Rethinking post-infarction remodelling: identification and validation of data-driven cardiac MRI phenotypes for risk stratification

AIMS: Conventional definitions of adverse left ventricular remodeling (ALVR) following ST-elevation myocardial infarction (STEMI), which are based on longitudinal changes in left ventricular end-diastolic volume index and ejection fraction, have demonstrated limited prognostic value. This study aimed to identify prognostic relevant remodeling phenotypes using a data-driven approach on longitudinal cardiac magnetic resonance (CMR) data. METHODS AND RESULTS: In this prospective longitudinal study, a derivation cohort (n=337) and an independent validation cohort (n=190) of patients with reperfused STEMI were analyzed. Patients underwent CMR at <7 days and 6-month follow-up. Unsupervised clustering algorithm was applied using baseline biventricular, biatrial, and infarct parameters, plus their longitudinal changes. Major adverse cardiovascular events (MACE) included all-cause mortality, recurrent MI, and heart failure. Three reproducible post-MI phenotypes were identified: Low-Risk (n=175), Early Remodeling (n=96, severe acute injury with minimal longitudinal change), and Atrial-dominant Remodeling (n=66, bi-atrial and ventricular enlargement). Phenotype-stratified MACE-free survival differed significantly in both cohorts (derivation P<.001; validation P=.002). In the validation cohort, both Early Remodeling (HR 4.73; 95%CI 1.77-12.64) and Atrial-dominant Remodeling (HR 4.02; 95%CI 1.49-10.88) were associated with elevated MACE risk versus the Low-Risk group; the Atrial-dominant phenotype was further associated with heart failure hospitalization (subdistribution HR: 5.92; 95%CI:1.04-33.56). Adding the derived phenotypes to baseline CMR parameters improved MACE discrimination (C-index 0.755 vs 0.714; ΔC 0.041, 95% CI 0.006-0.074), whereas conventional ALVR did not. CONCLUSION: Three reproducible post-MI remodeling phenotypes were identified, which provides superior risk stratification for MACE over conventional ALVR criteria and insights into post-MI heart failure.