Abstract WP39: Subhairline Electroencephalography For Detection Of Large Vessel Occlusion Stroke

Abstract

Introduction: Endovascular thrombectomy (EVT) for large vessel occlusion stroke of the anterior circulation (LVO-a) can only be performed in selected hospitals. A prehospital triage instrument with which paramedics can reliably identify patients with an LVO-a stroke would enable direct routing of these patients to the right hospital and would consequently improve patient outcome. Electroencephalography (EEG) has previously shown to discriminate between patients with and without an LVO-a stroke. In this study, the diagnostic accuracy of subhairline EEG for LVO-a stroke was evaluated. Methods: A 3-minute EEG recording was performed in 35 patients who were presented to the emergency room with a suspected stroke or known LVO-a stroke, as soon as possible after presentation, and prior to potential EVT. Recordings were performed using 9 self-adhesive electrodes placed on the forehead and behind the ears. We evaluated the diagnostic accuracies of EEG features quantifying frequency band power (relative delta, theta, alpha and lower beta powers and power ratios), brain symmetry (pairwise derived Brain Symmetry Index [pdBSI]) and connectivity (Magnitude Squared Coherence [MSC] and Weighted Phase Lag Index) for LVO-a stroke using receiver operating characteristic analysis. Optimal cut-off points were determined as the maximum sensitivity at a specificity of ≥ 80% for LVO-a stroke. Results: Median age was 75 (IQR 67-81) years and 20/35 (57%) patients were male. 13/35 (37%) patients had an LVO-a stroke, 12/35 (34%) a non-LVO-a ischemic stroke, 4/35 (11%) a transient ischemic attack and 6/35 (17%) a stroke mimic. Median onset-to-EEG-time was 250 (IQR 118-535) minutes. Recordings of all patients were of sufficient quality to calculate frequency band power and brain symmetry measures. The connectivity measures could not be calculated for 3 LVO-a stroke patients. The highest diagnostic accuracy for LVO-a stroke was reached by the MSC in the delta frequency band (AUC 0.80; sensitivity 80%; specificity 82%) and the pdBSI in the theta frequency band (AUC 0.78; sensitivity 77%; specificity 82%). Conclusions: Subhairline EEG is a promising instrument for detection of LVO-a stroke, but validation in a larger study population and in the prehospital setting is necessary.