Prognostic assessment based on the trajectory of artificial intelligence-enabled electrocardiograms for age estimation

Abstract

Abstract Background Artificial Intelligence (AI) estimation of age using data from the 12-lead electrocardiogram (ECG) has demonstrated its utility as a biomarker in predicting the future risk of cardiac diseases and mortality. However, expected long-term outcomes based on temporal changes in the AI-ECG age gap remain unclear. Purpose This study aims to utilize the difference between ages predicted by AI-ECG and actual ages longitudinally for follow-up and to assess whether there are significant clinical differences in mortality among specific groups. Methods Among patients at a hospital who had not been used for the AI-ECG age prediction model (121,702 individuals, 522,261 ECGs), 6,685 individuals with 31,858 ECGs who had an index ECG and the subsequent ECG in the fifth year, with at least one intervening ECG were included for analysis. The absolute age gap, defined as the absolute difference between ECG-predicted and chronological age, was limited to ±10 for analysis. Latent class trajectory modeling was then utilized to classify groups based on the AI-ECG age gap. Results Data-driven analysis revealed four distinct trajectory groups based on age gap trajectories: Consistently Young ECG (27.2%), Reversed ECG Aging (35.3%), Accelerated ECG Aging (8.0%), and Consistently Old ECG (29.5%). Notably, the Accelerated ECG Aging group exhibited a significantly higher cumulative incidence of all-cause death (adjusted hazard ratio (HR) 1.66 (95% confidence interval (CI) 1.30-2.12, p-value <0.001) and non-cardiac death (adjusted HR 1.83 (95% CI 1.38-2.43, p-value <0.001) compared to the Consistently Young ECG group. Conclusion Among 4 distinct trajectories based on the changes in the AI-ECG age gap, Accelerated ECG Aging was associated with a heightened risk of all-cause mortality. Additional in-depth analysis of factors associated with Accelerated ECG Aging is warranted.