Predictive Value of Intraoperative Neuromonitoring in Brainstem Cavernous Malformation Surgery

OBJECTIVE: To evaluate the predictive value of intraoperative neuromonitoring (IONM) in brainstem cavernous malformation (BSCM) surgery. METHODS: Surgically treated patients with BSCM were included. All patients received IONM consisting of motor evoked potentials (MEPs) and somatosensory-evoked potentials (SSEPs). Neurologic examination was conducted preoperatively and at discharge and follow-up >3 months after BSCM removal. Demographic, radiographic, and clinical features were assessed. Study end points were new motor or somatosensory deficits and functional disability. RESULTS: A total of 62 patients were included. MEP decrease was associated with new motor deficits at discharge (P = 0.022), and SSEP decrease was associated with new somatosensory deficits at discharge (P < 0.001) and follow-up (P < 0.001). Sensitivity and specificity values for MEPs (discharge: 31% and 93%; follow-up: 33% and 91%) and SSEPs (discharge: 82% and 80%; follow-up: 85% and 79%) were calculated, respectively. Receiver operating characteristic analyses with area under the curve (AUC) metrics revealed acceptable performance of MEPs (AUC, 0.75; P = 0.022) and SSEPs (AUC, 0.72; P = 0.004) in predicting early deficits. Intraoperative decrease of MEPs (P = 0.047) and SSEPs (P = 0.017) was associated with early functional disability. Surgery-related subdural air accumulation impaired IONM reliability in predicting early (P = 0.048) and long-term (P = 0.013) deficits. CONCLUSIONS: Established IONM warning criteria may be valid for BSCM removal. However, surgical approaches in the sitting position significantly limit the predictive value of IONM, to some extent because of intraoperative pneumocephalus.