Efficiency of an artificial intelligence powered electrocardiogram model for early detection of occlusion myocardial infarction in acute coronary syndrome: a multicenter cohort study

Abstract Background Most of the patients with acute coronary syndromes (ACS) present without typical ST elevation. One third of non ST elevation myocardial infarction (NSTEMI) patients have an acutely occluded culprit coronary artery occlusion myocardial infarction (OMI), leading to poor outcomes due to delayed identification and invasive management. Purpose This study evaluated the accuracy and time to diagnosis of OMI by comparing an artificial intelligence (AI) powered electrocardiogram (EKG) model, state of the art diagnostic criteria, and EKG experts in patients with ACS presenting to the emergency department for serial EKG evaluations. Methods AI model was developed using 18 616 EKGs from 10 543 patients with suspected ACS from an international database with clinically validated outcomes. The model was evaluated in an international cohort and compared with STEMI criteria and EKG experts in detecting OMI. In patients with multiple EKGs prior to coronary angiography, a maximum interpretation per patient was retained for the benchmarking. The time to diagnose OMI was noted for each criterion by measuring the duration from the patient's initial EKG to the accurate identification of OMI on subsequent EKGs. If the detection method detected OMI on the first EKG, the time to diagnosis is 0. In cases where the criteria were unable to detect OMI in any EKG before coronary angiography (CAG), the time to diagnosis was considered equivalent to the time to CAG. Results The mean age of the study population was 66 ± 14 years, 65.9% were males; 22.9% patients presented with OMI. The mean time to OMI diagnosis was significantly shorter for the OMI AI model compared with STEMI criteria, 2.3 vs. 5.3 hours, respectively (P0.001), but comparable with EKG experts, with a mean time of 2.9 hours (P=0.8). Patients with OMI received interventions at a similar rate regardless of the presence of STEMI criteria and outcome definition [primary outcome definition, 97.3 vs. 95.9% (P=0.570); strictest OMI outcome (TIMI 0-1 flow only), 96.3 vs. 92.4% (P=0.358) (Figure 1). Conclusion The present novel EKG AI model demonstrates superior accuracy and efficiency to detect acute OMI when compared with STEMI criteria. This suggests its potential to improve ACS triage, ensuring appropriate and timely referral for immediate revascularization.Figure 1