Utility of somatosensory- and motor-evoked potential change thresholds in surgical treatment for thoracic spinal stenosis based on different pathologies

BACKGROUND CONTEXT: Thoracic spinal stenosis (TSS) is secondary to different pathologies that differ in clinical characteristics and surgical outcomes. PURPOSE: This study aimed to determine the optimal warning thresholds for combined somatosensory-evoked potentials (SSEP) and motor-evoked potentials (MEP) for predicting postoperative neurological deterioration in surgical treatment for TSS based on different pathologies. Additionally, we explored the correlation between SSEP/MEP monitoring and postoperative spinal neurological function. STUDY SETTING: Retrospective study. PATIENT SAMPLE: Two hundred five patients. OUTCOME MEASURES: We obtained perioperative modified Japanese Orthopedic Association (mJOA) scores to assess spinal neurological function. METHODS: The data collected in this study included demographic data, intraoperative neurophysiological monitoring (IONM) signals, and perioperative neurological function assessments. To determine the optimal IONM warning threshold, a receiver operating characteristic (ROC) curve was used. Additionally, Pearson correlation analysis was conducted to determine the correlation between IONM signals and clinical neurological conditions. RESULTS: A total of 205 consecutive patients were eligible. Forty-one patients had thoracic disc herniation (TDH), 14 had ossification of the posterior longitudinal ligament (OPLL), 124 had ossification of the ligamentum flavum (OLF), and 26 had OPLL+OLF. The mean mJOA scores before surgery and 3 months after surgery were 7.0 and 7.9, respectively, resulting in a mean mJOA recovery rate (RR) of 23.1%. The average postoperative mJOA RRs for patients with TDH, OPLL, OLF, and OPLL+OLF were 24.8%, 10.4%, 26.8%, and 11.2%, respectively. Patients with OPLL+OLF exhibited a more stringent threshold for IONM changes. This included a lower amplitude cutoff value (a decrease of 49.0% in the SSEP amplitude and 57.5% in the MEP amplitude for short-term prediction) and a shorter duration of waveform change (19.5 minutes for SSEP and 22.5 minutes for MEP for short-term prediction). On the other hand, patients with TDH had more lenient IONM warning criteria (a decrease of 49.0% in SSEP amplitude and 77.5% in MEP amplitude for short-term prediction; durations of change of 25.5 minutes for SSEP and 32.5 minutes for MEP). However, OPLL patients or OLF patients had moderate and similar IONM warning thresholds. Additionally, there was a stronger correlation between the SSEP amplitude variability ratio and the JOA RR in OPLL+OLF patients, while the correlation was stronger between the MEP amplitude variability ratio and the JOA RR for the other three TSS pathologies. CONCLUSIONS: Optimal IONM change criteria for prediction vary depending on different TSS pathologies. The optimal monitoring strategy for prediction varies depending on TSS pathologies.