Machine Learning Model Identifies High-Risk Vasospasm After aSAH

Background: Cerebral angiographic vasospasm (AV) can occur in approximately one-third of patients with aneurysmal subarachnoid hemorrhage (aSAH) and is associated with high morbidity and mortality. Transcranial Doppler (TCD) is a noninvasive and highly sensitive method for detecting AV; however, it may not identify vasospasm until patients become symptomatic from hypoperfusion or develop completed ischemic stroke. Objective: To develop a machine learning model to stratify aSAH patients according to their risk of developing AV using longitudinal TCD data rather than static daily values. Design: Single-center retrospective analysis. Methods: Patients admitted with aSAH between 2015 and 2019 were included. Patients with ruptured aneurysms who survived ≥7 days after admission were included. Delayed cerebral ischemia (DCI) was defined as a neurological decline of ≥2 points in Glasgow Coma Scale score or new focal neurological deficits not explained by intracranial hemorrhage, seizur more...