Postoperative radiotherapy in the management of spinal cord ependymoma.

Fifty-eight patients with histologically verified spinal cord ependymomas were treated at the Royal Marsden Hospital and Atkinson Morley's Hospital between 1950 and 1987. The median age in this series was 40 years (range 1 to 79 years) and the male:female ratio was 1.8:1. Ten patients had tumors in the cervical cord and 10 in the thoracic cord; 14 tumors involved the conus medullaris and 24 the cauda equina. Forty ependymomas were grade I and 13 were grades II to IV (in five patients there was insufficient material for grading). Eleven patients underwent biopsy only, 33 had partial or subtotal resection, and 14 had complete resection. Forty-three patients received postoperative radiotherapy. The median follow-up period was 70 months (range 3 to 408 months). Cause-specific survival rates were 74% and 68% at 5 and 10 years, respectively. On univariate analysis, age, histological grade, postoperative neurological function, and era of treatment were significant prognostic factors for survival. The histological grade was the only significant independent prognostic factor. The relative risk of death from ependymoma was 9.0 for patients with tumor grades II to IV compared to grade I (p less than 0.005, 95% confidence interval 2.7 to 30). The survival rates of patients following complete excision were significantly better compared to those after incomplete surgery (p less than 0.025). The majority of completely resected neoplasms were low-grade cauda equina tumors. Despite incomplete surgery, 5- and 10-year progression-free survival rates following radical radiotherapy were both 59%, and cause-specific survival rates were 69% at 5 years and 62% at 10 years. This suggests that radiotherapy may achieve long-term tumor control in over half of those patients with residual spinal ependymoma.