Clinical and artificial intelligence assessed echocardiographic predictors of outcomes after acute myocardial...

Early risk stratification in patients after acute myocardial infarction (AMI) is critical for guiding therapy and resource allocation. While left ventricular ejection fraction (LVEF) is routinely assessed by echocardiography, novel markers offer additional prognostic utility but are not widely assessed due to time constraints or limited expertise. Artificial intelligence (AI) enables rapid, fully automated analysis of echocardiograms, producing standardized, comprehensive measurements. We evaluated the utility of AI-derived echocardiographic parameters on top of clinical variables in predicting outcomes post-AMI.

Consecutive AMI patients undergoing invasive coronary angiogram were included. Echocardiograms were analyzed using Us2.ai software. Independent predictors of one-year all-cause mortality and major adverse cardiac events (MACE) were assessed using Cox regression. Clinical, echocardiographic, and combined models were compared.

Among 1001 patients, aged 64 years (54, 72) and predominantly male (78.1%), 161 (16.1%) died during follow-up. AI-echocardiographic markers independently associated with one-year all-cause mortality or MACE included lower LVEF, greater LV wall thickness, lower LV mass, greater LA area, lower LA reservoir strain and lower aortic valve area. For one-year mortality, the combined model demonstrated superior discrimination compared with the clinical model alone (AUC 0.85 vs. 0.81; p = 0.018). Similarly, for one-year MACE, the combined model showed improved discrimination compared with the clinical model (AUC 0.80 vs. 0.74; p < 0.001) and yielded the lowest Akaike's and Bayesian Informations.

Combining AI-derived echocardiographic parameters, together with traditional clinical risk factors, provides incremental prognostic value post-AMI. AI tools that automate complex assessments accurately and reproducibly may enhance risk stratification.