DISCUSSION:
The World Health Organisation classification of CNS tumours recognizes a range of ependymal neoplasms, including clear cell, papillary, tanycytic, and perispinal myxopapillary variants [4, 5, 7, 9]. Typically, giant cells characterize none of these; most examples of ependymoma contain isomorphic cells, and even moderate cytologic pleomorphism tends only to be found in anaplastic tumors. However, several features of this tumor would place it among the ependymomas; it is well demarcated, contains foci of uniform cells that are associated with rosettes and canals, and electron microscopy has demonstrated cilia, microvilli and complex intercellular junctions, all of which are typical of this group of tumors [4]. Some uncommon gliomas, such as the pleomorphic xanthoastrocytoma (PXA) and giant cell glioblastoma, contain gigantic cells among a strikingly pleomorphic tumor cell population [3, 6], but the characteristics of the present tumor do not fit with these diagnoses, or the histopathologic features of a subependymal giant cell astrocytoma (SEGA). The SEGA, PXA and giant cell glioblastoma do not contain rosettes and canals. Other characteristics of the PXA, such as lipidized giant cells and a reticulin-rich architecture, were absent from this tumor. Also absent were the anaplastic features that would be expected in a giant cell glioblastoma.
Giant cell ependymomas have been reported, but are very rare [2, 8, 10]. Two have occurred in the filum terminale, one with features of a myxopapillary ependymoma and foci of giant cell formation, and one that was composed entirely of giant cells [10]. Two supratentorial examples have also been described, one intraventricular with features of anaplasia [2], and one extraventricular intraparenchymal example [8]. Both of these tumors contained pseudorosettes and cytological features that clearly identified them as ependymomas. We are unaware of any report of a giant cell ependymoma in the posterior fossa.
The biologic behavior of such a rare variant of ependymoma is difficult to predict. The relatively good outcome associated with the PXA suggests that the presence of grossly atypical cells in gliomas is not necessarily associated with a poor prognosis [6]. Ki-67 labeling index appears to be a prognostic indicator for ependymomas [1], and the relatively low Ki-67 LI in this case has so far (12 months) been associated with no disease progression.
REFERENCES
Contributed by Piyali Pal, MD, Helen Fernandes, MD, David W. Ellison, MD, PhD