Phenotypic Selectivity of Artificial Intelligence-Enhanced Electrocardiography in Cardiovascular Diagnosis and Risk...

Artificial intelligence (AI)-enhanced ECG (AI-ECG) models are often designed to detect specific anatomical and functional cardiac abnormalities. Understanding the selectivity of their phenotypic associations is essential to inform their clinical use. Here, we sought to assess whether AI-ECG models function as condition-specific classifiers or broader cardiovascular risk markers.

We included 4 distinct study populations drawn from both electronic health records and prospective cohort studies. We deployed 6 image-based AI-ECG models: 5 validated models for the detection of left ventricular systolic dysfunction, aortic stenosis, mitral regurgitation, left ventricular hypertrophy, and a composite model for structural heart disease; and 1 negative control AI-ECG model for biological sex. Additionally, we developed 6 experimental models designed to identify noncardiovascular conditions. Diagnosis codes from electronic health records and cohorts were transformed into interpretable phenotypes using a phenome-wide association study framework. We assessed associations of AI-ECG probabilities with cross-sectional phenotypes using logistic regression and with new-onset cardiovascular diseases using Cox regression. Pearson correlation coefficients were calculated to compare phenotypic signatures.

The study included one random ECG from 233 689 individuals (mean age 59±18 years, 130 084 [56%] women) across sites. Each of the 5 AI-ECG models for structural and functional cardiac disorders was more likely to be associated with cardiovascular phenotypes compared with other phenotype groups (odds ratios ranging from 2.16 to 4.41, P<10⁻⁶), whereas the sex model did not show a similar pattern. All AI-ECG models were significantly associated with their respective target phenotype but also showed similar or stronger associations with a broad range of other cardiovascular phenotypes. Phenotypic associations were similar across AI-ECG models trained for different conditions, which was not observed in models for noncardiovascular conditions. Correlation of phenotype association patterns between models was high (0.67-0.96). This pattern was consistent across all models and external data sets and in both cross-sectional and prospective analyses.

Despite being developed to detect specific cardiovascular conditions, AI-ECG models detect the presence and predict the future development of a broad range of cardiovascular diseases with similar propensity. This challenges their role as binary diagnostic tools and instead supports their use as broader cardiovascular biomarkers.