Chondroid chordoma. A hyalinized chordoma without cartilaginous differentiation.

"Chondroid chordoma" is a controversial and confusing entity that was originally described by Heffelfinger and colleagues as a biphasic malignant neoplasm possessing elements of both chordoma and cartilaginous tissue. Because the premise for this distinction was based strictly on histomorphologic criteria, the light microscopic, immunohistochemical, and electron microscopic features of the chondroid and chordoid areas of five chondroid chordomas of the skull base were evaluated separately, and compared to five typical chordomas and six low grade chondrosarcomas. Using light microscopy, chondroid chordoma revealed areas that resembled typical chordoma (chordoid areas) and areas that resembled low grade chondrosarcoma (chondroid areas). However, both the chordoid and chondroid areas had an epithelial phenotype and stained strongly for cytokeratin and EMA as well as S-100. 5'-nucleotidase, an enzyme that has been described in chordoma but not in chondrosarcoma, was found in both the chordoid and chondroid areas of one chondroid chordoma. Electron microscopic studies of both the chordoid and chondroid areas in four of the tumors demonstrated both tonofibrils and desmosomes. Chordoma demonstrated immunohistochemical and electron microscopic features that were nearly identical to chondroid chordoma. Chordoma was cytokeratin, EMA, S-100, and 5'-nucleotidase positive. Ultrastructurally, chordoma exhibited variably-sized vacuoles, abundant rough endoplasmic reticulum (RER), and desmosomes with tonofilaments. In contrast to chondroid chordoma, chondrosarcoma consistently stained for only S-100 protein and was cytokeratin, EMA and 5'-nucleotidase negative. Ultrastructurally, chondrosarcoma demonstrated a flocculogranular matrix, glycogen, abundant RER, and scalloped cellular outlines, but lacked desmosomes with tonofilaments. These findings indicate that "chondroid chordoma" is a variant of chordoma with histologic features that may mimic chondrosarcoma. Despite the resemblance of these hyalinized areas to cartilaginous tissue, these tumors retain their epithelial phenotype. Biphasic differentiation is not present. These findings undermine the original premise for distinguishing "chondroid chordoma" from typical chordoma. The authors propose that these tumors be classified as "hyalinized chordomas," rather than "chondroid chordoma," to clarify their histogenesis and avoid confusion with chondrosarcomas of the base of the skull.