Key result
Random forest and XGBoost models predict reduced LVEF from standard ECGs with an external AUC >0.80.
Why the study?
Deep learning models accurately detect reduced LVEF from ECGs, but their complexity limits clinical use.
Do simplified machine learning models using basic 12-lead ECG parameters accurately detect reduced LVEF (<40%) in patients with and without atrial fibrillation?
Do simplified machine learning models using basic 12-lead ECG parameters accurately detect reduced LVEF (<40%) in patients with and without atrial fibrillation?
Answers are grounded in this paper and cite their sources.
Simplified machine learning models using standard 12-lead ECG parameters and demographics can accurately detect LVEF <40%, providing a highly accessible, web-based screening tool for left ventricular systolic dysfunction.
Background Deep learning (DL) models have shown high accuracy in detecting reduced left ventricular ejection fraction (LVEF) from electrocardiograms (ECGs). However, their complexity limits clinical use. To address this, we aimed to develop and validate simplified machine learning (ML) models using numerical parameters from 12‐lead ECGs to detect LVEF < 40% and to implement them in a user‐friendly web application. Methods We retrospectively analyzed ECG and transthoracic echocardiography data from 21 471 patients across two institutions. The dataset was divided into a development cohort (non‐atrial fibrillation [non‐AF]: n = 12 922; AF: n = 1281) and an external validation cohort (non‐AF: n = 6284; AF: n = 984). Four machine learning algorithms—random forest (RF), extreme gradient boosting (XGBoost), support vector machine, and generalized additive models with LASSO—were evaluated for predicting LVEF as a continuous variable and binary outcome (< 40%). Results For continuous LVEF prediction, RF achieved R 2 values of 0.68 (non‐AF) and 0.74 (AF) in internal validation but performed poorly in external validation. Other models showed R 2 values below 0.40 in internal validation. For binary classification, all models achieved area under the curve (AUC) values > 0.90 in the non‐AF group during internal validation. RF and XGBoost showed strong performance in the AF group (AUC > 0.90 internally) and adequate accuracy externally (AUCs of 0.80–0.81 in AF and 0.90 in non‐AF). Conclusions We developed a simple web‐based tool for preliminary screening of reduced LVEF using 12‐lead ECG parameters.
No takes yet. Share an insight, caveat, or question.
Kawakami et al. (2026) studied this question. Random forest and XGBoost models predicted reduced LVEF from 12-lead ECGs with AUCs >0.90 internally and 0.80-0.90 externally.
Synapse has enriched 10 related papers. Consider them for comparative context: