Chordoid glioma of the third ventricle: a report of two new cases, with further evidence supporting an ependymal differentiation, and review of the literature.

The term chordoid glioma of the third ventricle was first used to describe a rare and slowly growing neoplasm of uncertain histogenesis, with chordoid appearance, occurring preferentially in middle-aged women. Herein we report two additional examples of this novel entity together with a literature review based on the 25 cases previously published. Our review fully confirms the strikingly stereotyped clinical, neuroradiologic, and pathologic features of this unique tumor. The female/male ratio was 1.7:1, and the age range was 24-70 years (mean 44.9 years). In all 27 cases imaging findings were similar showing a well-defined mass (mean 2.8 cm in largest dimension), ovoid in shape, hyperdense on CT scans, with uniform and intense contrast enhancement, arising in the hypothalamic/suprasellar/third ventricular region. Histologically, the main consistent characteristics were cords and clusters of epithelioid cells within an abundant mucinous and often vacuolated background. Mitoses were sparse or absent and anaplastic features, endothelial proliferation, and necrosis were not identified. Lymphoplasmacytic infiltrates with Russell bodies were frequent throughout the tumor and its interface with adjacent brain parenchyma. Most of the tumor cells revealed a strong and diffuse expression of vimentin and glial fibrillary acidic protein. Additionally, the vast majority of tumors showed focal coexpression of cytokeratins, CD34, S-100 protein, and epithelial membrane antigen; the MIB-1 labeling indices were uniformly low. Surprisingly for a glioma assigned WHO grade II, the 19 patients with an available but short follow-up (mean 22.5 months; range 6-68 months) experienced a rather poor outcome (three recurrences and seven deaths), probably reflecting the anatomic site of the neoplasm that precludes a complete surgical excision rather than its histologic composition. Ultrastructural examination of 10 cases demonstrated findings in line with a glial derivation and a putative ependymal origin such as cytoplasmic intermediate filaments, microvilli, intermediate junctions or desmosomes, and focal basal lamina formation. In our case no. 1, and for the first time in this tumor, we observed sparse and abnormal cilia in an aberrant juxtanuclear location, a further argument for considering chordoid glioma as a subtype of ependymoma. However, a better understanding of the biologic behavior and histogenesis of this distinctive clinicopathologic entity needs to be investigated with a larger series. Nevertheless, taking into account its strikingly consistent anatomic localization, its unique histopathologic and immunohistochemical profile, in conjunction with the most recent and convincing ultrastructural arguments, we suggest that chordoid glioma of the third ventricle could be better classified as chordoid ependymoma of the lamina terminalis area.