Malignant fibrous histiocytoma and fibrosarcoma of bone: a re-assessment in the light of currently employed morphological, immunohistochemical and molecular approaches.

Malignant fibrous histiocytoma (MFH) and fibrosarcoma (FS) of bone are rare malignant tumours and contentious entities. Sixty seven cases labelled as bone MFH (57) and bone FS (10) were retrieved from five bone tumour referral centres and reviewed to determine whether recent advances allowed for reclassification and identification of histological subgroups with distinct clinical behaviour. A panel of immunostains was applied: smooth muscle actin, desmin, h-caldesmon, cytokeratin AE1-AE3, CD31, CD34, CD68, CD163, CD45, S100 and epithelial membrane antigen. Additional fluorescence in situ hybridisation and immunohistochemistry were performed whenever appropriate. All cases were reviewed by six bone and soft tissue pathologists and a consensus was reached. Follow-up for 43 patients (median 42 months, range 6-223 months) was available. Initial histological diagnosis was reformulated in 18 cases (26.8 %). Seven cases were reclassified as leiomyosarcoma, six as osteosarcoma, three as myxofibrosarcoma and one each as embryonal rhabdomyosarcoma and interdigitating dendritic cell sarcoma. One case showed a peculiar biphasic phenotype with epithelioid nests and myofibroblastic spindle cells. Among the remaining 48 cases, which met the WHO criteria for bone FS and bone MFH, we identified five subgroups. Seven cases were reclassified as undifferentiated pleomorphic sarcoma (UPS) and 11 as UPS with incomplete myogenic differentiation due to positivity for at least one myogenic marker. Six were reclassified as spindle cell sarcoma not otherwise specified. Among the remaining 24 cases, we identified a further two recurrent morphologic patterns: eight cases demonstrated a myoepithelioma-like phenotype and 16 cases a myofibroblastic phenotype. One of the myoepithelioma-like cases harboured a EWSR1-NFATC2 fusion. It appears that bone MFH and bone FS represent at best exclusion diagnoses.

Osteofibrous dysplasia, osteofibrous dysplasia-like adamantinoma and adamantinoma: correlation of radiological imaging features with surgical histology and assessment of the use of radiology in contributing to needle biopsy diagnosis.

OBJECTIVES: The aim of this study was to correlate the imaging features with surgical histology for tibial osteofibrous dysplasia (OFD), osteofibrous dysplasia-like adamantinoma (OFD/LA) and classical adamantinoma and to determine the additional role of imaging in suggesting a correct diagnosis in cases of needle biopsy misdiagnosis. MATERIALS AND METHODS: This is a retrospective audit of 24 patients presenting over a 9-year period to a specialist orthopaedic oncology unit. Radiographic and axial magnetic resonance imaging (MRI) characteristics were recorded for each patient. The needle biopsy diagnosis and resection specimen histological diagnoses were retrospectively reviewed and compared with the imaging findings. RESULTS: The 24 cases comprised five OFD, 11 OFD/LA and eight adamantinoma based on surgical resection histology. The mean length of OFD was 6.1 cm (range 2-8.5 cm), for OFD/LA was 6.5 cm (range 2-13 cm) and for adamantinoma was 13.2 cm (range 6.5-26 cm). Seven of eight adamantinomas had moth-eaten margins compared to five of 11 OFD/LA and two of five OFDs. Three of eight adamantinomas demonstrated cortical destruction, with seven of eight cases completely involving the marrow cavity. In comparison, only one of 11 OFD/LA cases and one of five OFD cases demonstrated cortical destruction, and complete marrow involvement was rare. Four of 19 cases had a different needle biopsy result compared to the final histology, three cases being upgraded from an OFD/LA or OFD to classical adamantinoma. The radiological features of these three cases were more in keeping with a diagnosis of adamantinoma. CONCLUSIONS: A diagnosis of classical adamantinoma is suggested by an extensive lesion with moth-eaten margins and complete involvement of the medullary cavity on axial MR imaging. Misdiagnosis on needle biopsy may occur in up to one fifth of cases, and radiological features can assist in making the correct diagnosis.