DOI: 10.3390/jcdd13110552 ISSN: 2308-3425

Incremental Value of Detailed Electrocardiographic Features for Predicting Left Ventricular Remodeling After Anterior STEMI

Xiaopeng Dong, Wei Zhang, Dongxu Sun, Zhifei Bi, Ximeng Tong, Yuyang Wang, Huiyang Wang, Ruihao Liu, Changhao Sun, Hongji Qian, Jiayue Yang, Yunlong Xia, Xiaolei Yang, Bo Zhang

Left ventricular remodeling (LVR) after anterior ST-segment elevation myocardial infarction (STEMI) is associated with adverse cardiovascular outcomes. We aimed to determine whether detailed electrocardiographic (ECG) features provide incremental predictive value beyond clinical characteristics for LVR after anterior STEMI. We retrospectively analyzed 1079 patients with anterior STEMI, including a derivation cohort of 831 patients (training, n = 665; internal validation, n = 166) and a temporal internal validation cohort of 248 patients. An exploratory cohort of 730 patients with non-anterior STEMI was also included. LVR was defined as a ≥20% increase in left ventricular end-diastolic volume from baseline to 6 months. After feature selection, two extreme gradient boosting models were constructed for risk prediction at hospital discharge: Model 1 included three clinical predictors, whereas Model 2 additionally included eight ECG features. No ECG features were retained in the exploratory non-anterior STEMI analysis. Model 2 had a higher area under the receiver operating characteristic curve (AUC) than Model 1 in both the internal validation cohort (0.869 vs. 0.787; p = 0.039) and the temporal validation cohort (0.792 vs. 0.638; p = 0.001). Model 2 also showed greater estimated net benefit in decision curve analysis and improved risk reclassification. Detailed ECG features may provide additional prognostic information beyond the clinical predictors examined in this study. The web-based calculator is intended for research use only, pending external validation.

More from our Archive