Abstract WP117: Neurophysiological Detection of Stroke and Large Vessel Occlusions (LVOs)

Abstract

Introduction: Approximately 1.9 million neurons are lost with each passing minute of a cerebral large vessel occlusion (LVO). Proper triage is imperative for transfer to an appropriate center and further care. EEGs and SSEPs are the standard of care for intraoperative detection of stroke. Asymmetric changes in electrographic activity can identify neural ischemia with high accuracy. In this study, we tested the predictive power of a brain functional symmetry index (BFSI) to identify stroke with EEG and SSEP. Methods: We conducted a cross-sectional study utilizing ischemic stroke patients admitted to two university hospitals. EEG results were recorded by leads placed on the head. Evoked potentials were acquired via electrodes providing somatosensory stimulations bilaterally to the wrists. Patients were further categorized into LVOs versus non-LVOs. During analysis, measures were taken to also account for small sample size of both stroke and LVO identification. The predictive power of the algorithm to identify stroke and LVO were examined. Results: A total of 38 patients were enrolled of which 19 patients had ischemic strokes. Of the strokes, 8 were LVOs. Demographics of control and stroke patients were analyzed using t-test and chi-squared analysis. BFSI correctly classified 17/19 strokes (89%) and 17/19 controls (89%). Similar results were obtained when predicting for larger sample sizes with 84% sensitivity, 79% specificity, 80% positive predictive value, and 83% negative predictive value. Of the LVOs, the BFSI correctly identified all LVOs (100% sensitivity) and produced a specificity of 57%. Again, similar results were demonstrated when accounting for larger sample size with sensitivity of 100% and specificity of 53%. Conclusion: The predictive power of BFSI demonstrates high accuracy for ischemic strokes, especially LVOs. These results not only demonstrate a dependable means of detecting stroke in a controlled environment, it also serves as a blueprint for rapid detection of LVOs in the prehospital setting.