Ependymomas Richard A. Prayson and Gene H. Barnett

Ependymomas represent a group of neoplasms derived from the ependymal lining of the central nervous system. These tumors generally arise proximal to the ventricular system and intramedullary region of the spinal cord. Ependymomas account for anywhere between 3% to 9% of glial tumors and comprise between 6% and 12% of intracranial tumors of childhood. They may arise at any age; however, peak incidences occur during the first decade of life and 4th decade of life. Signs and symptoms at the time of presentation are often related to the tumor's location proximal to the ventricular system and most frequently are a manifestation of hydrocephalus and increased intracranial pressure including headache, nausea, dizziness, and vomiting. Posterior fossa neoplasms may present with cerebellar ataxia, dizziness, paresis and visual disturbances. Supratentorial-based tumors may manifest with focal neurologic deficits and seizures. Radiographically and grossly, the tumors appear to be well circumscribed lesions. Variable degrees of contrast enhancement are observed on MRI. The gross and radiographic appearance of the neoplasm may be altered by areas of cystic degeneration, intratumoral hemorrhage, and calcifications. Occasionally with higher grade tumors, infiltration of adjacent brain structures are noted. Intramedullary ependymomas of the spinal cord are often associated with syrinx formation. Histologically, ordinary ependymomas are characterized by the formation of perivascular pseudorosettes and ependymal rosettes. Tumor cells are generally monomorphic in their appearance. At low magnification, ependymomas frequently have a pattern marked by variable cellularity with the hypocellular, fibrillary zones centered around blood vessels in the formation of perivascular pseudorosettes. Mitotic figures, particularly in low grade tumors, are relatively rarely encountered. Vascular proliferative changes and necrosis are not features of low grade ependymoma. The formation of true ependymal rosettes is marked by the lining of luminal spaces by ependymal cells. Frequently, the cells lining these spaces have a more epithelioid appearance. Mixtures of epithelioid and glial patterns of ependymoma are well described and have no known prognostic significance. Areas of myxoid or cystic degeneration, hemorrhage, microcalcification and metaplastic cartilage and bone formation have been variously described in these tumors. There is little agreement in the literature regarding precise criteria for grading ependymal tumors. The most commonly utilized grading schemas are two-tier systems in which tumors are designated as low grade ependymoma (WHO grade II) and anaplastic ependymoma (WHO grade III). The precise criteria which can be utilized to distinguish low from high grade ependymomas is not well established. Similar to other glioma types, higher grade ependymomas tend to be more cellular than low grade tumors, are marked by more nuclear pleomorphism, have increased mitotic activity, and more frequently demonstrate evidence of vascular proliferation and necrosis. The number of these features that need to be present and the degree to which they need to be present in a given tumor in order to designate it as high grade is debated. Consequently, the reproducibility of diagnoses in this arena is not great. Attempts at both three-tier and four-tier systems have not met with universal acceptance and also suffer from problems of reproducibility. A number of other issues further complicate grading. Similar to other glial tumors, ependymomas may be heterogeneous; therefore, sampling of the tumor becomes an important factor in obtaining a precise diagnosis. A number of studies have demonstrated the importance of a variety of clinical and surgical parameters in predicting tumor behavior. In many of these studies, these other parameters have proven to be more important then histologic grade in predicting the behavior of a given neoplasm. In general, outcome is better in adults versus children with ependymoma. This may, in part, be reflective of the fact the majority of pediatric-aged ependymomas arise in the posterior fossa region which historically has been a more difficult target for surgical excision. Also, there is some suggestion that tumors that are histologically higher grade are more commonly encountered in the pediatric-age group. A particularly poor prognosis is observed in children with tumors arising in the first two years of life. In several studies, the extent of surgical resection has been noted to be an independent prognostic factor. Attempts at predicting tumor behavior based solely on histology have been frought with problems. Several studies have attempted to examine the potential utilization of cell proliferation markers in the evaluation of ependymomas. In general, higher rates of cell proliferation are observed in higher grade tumors; however, a reliable correlation of proliferation indices with prognosis has yet to be established. Genetic studies of ependymal tumors have not yet yielded information useful for prognostication. The association of spinal ependymomas and neurofibromatosis type II is well known. Approximately 30% of ependymomas are associated with abnormalities involving chromosome 22, with the majority of these represented by monosomy 22 or deletions/translocations involving the long arm of chromosome 22. A variety of other chromosomes have been noted to contain abnormalities in a small percentages of ependymomas including sites on chromosomes 7, 9, 10, 13 and 17. In contrast to other glioma types, mutations or deletions associated with tumor suppressor gene CDKN2a and CDKN2b are not common and amplification of CDK4 or CCND1 are not frequently observed. Epidermal growth factor receptor amplification and mutations of the TP53 tumor suppressor gene are not frequent findings. The immunohistochemical profile of most ependymomas is somewhat nondescript. Most tumors demonstrate immunoreactivity with GFAP, S-100 protein, and vimentin antibodies. Many of the lower grade ependymomas, particularly those with more epithelioid features, are likely to demonstrate focal immunoreactivity with epithelial membrane antigen and cytokeratin antibodies. There generally is no evidence of neuronal antigen expression immunohistochemically in these tumors. The ultrastructural appearance of ependymomas is quite distinctive and can be helpful in distinguishing ependymoma from other glioma types, particularly astrocytoma, when biopsy material is limited and histologic evidence of rosettes and pseudorosettes is not evident. Ultrastructurally, ependymomas are marked by the presence of ciliary body attachments known as blepharoplasts. Frequently, microvilli are located at the luminal surface and junctional complexes at the lateral surfaces. Histologically, a number of variants of ependymoma have been described that provide, at times, a diagnostic challenge. Variants of low grade (WHO grade II) ependymoma include papillary, clear cell and tanycytic types. The papillary ependymoma variant is marked by the presence of well-formed papillae which contain glial vascular cores. Distinction of this variant from choroid plexus papillomas and papillary meningiomas is important. Both of the latter two lesions are marked by fibrovascular cores and have different immunohistochemical profiles. As previously mentioned, ependymomas may focally have clear cell areas resembling low grade oligodendroglioma. The distinction of this lesion from oligodendroglioma or central neurocytoma is often dependant on recognizing more classic areas of ependymal differentiation. Ultrastructural identification of characteristic ependymal features may also be useful. Tanycytic ependymomas are marked by spindled appearing cells, arranged in fascicles. Rosettes and pseudorosettes are often not well delineated in these lesions and confusion of this entity with an astrocytic tumor is common. Again, the ultrastructural appearance of the tumor may be helpful in distinguishing it from an astrocytoma. Rare examples of other ependymoma variants including a lipomatous giant cell type, tumors with extensive cellular vacuolization, melanotic type and signet ring cell type have been described. Two other lesions which are generally considered related to the ependyma, but are important to distinguish from ordinary ependymomas because of their more favorable prognosis (WHO grade I tumors) include the myxopapillary ependymoma and subependymoma: Table 10: Differential Diagnostic Features of Ependymal Neoplasms and Astrocytoma   Ependymoma Astrocytoma MyxopapillaryEpendymoma Subependymoma Age Childhood & Adulthood Mostly adulthood Adulthood Adulthood Location Intraventricular, spinal cord White matter, spinal cord Filum terminale Intraventricular, rarely spinal cord Infiltrative ± + - - Rosettes/pseudorosettes + - - - Mucoid stroma - - + - Papillary architecture + rare - ± - GFAP/S-100 + + + + EMA/Keratin ± -* - - Cell proliferation indices Variable-correlates with grade Variable-correlates with grade Low Low Cilia/microvilli/blepharoplasts + - ± ± * Some cross reactivity of certain keratin markers (such as AE1/3) with astrocytomas has been observed. Myxopapillary ependymomas typically arise in the filum terminale region, often presenting with a long history of lower back pain. Grossly, they appear to be well circumscribed, lobulated masses. Microscopically, they have a distinctive architectural pattern marked by the presence of papillary structures which contain central vessels surrounded by a mucoid stroma and ependymal-type cells. True rosettes and pseudorosettes are not a salient feature of the myxopapillary ependymoma variant. Similar to ordinary ependymomas, the tumor cells stain with antibodies to GFAP, S-100 protein, and vimentin. As expected, rates of cell proliferation are rather low in these tumors, typically less than 2% in most cases. Prognosis is excellent, particularly with gross total resection of the neoplasm. The subependymoma usually presents as an incidental finding at time of autopsy. Most, arise either in the region of the 4th ventricle or lateral ventricles. Rarely, the tumor becomes large enough or is situated in a position that results in symptoms. Both grossly and microscopically, it is a well demarcated lesion, marked by clusters of monomorphic appearing nuclei arranged against a dense fibrillary background ("islands of blue in a sea of pink"). Mitotic activity, nuclear pleomorphism, and necrosis are not features of this tumor. Microcystic degeneration is a fairly common finding of most of these neoplasms. Interestingly, tumor cells demonstrate ultrastructural features of both ependymal and astrocytic derivation. Similar to the myxopapillary ependymoma, rates of cell proliferation are rather low in these tumors, with MIB-1 labeling indices generally being less than 1%. Surgical removal of the tumor is usually considered curative and the prognosis for these tumors is excellent. Selected References Áfra D, Müeller W, Slowik F, et al. Supratentorial lobar ependymomas: reports on the grading and survival periods in 80 cases, including 46 recurrences. Acta Neurochir 69:243-251, 1983. Ang LC, Taylor AR, Bergin D, et al. An immunohistochemical study of papillary tumors in the central nervou system. Cancer 65:2712-2719, 1990. Brown DF, Chason DP, Schwartz LF, et al. Supratentorial giant cell ependymoma: A case report. Mod Pathol 11:398-403, 1998. Burger PC, Shibata T, Kleihues P. The use of monoclonal antibody Ki-67 in the identification of proliferating cells: Application to surgical neuropathology. Am J Surg Pathol 10:611-617, 1986. Ernestus RI, Wilcke O, Schroder R. Supratentorial ependymomas in childhood: Clinicopathological findings and prognosis. Acta Neurochir 111-96-102, 1991. Hamilton RL, Pollack IF: The molecular biology of ependymomas. Brain Pathol 7:807-822, 1997. Hirato J, Nakazato Y, Iijima M, et al. An unusual variant of ependymoma with extensive tumor cell vacuolization. Acta Neuropathol 93:310-316, 1997. Hirose Y, Aldape K, Bollen A, et al. Chromosomal abnormalities subdivide ependymal tumors into clinically relevant groups. Am J Pathol 158:1137-1143, 2001. Ilgren EB, Stiller CA, Hughes JT, et al. Ependymomas: a clinical and pathologic study. Part II. Survival features. Clin Neuropathol 3:122-127, 1984. Jallo GI, Zagzag D, Epstein F. Intramedullary subependymoma of the spinal cord. Neurosurgery 38:251-257, 1996. Kawano N, Yada K, Yagishita S. Clear cell ependymoma: a histological variant. Virchows Arch [A] 415:467-472, 1989. Kernohan JW, Mabon RF, Svien JH, et al. A simplified classification of gliomas. Proc Staff Meet Mayo Clin 24:71-75, 1949. Mørk SJS, Løken AC. Ependymoma: a follow-up study of 101 cases. Cancer 40:907-915, 1977. Mathews T, Moossy J. Gliomas containing bone cartilage. J Neuropathol Exp Neurol 33:456-471, 1974. Mazewski C, Soukup S, Ballard E, et al. Karyotype studies in 18 ependymomas with literature review of 107 cases. Cancer Genet Cytogent 113:1-8, 1999. Min K-W, Scheithauer BW. Clear cell ependymoma: a mimic of oligodendroglioma: clinicopathologic and ultrastructural considerations. Am J Surg Pathol 21:820-826, 1997. Nagashima T, Hoshino T, Cho KG, et al. The proliferative potential of human ependymomas measured by in situ bromodeoxyuridine labeling. Cancer 61:2433-2438, 1988. Needle MN, Goldwein JW, Grass J, et al. Adjuvant chemotherapy for the treatment of intracranial ependymoma of childhood. Cancer 80:341-347, 1997. Prayson RA, Suh JH. Subependymomas: clinicopathologic study of 14 tumors including comparative MIB-1 immunohistohemistry with other ependymal neoplasms. Arch Pathol Lab Med 123:306-309, 1999. Prayson RA. Myxopapillary ependymomas: a clinicopathologic study of 14 cases including MIB1 and p53 immunoreactivity. Mod Pathol 10:304-310, 1997. Prayson RA. Clinicopathologic study of 61 patients with ependymoma including MIB-1 immunohistochemistry. Ann Diagn Pathol 3:11-18, 1999. Prayson RA. Cyclin D1 and MIB-1 immunohistochemistry in ependymomas: a study of 41 cases. Am J Clin Pathol 110:629-634, 1998. Rawlings VE III, Giangaspero F, Burger PC, et al. Ependymomas: a clinicopathologic study. Surg Neurol 29:271-281, 1988. Rosenblum MK, Erlandson RA, Aleksic SN, et al. Melanotic ependymoma and subependymoma. Am J Surg Pathol 14:729-736, 1990. Ross DA, McKeever PE, Sandler HM, et al. Myxopapillary ependymoma: results of nucleolar organizing region staining. Cancer 71:3114-3118, 1993. Ruchoux M-M, Kepes JJ, Dhellemmes P, et al. Lipomatous differentiation in ependymomas. A report of three cases and comparison with similar changes reported in other central nervous system neoplasms of neuroectodermal origin. Am J Surg Pathol 22:338-346, 1988. Sara A, Bruner JM, Mackay B. Ultrastructure of ependymoma. Ultrastruct Pathol 18:33-42, 1994. Scheithauer BW. Symptomatic subependymoma: report of 21 cases with review of the literature. J Neurosurg 49:689-696, 1978. Schiffer D, Chi b A, Cravioto H, et al. Ependymoma: Internal correlations among pathological signs: The anaplastic variant. Neurosurgery 29:206-210, 1991. Schwartz TH, Kim S, Glick RS, et al. Supratentorial ependymomas in adult patients. Neurosurgery 44:721-731, 1999. Sonneland PRL, Scheithauer BW, Onofrio BM. Myxopapillary ependymoma. A clinicopathologic and immunohistochemical study of 77 cases. Cancer 56:883-893, 1985. Zuppan CW, Mierau GW, Weeks DA, et al. Ependymoma with signet ring cells. Ultrastruct Pathol 18:43-46, 1994.