Factors associated with progression-free survival and long-term neurological outcome after resection of intramedullary spinal cord tumors: analysis of 101 consecutive cases.

Object: With the introduction of electrophysiological spinal cord monitoring, surgeons have been able to perform radical resection of intramedullary spinal cord tumors (IMSCTs). However, factors associated with tumor resectability, tumor recurrence, and long-term neurological outcome are poorly understood.

Methods: The authors retrospectively reviewed 101 consecutive cases of IMSCT resection in adults and children at a single institution. Neurological function and MR images were evaluated preoperatively, at discharge, 1 month after surgery, and every 6 months thereafter. Factors associated with gross-total resection (GTR), progression-free survival (PFS), and long-term neurological improvement were assessed using multivariate regression analysis.

Results: The mean age of the patients was 41 +/- 18 years and 17 (17%) of the patients were pediatric. Pathological type included ependymoma in 51 cases, hemangioblastoma in 15, pilocytic astrocytoma in 16, WHO Grade II astrocytoma in 10, and malignant astrocytoma in 9. A GTR was achieved in 60 cases (59%). Independent of histological tumor type, an intraoperatively identifiable tumor plane (OR 25.3, p < 0.0001) and decreasing tumor size (OR 1.2, p = 0.05) were associated with GTR. Thirty-four patients (34%) experienced acute neurological decline after surgery (associated with increasing age [OR 1.04, p = 0.02] and with intraoperative change in motor evoked potentials [OR 7.4, p = 0.003]); in 14 (41%) of these patients the change returned to preoperative baseline within 1 month. In 31 patients (31%) tumor progression developed by last follow-up (mean 19 months). Tumor histology (p < 0.0001) and the presence of an intraoperatively identified tumor plane (hazard ratio [HR] 0.44, p = 0.027) correlated with improved PFS. A GTR resulted in improved PFS for hemangioblastoma (HR 0.004, p = 0.04) and ependymoma (HR 0.2, p = 0.02), but not astrocytoma. Fifty-five patients (55%) maintained overall neurological improvement by last follow-up. The presence of an identifiable tumor plane (HR 3.1, p = 0.0004) and improvement in neurological symptoms before discharge (HR 2.3, p = 0.004) were associated with overall neurological improvement by last follow-up (mean 19 months).

Conclusions: Gross-total resection can be safely achieved in the vast majority of IMSCTs when an intraoperative plane is identified, independent of pathological type. The incidence of acute perioperative neurological decline increases with patient age but will improve to baseline in nearly half of patients within 1 month. Long-term improvement in motor, sensory, and bladder dysfunction may be achieved in a slight majority of patients and occurs more frequently in patients in whom a surgical plane can be identified. A GTR should be attempted for ependymoma and hemangioblastoma, but it may not affect PFS for astrocytoma. For all tumors, the intraoperative finding of a clear tumor plane of resection carries positive prognostic significance across all pathological types.