Pediatr Neurosurg 1996 ;24(3):119-25
Bristol Royal Hospital for Sick Children, UK.
A retrospective study of 31 children presenting in the period 1976-1993 with intracranial ependymomas in the south-west region of the UK was performed. The diagnosis was confirmed by histological review and several histological variables recorded, including pleomorphism, cellularity, numbers of mitotic figures and extent of necrosis. Histological features, presentation and treatment were correlated with outcome. Eleven patients had supratentorial primaries and 20 infratentorial primaries. Seven patients had macroscopically complete resections of which 5 had supratentorial primaries. Eight patients had no other therapy apart from surgery. Six patients received craniospinal irradiation, 11 had involved-field irradiation and 14 did not have radical radiotherapy. Twelve patients received chemotherapy. Eleven patients are relapse-free at a median of 43 months. Of the 20 relapses, 19 were local and only 1 metastatic. The grade of the tumor assigned at presentation had no correlation with outcome. On histological review, the only feature that had a significant correlation (p = 0.05) with survival was tumor cellularity. There was a survival advantage for a supratentorial primary site (p = 0.04). The event-free survival was significantly longer for patients diagnosed after 1986 than before (p = 0.006). Survival was not significantly affected by radiotherapy or chemotherapy. Children who had had a macroscopically complete excision of the tumor had a 5-year survival of 69% as opposed to 47% for those who had incomplete surgery, but this difference is not significant (p = 0.13). The data suggest that: (1) the event-free survival rate has improved significantly with time; (2) there is little correlation between a variety of pathological features and outcome, and (3) failure at the primary site is the major obstacle to improved cure rates. The role of complete resection, including second-look surgery, needs further evaluation.