Intraoperative Neuromonitoring in Patients with Intramedullary Spinal Cord Tumor: A Systematic Review, Meta-Analysis, and Case Series

Abstract

Intramedullary spinal cord tumors (IMSCT) comprise the least common types of spinal neoplasms; surgery is mostly conducted with intraoperative neuromonitoring (IONM). However, although IONM is widely used to prevent neurologic damage and many single-center studies have been reported, the added value, in terms of overall sensitivity and specificity, of different monitoring techniques used in IMSCT surgery remains to be clearly defined. To summarize and review the reported evidence on the use of IONM in IMSCT surgery, including our own case series (n= 75 patients). We systematically searched the literature published from January 2000 to February 2018. Articles included patients with IMSCT who underwent surgery with neuromonitoring. We estimated the sensitivity, specificity, 95% confidence interval (CI), positive likelihood ratio, and negative likelihood ratio. Of 1385 eligible articles, 31 were included in the systematic review. Fifteen of these articles were used for a meta-analysis, complemented with our own case series. All neuromonitoring modalities showed acceptable but not optimal test characteristics. For the indications used in the different studies, the motor evoked potentials approach showed the highest sensitivity (0.838; 95% CI, 0.703-0.919) and the best specificity (0.829; 95% CI, 0.668-0.921) for detecting neuronal injury. In our own case series, the extent of resection was significantly smaller in the false-positive group than in the true-negative group (P= 0.045). IONM showed high but not perfect sensitivity and specificity. IONM prevents neurologic damage but can also limit the extent of tumor resection. Prospective studies will have to define the definitive value of IONM.