Development and validation of a scale for the assessment of left ventricular remodeling phenotypes in ambulatory patients with heart failure following acute myocardial infarction

Background . Predicting the progression of heart failure in patients after acute myocardial infarction, despite modern therapy, remains challenging, as relying solely on left ventricular ejection fraction does not capture the full spectrum of structural and functional myocardial changes. Existing phenotyping approaches that enable more precise risk stratification are often cumbersome and impractical in routine ambulatory practice. Therefore, there is a need for a simple and accessible tool based on standard echocardiographic parameters. Objective . To develop the Left Ventricular Remodeling Phenotype Assessment Scale (ШОФР / R-PAS) in patients with heart failure after acute myocardial infarction based on cluster analysis of five standard echocardiographic parameters, to assess the distribution of phenotypes obtained using the scale in the training and independent cohorts and their prognostic significance for major adverse cardiovascular events. Material and methods . This retrospective cohort study included two stages: development of the ШОФР / R-PAS scale in a training sample of 105 patients with heart failure after acute myocardial infarction (AMI, stage 1), and external validation in an independent cohort of 103 patients (stage 2). Data were obtained from electronic medical records of the Unified Medical Information and Analytical System of Moscow. The scale was developed based on k-means cluster analysis of five echocardiographic parameters (left ventricular ejection fraction (LVEF), end-diastolic and end-systolic volumes (EDV, ESV), left atrial size (LA), left ventricular myocardial mass index (LVMMI)) using sex-specific threshold values. The prognostic significance of phenotypes for two-year risk of major adverse cardiovascular events (MACE) was assessed in the combined sample (n=208) using Cox models, discrimination analysis (C-index), calibration, and risk reclassification metrics (NRI, IDI). Results . Three left ventricular remodeling phenotypes were identified through cluster analysis: phenotype 1 ("preserved geometry"), phenotype 2 ("moderate remodeling"), and phenotype 3 ("advanced remodeling"). Sex-specific threshold values for LVEF, EDV, ESV, LA size, and LVMMI were incorporated into the ШОФР / R-PAS scale, allowing phenotype assignment without repeated cluster analysis. In the independent cohort, the proportions of phenotypes 1, 2, and 3 were 40.8 %, 33.0 %, and 26.2 %, respectively, with no statistically significant difference compared with the training cohort (p=0.675). First MACE occurred in 46 patients over 2 years; phenotype 3 was associated with a threefold increase in risk compared to phenotype 1 (HR 3.11; 95 % CI 1.54–6.25; p=0.001) and remained an independent predictor after adjustment for clinical factors (HR 2.23; 95 % CI 1.07–4.66; p=0.03). Adding the phenotype to the baseline clinical model improved its predictive capacity: prognostic accuracy (C-index) increased from 0.66 to 0.69, representing a clinically significant gain. Reclassification (NRI = 0.51) and discrimination improvement (IDI = 0.031) metrics confirmed that the phenotype helps more accurately separate patients with high and low risk, particularly through better identification of those who will not experience events. Conclusion . The ШОФР / R-PAS scale is a reproducible and prognostically significant tool for phenotypic risk stratification in patients with heart failure after AMI. Phenotype 3 ("advanced remodeling") is associated with an independent increase in MACE risk, and adding the phenotype to the clinical-anamnestic model improves discrimination and risk reclassification. The scale is based on routine echocardiographic parameters and can be used in clinical practice without specialized software.