Interpretation:

In ED patients with elevated hs-cTn and suspected myocardial injury, 6-month clinical outcomes were similar when initial evaluation and management was assisted by an AI algorithm compared with usual care. The algorithm, based on the Fourth Universal Definition of MI, provided diagnostic probabilities of type 1 versus 2 MI and acute versus chronic myocardial injury, associated prognostic assessments, and evidence-based management recommendations. No increase in short-term adverse events was noted despite lower rates of invasive angiography in patients not classified as type 1 MI, suggesting the relative safety of the tool when used to support independent clinical assessment.

Importantly, although event rates were similar between groups, there were higher rates of evidence-based therapies such as statins and antiplatelet agents in the AI arm. The overall rate of antiplatelet use in type 1 MI-classified patients was still relatively low, which may reflect subsequent reclassification based on additional diagnostic data. Nonetheless, the current data suggest that AI-based decision support tools may reinforce adherence to guideline-directed medical therapies in critical diagnoses such as type 1 MI without an increased risk for harm. It is likely that these AI algorithms will get refined over time. Cost-effectiveness and scalability also need to be evaluated in future studies.