Neurophysiological Monitoring During Arteriovenous Malformation Embolization

Abstract

Abstract BACKGROUND Neurophysiological monitoring (NPM) is frequently performed during arteriovenous malformation (AVM) embolization. However, the ability of NPM to predict neurological deficits or improve surgical decision making in this setting has not been studied. OBJECTIVE To review our use of NPM during AVM embolization to better define its utility. METHODS We retrospectively examined AVM embolization cases from 2004 to 2017. We recorded patient and AVM characteristics as well as outcomes. We then reviewed NPM results from each case, including somatosensory evoked potentials and electroencephalogram. Our primary outcome was postoperative neurological deficit, and secondary outcomes were discharge and 30-d modified Rankin Score (mRS). RESULTS There were 173 embolizations in 74 patients. Mean patient age was 40 yr. There were 8 (5%) transient and 2 (1.3%) permanent neurological complications. Among those with neurological complications, 3 had NPM changes during the operation (positive predictive value [PPV] = 50%). This improved to 67% for permanent NPM change. Three patients had NPM changes but did not suffer clinical deficits postoperatively (negative predictive value = 90%). The predictive value of the test was improved for discharge but not 30-d mRS, and the test performance improved dramatically with increased pretest probabilities (likelihood ratio [LR](+) = 14.5, LR(–) = 0.715). CONCLUSION We present a large series of AVM embolization operations performed with NPM. The PPV of NPM changes was moderate but improved dramatically with increased pretest probabilities. The rate of permanent neurological complications was among the lowest reported in the literature, suggesting NPM may lead to improved intraoperative decision making.