Value of uni-portal arthroscope-assisted multimodal neurophysiological monitoring for predicting neurological injury in transforaminal lumbar interbody fusion

English Keywords:Lumbar spinal stenosis  Arthroscope-assisted uni-portal spinal surgery  Transforaminal lumbar interbody fusion  Intraoperative neurophysiological monitoring  Postoperative sensory deficits  Predictive value

English Abstract:

  【Abstract】 Objectives: To evaluate the association between intraoperative neurophysiological monitoring(IONM) alert events and postoperative sensory deficits(PSD), as detected by somatosensory evoked potentials(SSEPs), transcranial electric motor evoked potentials(TCeMEPs), and electromyography(EMG), during uni-portal non-coaxial spinal endoscopic surgery(UNSES)/arthroscope-assisted uni-portal spinal surgery(AUSS) via transforaminal lumbar interbody fusion(TLIF). Methods: This retrospective study included 98 patients with lumbar spinal stenosis who underwent AUSS-TLIF between September 2024 and September 2025. Baseline characteristics, IONM parameters, and postoperative outcomes were assessed. Binary discriminant analysis and logistic regression models were used to evaluate the predictive ability of IONM parameters for PSD. Receiver operating characteristic(ROC) curves were generated and area under the curve(AUC) was calculated to determine the predictive value and threshold of EMG. Based on a composite warning criterion of EMG activity duration ≥5.45s or MEP amplitude reduction >80%, the clinical performance of the IONM system was evaluated, including sensitivity, specificity, positive predictive value, negative predictive value, and false-negative rate. Results: Binary discriminant analysis showed an explanatory power of 73.5%, and logistic regression showed an explanatory power of 75.5%. The most predictive variables included EMG activity duration(0.795), percentage decrease in MEP amplitude(0.725), and operative time(0.655). ROC analysis identified an EMG activity duration threshold of 5.45s, with the AUC of 0.867(95%CI: 0.795-0.939, P<0.001). The highest sensitivity was observed at a threshold of ≥4.0s(90.2%), and the highest specificity was observed at ≥6.5s(92.4%). Using the composite warning criterion, the IONM system had a sensitivity of 80.0%, a specificity of 84.9%, a positive predictive value of 64.5%, a negative predictive value of 92.5%, and a false-negative rate of only 5.1%. Conclusions: EMG activity duration can serve as a monitoring indicator that provides early warning of intraoperative nerve injury, guiding timely surgical intervention and thereby improving postoperative sensory outcomes.