‌Timely detection of structural heart disease (SHD) can enable early intervention and improve patient outcomes. At European Society of Cardiology (ESC) Congress 2025, findings were presented from a prospective study evaluating an artificial intelligence-enhanced 12-lead ECG (AI-ECG) for SHD detection.

The study enrolled 879 outpatients aged ≥45 years with cardiovascular risk factors or ischemic heart disease who were undergoing transthoracic echocardiography (TTE). Clinically significant SHD included left ventricular systolic dysfunction (LV ejection fraction <40%), severe left-sided valvular disease, or severe left ventricular hypertrophy. ECG images were captured with a smartphone and analyzed by the AI-ECG model.

Overall, 37% of participants had one or more SHDs. The AI-ECG model achieved an AUROC of 0.866 for composite SHD, 0.890 for LVSD, and 0.817 for severe valvular disease. The model also reduced the number needed to screen to identify one patient with SHD by 51.4%, indicating improved diagnostic efficiency.

These results suggest AI-ECG can enhance triage and early diagnosis of SHD, particularly in outpatient and resource-limited settings.The AI-ECG approach can guide timely treatment and optimize the use of echocardiography.