Key result
Adding ECG features to clinical predictors significantly improves prediction of post-STEMI left ventricular remodeling.
Why the study?
Left ventricular remodeling after anterior STEMI is associated with adverse cardiovascular outcomes, and whether detailed ECG features provide incremental predictive value beyond clinical characteristics remained to be determined.
Does adding detailed electrocardiographic features to a clinical risk prediction model improve the prediction of left ventricular remodeling at 6 months in patients with anterior STEMI?
Cohort (n=1,809)
Does adding detailed electrocardiographic features to a clinical risk prediction model improve the prediction of left ventricular remodeling at 6 months in patients with anterior STEMI?
Incorporating eight detailed ECG features into a clinical machine learning model at hospital discharge significantly improves the prediction of 6-month left ventricular remodeling in patients with anterior STEMI.
Answers are grounded in this paper and cite their sources.
ECG features add no incremental LVR prediction value beyond clinical and biochemical data after anterior STEMI; leaves open their role in prospective risk models.
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.
No takes yet. Share an insight, caveat, or question.
Dong et al. (2026) conducted a cohort in ST-segment elevation myocardial infarction (STEMI) (n=1,809). Extreme gradient boosting model combining clinical predictors and 8 detailed ECG features (Model 2) vs. Extreme gradient boosting model with 3 clinical predictors alone (Model 1) was evaluated on Left ventricular remodeling (≥20% increase in left ventricular end-diastolic volume from baseline to 6 months) (AUC 0.869 vs 0.787 (internal validation); AUC 0.792 vs 0.638 (temporal validation), p=0.039 (internal validation); 0.001 (temporal validation)). Adding 8 ECG features to clinical predictors improved 6-month LVR prediction after anterior STEMI in internal (AUC 0.869 vs 0.787; p=0.039) and temporal validation (0.792 vs 0.638; p=0.001).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: