Abstract Background and aims Timely detection of acute ischemic stroke is crucial for ensuring an appropriate standard of care and guiding reperfusion therapies. While MRI remains highly sensitive for identifying early ischemic changes, its interpretation is subject to inter-observer variability and may be less accessible or efficient in resource-limited settings. Emerging AI models demonstrate high specificity, sensitivity, and accuracy in detecting stroke lesions, offering the potential to enhance diagnostic efficiency and streamline clinical workflows in stroke management. Methods A systematic review of acute stroke diagnostic-focused AI studies from January 2014 to September 2025 in PubMed, MEDLINE, CENTRAL, Google Scholar, and Cochrane Library using terms: ’artificial intelligence’ or and ’ischemic stroke’ and ‘DWI’ and ‘FLAIR’ was performed. The overall certainty of the evidence was evaluated using the GRADE approach. Results AI models demonstrate high accuracy, sensitivity, and specificity in detecting acute ischemic lesions on MRI, particularly in cases of large infarcts, thereby supporting timely decision-making for patients eligible for reperfusion therapies. Conclusions AI models enable rapid and accurate identification of stroke lesions, serving as a valuable adjunct in the early detection of ischemic changes. Conflict of interest Nothing to disclose
Guerra et al. (Fri,) studied this question.