Osteofibrous dysplasia of long bones--a reactive process to adamantinomatous tissue

The most controversial aspect of osteofibrous dysplasia (OFD) is its possible histogenetic relationship to adamantinoma of long bone. Evidence is recently beginning to accumulate that OFD may be a reactive process to regressive adamantinoma. To verify the concept, 13 lesions of OFD were studied again by immunohistochemistry for cytokeratins of different molecular masses, as well as by conventional stainings. In addition, 2 adamantinomas and 6 fibrous dysplasias of the tibia were studied for reference. A small number of spindle- or ovoid-shaped cells scattered individually in the fibro-osseous stroma showed positive reactions for cytokeratins of 55-57 kDa in 2 lesions, and for those of 45-56.5 kDa in 8 lesions of 13 OFDs, although no definite epithelial island could be detected even by immunohistochemistry. Adamantinomas also showed single cytokeratin-positive cells dispersed in fibroblastic stroma, in addition to epithelial islands positive for cytokeratins of both 55-57 kDa and 45-56.5 kDa. All cases of fibrous dysplasia were negative for cytokeratins. During the observation, no case of OFDs progressed to classic adamantinoma. The present study, demonstrating the existence of an intermediate stage between "differentiated adamantinoma" and total elimination of adamantinomatous components, gives further support for the concept that OFD is a secondary reactive process to adamantinomatous tissue. In practice, the existence of single scattered cytokeratin-immunoreactive cells in otherwise typical OFDs may not indicate the truly malignant behaviour of classic adamantinoma, unless discrete epithelioid cell nests are also found.