Surveillance neuroimaging in childhood intracranial ependymoma: how effective, how often, and for how long?

Object. The authors examined images obtained in 52 children with intracrani al ependymomas to determine risk factors for tumor recurrence and to assess the impact of surveillance imaging on patient outcome. Methods. Data obtained in all children with intracranial ependymomas were p rospectively entered into a database from January 1987 to June 2000. The im aging and clinical details in all patients were reviewed. Fifty-two childre n with histologically proven intracranial ependymomas were treated at the a uthors' institution; recurrences developed in 28 (54%) of them, with a medi an time from surgery to first recurrence of 14.5 months (range 3-65 months) . Of these tumor recurrences, 43% were asymptomatic and were noted on surve illance imaging. Seventeen children died, all of whom had recurrences. inco mplete excision of the primary tumor was significantly associated with redu ced time to recurrence (p = 0.0144) and time to death (p = 0.0472). The age of the patient, location of the primary tumor, histological findings,and t he presence or absence of spinal metastases on preoperative imaging were no t significantly associated with outcome. The risk of death at any given tim e was 12-fold greater in patients in whom a recurrence was identified due t o symptoms rather than on surveillance images (p = 0.016). Conclusions. Recurrent childhood ependymoma has a poor prognosis. The exten t of the initial local tumor resection is the factor most closely associate d with outcome. Surveillance imaging reveals a substantial number of asympt omatic recurrences, and survival appears to be improved in these patients c ompared with those identified by symptoms. The improvement in survival is t hought to be greater than that expected just from earlier diagnosis.