Combined Motor and Sensory Intraoperative Neuromonitoring for Cervical Spondylotic Myelopathy Surgery Causes Confusion: A Level-1 Diagnostic Study.

Abstract

Level-1 diagnostic study. The purpose of this study was to evaluate the sensitivity and specificity of combined motor and sensory intraoperative neuromonitoring (IONM) for cervical spondylotic myelopathy (CSM). Intraoperative neuromonitoring during spine surgery began with sensory modalities with the goal of reducing neurological complications. Motor monitoring was later added and purported to further increase sensitivity and specificity when used in concert with sensory monitoring. Debate continues, however, as to whether neuromonitoring reliably detects reversible neurologic changes during surgery or simply adds set-up time, cost, or mere medicolegal reassurance. Neuromonitoring data using combined motor and sensory evoked potentials for 540 patients with CSM undergoing anterior or posterior decompressive surgery were collected prospectively. Patients were examined postoperatively to determine the clinical occurrence of new neurologic deficit which correlated with monitoring alerts recorded per established standard criteria. The overall incidence of positive IONM alerts was 1.3% (N = 7) all of which were motor alerts. All were false positives as no patient had clinical neurological deterioration post-operatively. The false-positive rate was 1.4% (N = 146) for anterior surgeries and 1.3% (N = 394) for posteriors with no statistical difference between them (P = 1.0, Fisher exact test). There were no false-negative alerts, and all negatives were true negatives (N = 533). The overall sensitivity of detecting a new neurologic deficit was 0%, overall specificity 98.7%. Combined motor and sensory neuromonitoring for CSM patients created a confusing choice between the motor or sensory data when in disagreement in 1.3% of surgical patients. Criterion standard clinical examinations confirmed all motor alerts were false positives. Surgical plan was negatively altered by following false motor alerts early on, but disregarded in later cases in favor of sensory data. Neuromonitoring added set-up time and cost, but without clear benefit in this series.Level of Evidence: 4.