Artificial intelligence improves transfer times and ischemic stroke workflow metrics

Abstract

Introduction Rapid initiation of mechanical thrombectomy (MT) for the treatment of large-vessel occlusion (LVO) critically improves patient outcomes. Artificial intelligence algorithms aid in the identification of LVOs and improve door to puncture times as well as patient transfer times. Objectives We aimed to determine whether the implementation of an LVO detection algorithm that provides immediate active notification to the thrombectomy team provider's cell phone would improve ischemic stroke workflow at our institution and aid in patient transfer from outlying hospitals when compared to our prior system of passive computed tomography perfusion software analysis and radiologist interpretation and notification. Methods A retrospective review of our institutional thrombectomy registry was performed for all patients who underwent MT between January 2020 and March 2022. Demographic, radiographic, and stroke workflow metrics and notification times were collected. Transfer times and stroke metrics were compared pre- and post-implementation of the Viz.ai (Viz.ai, San Francisco, California, USA) smartphone application. Results Two hundred sixty-two patients underwent MT during the study period. Door-to-puncture time decreased 15 min ( p = 0.009) after the implementation of Viz.ai at our Comprehensive Stroke Center. Transfer time from outside hospitals that implemented Viz.ai was reduced by 37 min ( p = 0.04). There was no significant change in transfer time over the same time period in outlying hospitals that did not implement the Viz.ai software. Conclusion Active notification of the neurosurgical team significantly reduces patient transfer time and initiation of MT.

Ähnliche Arbeiten