Abstract 271: The Artificial Intelligence Revolution in Stroke Care: A Decade of Scientific Evidence in Review

Abstract

Introduction The role of artificial intelligence (AI) in clinical medicine is ever‐increasing and in time‐dependent, tissue‐based evaluation and treatment of ischemic stroke, AI is clearly proving its utility. In the past decade, scientific publications evaluating the use of AI in stroke care have increased exponentially and information on level of evidence is limited.We aim to systematically review the use of AI in different areas of stroke care, evaluate the level of evidence, and demonstrate the trend of scientific publications in the past decade. Methods A systematic search of electronic scientific databases from 2013 to 2022 was conducted with inclusion of published studies that incorporated AI in all areas of stroke care: prevention, diagnosis, acute care and rehabilitation. Studies published in languages other than English, not directly relevant to stroke care and AI, and any duplicate studies were excluded. Results Our study comprised of 623 studies divided into 101 (16.2%) review studies, 9 (1.4%) meta‐analyses, 140 (22.5%) methodology articles, 2 (0.3%) case reports, 2 (0.3%) case series, 31 (5%) case‐control studies, 277 (44.5%) cohort studies, 16 (2.6%) cross‐sectional studies, and 45 (7.2%) experimental studies. Role of AI in stroke diagnosis was the most frequently explored areas in the published literature (n=275, 44.1%). Our trend analysis showed that the number of studies focusing on AI in stroke significantly increased in the past decade from 1 study in 2012 to 242 studies in 2022 (p=0.0001). Conclusion Majority research focused on role of AI in stroke diagnosis and rehabilitation with more data and input needed in stroke treatment. Although several articles demonstrate AI’s precision and efficacy, utility across the spectrum of stroke care remains to be seen.