Prognostic Factors in Childhood Intracranial Ependymomas: The Role of Age and Tumor Location

Abstract

Despite several clinical reports on intracranial ependymomas in children, the factors which affect prognosis, and the possibility that certain combinations of factors might limit survival, are still a matter of debate. Between 1976 and 1996 we operated on 35 children with intracranial ependymomas. Postoperative irradiation was given to 27 patients, with associated chemotherapy in 6 cases. Mean follow-up was 62 months. In 12 patients a 5-year follow-up was possible. In October 1996, 18 patients (51.4%) were still alive, the longest disease-free follow-up being 20 years, and the shortest 8 months. We analyzed the prognostic relevance of eight factors. For each factor, different subgroups were distinguished and compared as follows: age at diagnosis (<4 vs. ≥4 years), sex, tumor location (supratentorial vs. infratentorial), tumor size (<4 vs. 4–7 vs. >7 cm), surgical removal (total vs. subtotal), histology (low-grade vs. anaplastic), morphology (solid vs. cystic), adjuvant therapies (treatment vs. no treatment). Two-way contingency tables were made to identify associations between variables. The only significant association was between age and tumor location (p = 0.022): in children under 4, tumors were almost invariably located in the posterior fossa (9 out of 10 cases) with a clear preference for the lateral recess (8 cases). Other correlations were not significant. Kaplan-Meier survival curves were compared to assess the prognostic relevance of each factor. Survival was significantly lower for children under 4, for those with posterior fossa tumors, and for patients with residual tumor (p < 0.05). A multivariate analysis compared variables which significantly affected survival, revealing that age is the most important factor affecting prognosis (p < 0.05), while tumor location and surgical removal do not add any significance to the effect of age on survival. We conclude that age has the strongest prognostic relevance in childhood intracranial ependymomas, while the effect of tumor location on survival may be related to the high incidence of lateral recess ependymomas in younger children.