Abstract Background and aims Timely diagnosis and endovascular treatment (EVT) are crucial for favorable outcomes in acute ischemic stroke. In regions with long distances and variable access to neurointerventional expertise workflow delays remain a challenge. Artificial intelligence (AI) based large vessel occlusion (LVO) detection on CT angiography may support and streamline workflows, reduce delays, and support equitable care. We evaluated the impact of implementing an AI-based stroke imaging support system on workflow metrics and clinical outcomes within a regional stroke network. Methods In this prospective, observational multicenter pilot study, EVT-treated patients from one comprehensive stroke center and affiliated hospitals in South-Eastern Norway were included. Patients were grouped as pre-AI (n=90), post-AI without AI use (n=80), and post-AI with active AI use (n=30). The primary endpoint was time from local CT acquisition to CSC contact. Secondary endpoints included CT-to-CSC arrival, CT-to-groin puncture, CSC arrival-to-groin puncture, and functional independence (modified Rankin Scale 0–2) at 90 days. Results Median CT-to-CSC contact times were 24, 29, and 22 minutes in the three groups (global p=0.039), with no significant adjusted pairwise differences. CT-to-groin puncture time was shorter in the AI-used group (152 minutes) compared with both pre-AI (179 minutes) and post-AI without use (182 minutes) (adjusted p=0.045 and p=0.007). Functional outcomes at 90 days did not differ between groups. Conclusions AI-based stroke imaging support was associated with modest workflow improvements but no measurable effect on functional outcome. Limited uptake and lack of automated alerting likely reduced impact, highlighting the importance of full clinical workflow integration. Conflict of interest Linn Heitmann: nothing to disclose
Building similarity graph...
Analyzing shared references across papers
Loading...
Linn Heitmann
Thor Håkon Skattør
Brian Anthony Enriquez
European Stroke Journal
Oslo University Hospital
Barnes Hospital
Building similarity graph...
Analyzing shared references across papers
Loading...
Heitmann et al. (Fri,) studied this question.
www.synapsesocial.com/papers/69fd7ef7bfa21ec5bbf07554 — DOI: https://doi.org/10.1093/esj/aakag023.1108