The prognostic significance of postoperative residual tumor in ependymoma.

Between 1970 and 1989, 29 patients with intracranial ependymomas were evaluated and treated at the Children's Hospital in Boston. With a median follow-up of 82 months, the actuarial survival rates at 5 and 10 years were 61 +/- 10% and 46 +/- 12%, respectively. Anaplastic histological findings were uncommon (2 of 29). Initial postoperative radiotherapy was given to 25 patients, with a median tumor dose of 5360 cGy. With a median time to recurrence of 22 months, local failure (within 2 cm of original enhancing mass) was the predominant pattern of relapse (15 of 16 failures). The presence of radiographic residual disease seen on postoperative magnetic resonance imaging or computed tomographic scans was the most important prognostic variable for patients with intracranial ependymoma. Analysis of the 19 patients who underwent postoperative imaging revealed a 75 +/- 15% 5-year freedom from progressive disease for 9 patients with no residual disease, as compared with 0% freedom from progressive disease for the 10 patients with gross residual disease (P = 0.03). In contrast, the surgical assessment of residual disease was not significant (P = 0.4). Age at presentation was also a significant prognostic factor. The overall actuarial survival rate at 12 years for infants 24 months or younger at diagnosis was 0%, as compared with 62 +/- 13% for older patients (P = 0.03). For non-anaplastic ependymomas, complete surgical resection followed by local-field, high-dose (greater than 54 Gy) radiotherapy appears to offer the greatest chance for long-term survival. Because of the markedly reduced survival rate for patients with radiologically apparent postoperative disease, maximal surgical resection and novel therapeutic endeavors appear warranted for this high-risk group. Future protocols should use postoperative imaging, not operative reports, to stratify patients with ependymoma.