Key result
AI-enabled ECG identifies left ventricular diastolic dysfunction with ~0.91 AUROC.
Why the study?
Does a deep-learning model applied to standard 12-lead ECGs accurately identify guideline-defined left ventricular diastolic dysfunction in patients undergoing paired ECG and echocardiographic examinations?
Observational (n=21,385)
No
Does a deep-learning model applied to standard 12-lead ECGs accurately identify guideline-defined left ventricular diastolic dysfunction in patients undergoing paired ECG and echocardiographic examinations?
An AI-enabled 12-lead ECG model accurately screens for and rules out guideline-defined left ventricular diastolic dysfunction with a 96.6% negative predictive value, offering a scalable triage tool to identify patients who warrant confirmatory echocardiography for early HFpEF recognition.
May support ECG screening for diastolic dysfunction; hypothesis-generating and requires prospective validation before adoption.
No abstract yet. The journal hasn't published one we can show.
No takes yet. Share an insight, caveat, or question.
Park et al. (2026) conducted an observational in Left ventricular diastolic dysfunction (LVDD) and early heart failure with preserved ejection fraction (HFpEF) (n=21,385). Deep-learning AI-enabled 12-lead electrocardiography (AI-ECG) vs. Echocardiography-defined LVDD (2016 ASE/EACVI recommendations) was evaluated on Identification of guideline-defined left ventricular diastolic dysfunction (LVDD) (AUROC 0.913 (sensitivity 83.8%, specificity 82.5%, NPV 96.6%)). AI-enabled 12-lead ECG detected guideline-defined left ventricular diastolic dysfunction with an AUROC of 0.913, 83.8% sensitivity, 82.5% specificity, and 96.6% negative predictive value.
Synapse has enriched 5 related papers. Consider them for comparative context: