Predictors of survival in patients with spinal ependymoma.

OBJECTIVES: We aimed to examine the impact of demographic and treatment factors on overall survival among spinal ependymoma patients. METHODS: Using data from Surveillance, Epidemiology, and End Results (SEER) from 1990-2008, we evaluated subjects with histologically confirmed spinal cord ependymoma. RESULTS: We identified 1,353 patients with spinal ependymoma (mean age 43.5 years). Among these, 26 had anaplastic ependymoma (AE), 374 had myxopapillary ependymoma (MPE), and 953 had other low-grade ependymoma (non-MPE). Median follow-up was 61 months. Overall survival at 5 years was 97% for MPE, 92% for low-grade non-MPE, and 73% for AE. Gross total resection (GTR) was achieved in 58% of subjects with MPE, 51% with low-grade non-MPE, and 50% with AE. Radiation occurred more often after subtotal resection (STR) than after GTR, and more often among those with anaplastic histology. Histology and extent of surgical resection were significant prognostic factors in multivariate analyses. Compared to MPE, subjects with low-grade non-MPE had a higher risk of mortality at 5 years (HR 2.35, P = 0.03), as did subjects with AE (HR 8.63, P < 0.01). Compared to GTR, STR was associated with an increased risk of mortality at 5 years (HR 2.2, P = 0.01), as was biopsy only (HR 2.05, P = 0.03) and no surgery (HR 4.97, P < 0.01). Among patients with either STR or GTR, adjuvant radiotherapy did not confer a survival benefit at 5 years (STR: HR 2.29, P = 0.07, and GTR: HR 2.2, P = 0.12). DISCUSSION: We found that lower grade histology and higher extent of surgical resection were significant prognostic factors for more favourable survival outcomes for spinal ependymoma patients.