlap in their radiographic features. This can cause diagnostic difficulties that may lead to misdiagnosis and inappropriate treatment. The problem is compounded by the fact that, because low-grade central osteosarcoma is so uncommon, most pathologists and radiologists rarely see it. Therefore, it is difficult to recognize.
When compared to conventional osteosarcoma, low-grade central osteosarcoma has a relatively good prognosis. Nevertheless, it is a malignant tumor with the potential, if treated with inadequate surgical margins, for local recurrence, dedifferentiation, and metastasis. This differs from the benign biologic behavior of fibrous dysplasia, which can be successfully treated with conservative surgery. These differences emphasize the importance of a correct diagnosis that will result in optimal patient care. This review article covers the clinicopathologic features of these two tumors with a focus on those that can be used to make the distinction.
Fibrous dysplasia is more common than low-grade central osteosarcoma. It represented 25% of the benign bone tumors and 7% of all biopsied bone tumors in the Mayo Clinic files. 2 The overall incidence of low-grade central osteosarcoma is less than 1% of primary bone tumors and only 1% to 2% of all osteosarcomas. 2,3 Therefore, when faced with this differential diagnosis, remember that the lesion in question is more likely to be fibrous dysplasia. Men and women are equally affected by both tumors. They also share a similar age distribution, with a peak incidence in the second and third decades of life.
Low-grade central osteosarcoma is most commonly located in the long bones with a distinct predilection for the distal femur and proximal tibia. In 1990, Kurt et al. 4 reported on a series featuring 80 patients with low-grade central osteosarcoma. It is the largest series reported in the literature. 4 Sixty-six (81%) of the 80 tumors were located in the long bones and 43 of these 66 were around the knee within the distal femur or proximal tibia. Only 12 (15%) of the lesions were located in flat bones. In the second largest series, from a different institution, Bertoni et al. 5 described 10 cases of low-grade central osteosarcoma. None of the tumors in their series were located in the flat bones, a finding similar to other case reports in the literature. 6–8 This is in contrast to the skeletal distribution for fibrous dysplasia, where the flat bones are commonly affected. The skull, jawbones, and ribs are favored skeletal sites for fibrous dysplasia. Fifty-five percent of the fibrous dysplasias in the Mayo Clinic files were located in these three sites combined. Given these differences in skeletal distribution, be reluctant to make a diagnosis of central low-grade osteosarcoma if the tumor involves a flat bone or a site in a long bone other than the distal femur or proximal tibia.
Because fibrous dysplasia and central low-grade osteosarcoma are so similar histologically, the radiographic features are an extremely important part of the diagnosis. One of the most helpful points to remember is that, though the radiologic features of central low-grade osteosarcoma are variable, they are worrisome enough to suggest the possibility of malignancy in most cases. In the study by Kurt et al., 4 radiographs were available for 74 of the 80 cases. Of these 74 cases, 61 had radiographic features that were considered indicative of malignancy. An additional 10 cases appeared suggestive of a malignant tumor, though they also showed overlapping benign features. In the study by Ellis et al., 9 all 8 central low-grade osteosarcomas showed radiographic findings that suggested malignancy. Nevertheless, there are examples where more aggressive features are subtle or even impossible to detect. These cases require careful analysis with close histologic correlation. The potential for radiographic misdiagnosis was highlighted in a review by Choong et al., 10 focusing on long-term follow-up of 20 patients with low-grade central osteosarcoma who had been seen at the Mayo Clinic. Choong et al. 10 reported that an initial radiographic diagnosis of a benign process was made in 7 of 20 cases.
Central low-grade osteosarcomas tend to be large metaphyseal or diametaphyseal intramedullary tumors. It is not uncommon to see extension into the end of the bone when the epiphyseal plate is closed. Only a small percentage of cases have been reported to be cortical based lesions. Similarly, when fibrous dysplasia affects a long bone, the shaft and metaphysis are typically involved. However, fibrous dysplasia rarely extends to the articular cartilage.
Although the majority of central low-grade osteosarcomas are poorly marginated, up to one-third may show intermediate or well-defined margins suggesting an indolent or benign lesion. In comparison, fibrous dysplasia involving the long bones is always a well-circumscribed, benign-appearing lesion with or without a sclerotic rim (Figure 1A–B). The radiographic density of both tumors is variable, depending on the relative proportions of bone and fibrous tissue. Fibrous dysplasia has a hazy ground glass appearance on plain films. In contrast, low-grade osteosarcoma typically contains areas of heavy mineralization with regions of amorphous, cloud-like, or fluffy mineralization. 4,9
Cortical destruction is the most convincing radiographic feature in support of malignancy. Fibrous dysplasia is not associated with cortical destruction, except in cases that are associated with a pathologic fracture. In contrast, the majority of central low-grade osteosarcomas show some degree of cortical disruption with or without soft tissue extension. In some cases, it is obvious, whereas in others, it is subtler and therefore potentially overlooked. Computed tomography and magnetic resonance imaging can be quite useful in delineating the extent of the tumor and may even pick up cortical abnormalities that are not evident on plain films (Figure 2A—C). In the series by Kurt et al., 4 85% of their cases showed varying degrees of cortical destruction. Approximately half of these tumors demonstrated obvious destruction, and in the remaining cases, it was more in a localized region or incomplete. Fifty-five percent had identifiable soft tissue extension. Only in extremely rare cases will fibrous dysplasia form an exophytic mass.
Low-grade central osteosarcoma and fibrous dysplasia are composed of a bland fibroblastic stroma with variable amounts of bone production. Despite their similarities, one can usually separate these entities if careful attention is paid to a number of microscopic features. These primarily include (1) presence or absence of permeation; (2) stromal growth pattern; (3) cytologic atypia; (4) mitotic activity; and (5) pattern of bone production. It is important to be able to evaluate all these features. Therefore, as stated by Franceschina et al., 11 it is very difficult, if not impossible, to make the distinction with limited amounts of tissue.
The most helpful factor for differentiating low-grade central osteosarcoma from fibrous dysplasia is a permeative growth pattern. Therefore, it is important to spend some time at low magnification to assess the growth pattern of the lesion. If tumor is seen surrounding pre-existing bony trabeculae or permeating marrow fat, it is not fibrous dysplasia (Figure 3). Similarly, if the tumor destroys the cortex and invades the soft tissue, it should be low-grade central osteosarcoma.
The hypocellular fibroblastic stroma of fibrous dysplasia and of low-grade central osteosarcoma is composed of collagen-producing spindle cells. Both tumors can show variable cellularity within an individual lesion and between different lesions. However, in low-grade central osteosarcoma, the spindle cells tend to be arranged in intersecting bundles reminiscent of the pattern seen in low-grade fibrosarcoma. Sometimes this feature is subtle, but it is a helpful clue to the correct diagnosis, because these spindling fascicles are not typically seen in fibrous dysplasia.
Low-grade central osteosarcoma will contain some degree of cytologic atypia, whereas fibrous dyplasia does not (Figure 4A&B). Although the atypia is minimal, it is still an important distinguishing feature. At low magnification, occasional to moderate numbers of nuclei stand out because of their larger size, irregular shape, and dark chromatin pattern. They may be present in some areas of the tumor and absent in others, but overall there are more than what is acceptable in a benign lesion. Moreover, the spindle-shaped nuclei in low-grade central osteosarcoma tend to be longer and more slender than the short plump nuclei of fibrous dysplasia. If anaplasia is obvious at low magnification, the tumor is more likely to be a high-grade osteosarcoma.
Mitotic figures are extremely uncommon in fibrous dysplasia. In comparison, although low-grade central osteosarcoma does not contain brisk mitotic activity, at least occasional mitotic figures are almost always identified. Kurt et al. 4 found one or two mitotic figures per 10 high-power fields (HPF) in 82% of their cases and 18% had 3 or 4 mitoses per 10 HPF. 4 Therefore, suspect low-grade central osteosarcoma even when only a few scattered mitotic figures are identified from field to field.
Variable patterns of bone production are found in low-grade central osteosarcoma, which has led to a histologic classification scheme that includes three primary types: (1) fibrous dysplasia-like; (2) parosteal osteosarcoma-like; and (3) desmoid-like. The fibrous dysplasia-like variant contains irregularly shaped spicules of woven bone indistinguishable from what is found in fibrous dysplasia (Figure 5A–B). They can even mimic the classic “Chinese letters” pattern of fibrous dysplasia. In these cases, one must rely on other histologic and radiographic features to make the distinction. In the parosteal osteosarcoma-like pattern, moderate to heavy amounts of bone are present as long longitudinal seams of lamellar-like bone (Figure 6A–B). The histologic features are indistinguishable from parosteal osteosarcoma. In these cases, it is important to have radiographic correlation to confirm the intramedullary location of the tumor. Because this pattern of bone production is not found in fibrous dysplasia involving the long bones, it can be a helpful diagnostic finding. The desmoid-like pattern is the least common variant. In these tumors the amount of bone production is minimal. Consequently, the fascicles of spindle-shaped cells resemble the appearance of a desmoid tumor. The presence of cytologic atypia is helpful in making a diagnosis of malignancy.
Cartilage differentiation can be seen in both fibrous dysplasia and low-grade central osteosarcoma. Benign giant cells have been reported in up to 36% of low-grade central osteosarcomas. 4 They can also be found in fibrous dysplasia. Conversely, increased numbers of foam cells and stromal myxoid changes would favor a diagnosis of fibrous dysplasia over low-grade central osteosarcoma.
TREATMENT AND PROGNOSIS
Fibrous dysplasia is a benign lesion that can be successfully treated with curettage and bone grafting. For lesions in expendable bones (such as the ribs), en bloc resection may be indicated. When the lesion is asymptomatic and structural integrity is not compromised, observation alone is often recommended. With adequate surgery, the recurrence rate is very low, particularly in adults. Malignancy in fibrous dysplasia is extremely rare. 12–14 Although the frequency is probably less than 1%, the incidence varies in the literature from 0.4% to 6.7%. 14
Low-grade central osteosarcoma behaves in a more indolent fashion than high-grade conventional osteosarcoma. Nevertheless, it is associated with a high incidence of local recurrence after inadequate surgical margins. Recurrences may exhibit a higher histologic grade or dedifferentiation with the potential for metastases. It is metastatic tumor from the higher-grade recurrence that can lead to death in patients with low-grade central osteosarcoma. Therefore, the treatment of choice is resection with a wide surgical margin. In a study on 20 patients with low-grade central osteosarcoma, Choong et al. 10 reported that intralesional resection in 12 patients was associated with local recurrences. Four of the 12 recurrences were of a higher grade, and 3 of these patients died of their disease. Radical margins were not associated with local recurrence, a finding similar to other reports in the literature. 4–6
Low-grade central osteosarcoma and fibrous dysplasia can pose a challenging diagnostic differential diagnosis due to similar histologic and radiographic features. The distinction requires careful attention to the clinical, radiographic, and histologic findings. Most examples of low-grade central osteosarcoma will show some radiographic evidence, suggesting more aggressive behavior than what is typically seen in benign lesions. However, it may be subtle and consequently overlooked, particularly if there is a low index of suspicion. In addition to plain films, CT and MRI can be helpful in determining the extent of tumor. The most helpful histologic features that support a diagnosis of low-grade central osteosarcoma over fibrous dysplasia include permeation of bone or marrow fat, soft tissue extension, cytologic atypia, and mitotic activity. The treatment of choice for low-grade central osteosarcoma is resection with wide surgical margins because of the high incidence of local recurrence with inadequate margins. Recurrent tumor is often associated with a higher histologic grade. Misdiagnosing low-grade central osteosarcoma as fibrous dysplasia can result in undertreatment, leading to an increased risk for the development of metastatic disease.
1. Unni KK, Dahlin DC, McLeod RA, et al. Intraosseous well differentiated osteosarcoma. Cancer 1977; 40: 1337–47.
2. Unni KK. Dahlin’s Bone Tumors. 5th ed. Philadelphia: Lippincott-Raven Publishers; 1996.
3. Mirra JM, Picci P, Gold RM. Bone Tumors: Clinical, Radiologic and Pathologic Considerations. Philadelphia: Lea and Febiger; 1989; 359–83.
4. Kurt AM, Unni KK, McLeod RA, et al. Low-grade intraosseous osteosarcoma. Cancer 1990; 65: 1418–28.
5. Bertoni F, Bacchini P, Fabbri N, et al. Low-grade intraosseous-type osteosarcoma, histologically resembling parosteal osteosarcoma, fibrous dysplasia, and desmoplastic fibroma. Cancer 1993; 71: 338–45.
6. Xipell JM, Rush J. Case report 340: well differentiated intraosseous osteosarcoma of the left femur. Skelet Radiol 1985; 14: 312–6.
7. Sundaram M, Herbold D, McGuire MH. Case report 370: low-grade intramedullary osteosarcoma. Skelet Radiol 1986; 15: 338–42.
8. Iemoto Y, Ushigome S, Fukunaga M, et al. Case report 679. Skelet Radiol 1991; 20: 379–82.
9. Ellis JH, Siegel CL, Martel W, et al. Radiologic features of well-differentiated osteosarcoma. Am J Radiol 1988; 151: 739–42.
10. Choong PFM, Pritchard DJ, Rock MG, et al. Low grade central osteosarcoma: a long-term follow-up of 20 patients. Clin Orthop 1996; 322: 198–206.
11. Franceschina MJ, Hankin RC, Irwin RB. Low-grad central osteosarcoma resembling fibrous dysplasia: a report of two cases. Am J Orthop 1997; 26: 432–40.
12. Huvos AG, Higinbotham NL, Miller TR. Bone sarcomas arising in fibrous dysplasia. J Bone Joint Surg Am 1972; 54: 1047–56.
13. Yabut SM Jr, Kenan S, Sisson HA, et al. Malignant transformation of fibrous dysplasia. A case report and review of the literature. Clin Orthop 1988; 228: 281–9.
© 2001 Lippincott Williams & Wilkins, Inc.
14. Ruggieri P, Sim FH, Bond JR, et al. Malignancies in fibrous dysplasia. Cancer 1994; 73: 1411–24.