Low baseline Tc-MEP amplitude is associated with false-positive alerts in spine surgery: a ROC-based analysis across spinal pathologies.

Abstract

This study aimed to determine whether absolute baseline transcranial motor evoked potential (Tc-MEP) amplitude is associated with false-positive (FP) alerts during intraoperative neurophysiological monitoring and to estimate pathology-specific cutoff values. We retrospectively reviewed 307 high-risk spine surgery cases monitored with Tc-MEPs from 2018 to 2020. For waveform-level analysis, 1,748 muscle waveforms from 287 FP/true-negative cases were analyzed. Alerts were defined as final amplitude ≤ 30% of baseline, and FP alerts were defined as alerts without new postoperative motor deficit. Receiver operating characteristic analyses were performed to estimate baseline-amplitude cutoff values for FP alerts in spinal cord tumor surgery, spinal deformity surgery, ossification of the posterior longitudinal ligament surgery, and the pooled cohort. FP alerts occurred in 253 of 1,748 waveforms (14.5%). Baseline Tc-MEP amplitude was lower in FP than in true-negative waveforms (351 ± 533 µV vs. 604 ± 1007 µV). FP alerts were more frequent in lower-extremity recordings. Exploratory ROC-derived cutoff values were 122 µV for spinal cord tumor surgery, 180 µV for spinal deformity surgery, 160 µV for ossification of the posterior longitudinal ligament surgery, and 170 µV for the pooled cohort; the corresponding AUCs were 0.68, 0.56, 0.61, and 0.61, respectively. Lower baseline Tc-MEP amplitudes were associated with FP alerts. Baseline amplitude assessment may help identify low-amplitude signals that warrant optimization and cautious interpretation before percentage-based alarm criteria are applied.