Valvular & structural heart disease
Artificial intelligence-enhanced electrocardiography to predict regurgitant valvular heart diseases: an international study
Authors' conclusion
This study developed AI-ECG models to diagnose and predict rVHDs and validated the models in a transnational and ethnically distinct cohort.
Study context
- Population
- Patients with paired ECG and transthoracic echocardiogram, Zhongshan Hospital, Shanghai, China; external outpatient set, Beth Israel Deaconess Medical Center, Boston, US988,618 ECG–echo pairs
- Participants
- 400,882 patients
- Comparison
- AI-ECG risk quartiles
- Outcome
- Future moderate or severe regurgitant valvular heart disease (VHD) (MR, TR, AR)
Methodological detailsinternational
- International
Performance
As reported in the abstract
Prediction of future significant rVHD (internal test set)
Highest risk quartile vs Lowest risk quartileHazard ratio (95% CI)P value
MR; age- and sex-adjusted Cox
7.6(5.8–9.9)
P < 0.0001AR
3.8(2.7–5.5)
TR
9.9(7.5–13.0)
- C-index, mitral regurgitation (MR)
- 0.774 (95% CI 0.753-0.792)
- C-index, aortic regurgitation (AR)
- 0.691 (95% CI 0.657-0.720)
- C-index, tricuspid regurgitation (TR)
- 0.793 (95% CI 0.777-0.808)
Further findings
External test set (34 214 outpatients, Boston)
- Findings confirmed in the transnational external test set
Imaging association
- AI-ECG predictions associated with subclinical chamber remodelling
Original publication
Artificial intelligence-enhanced electrocardiography to predict regurgitant valvular heart diseases: an international study. Eur Heart J. Published 16 July 2025. doi:10.
Summarised from the published abstract; every number is checked against it