Histologic characterization of 41 ependymomas with the help of a personal computer

Abstract

Fourty-one ependymomas were histologically analyzed in relation to patient age and sex and tumor location. A discriminant analysis model using Bayes' formula and a personal computer were employed. Ependymomas situated in the posterior fossa had a higher incidence in children. Ependymoblastomas were all situated above the tentorium and occurred only in young children. We identified three tumor groups on the basis of common histologic characteristics: benign ependymomas, anaplastic ependymomas, and ependymoblastomas. The main features useful for the grouping were the degree of differentiation and anaplasia. Ependymomas from the cauda equina showed histologic characteristics that allowed them to be differentiated from other benign ependymomas. In each group the particular histologic characteristics, age, and location were significant in terms of diagnosis and prognosis. This method of analysis may help to more precisely define ependymomas and may provide pathologists and clinicians with a quantifiable diagnostic tool that may be of help in establishing the appropriate treatment.