Artificial intelligence-augmented electrocardiogram for heart failure diagnosis and risk prediction: a narrative review

The electrocardiogram (ECG) is a widely available, low-cost diagnostic tool routinely used in cardiovascular care; however, its conventional interpretation has limited sensitivity for detecting heart failure (HF) and predicting disease progression. Artificial intelligence (AI) has emerged as a promising approach to enhance ECG interpretation by identifying latent patterns associated with cardiac dysfunction. This review summarizes the diagnostic performance and clinical utility of AI-augmented ECG models for HF diagnosis and risk prediction based on recent primary studies.