Use of Artificial Intelligence to Detect Cardiac Rhythm Disturbances in Athletes: A Scoping Review.

Abstract

BACKGROUND: Artificial intelligence (AI) is increasingly used to enhance electrocardiogram (ECG) interpretation in human medicine. In equine athletes, exercise-associated arrhythmias are common and linked to sudden cardiac death at rates higher than in humans. However, ECG interpretation in horses remains time-consuming and subjective, with the clinical relevance of mild rhythm disturbances often unclear. OBJECTIVES: Evaluate the application of AI to ECG interpretation for arrhythmia detection, with emphasis on current and potential use in athletic species, particularly horses. ANIMALS: About 17 studies were included: 13 involving humans, 3 in horses, and 1 in dogs. METHODS: A scoping review of relevant, peer-reviewed studies published between 2000 and 2024 was conducted to identify research applying AI to ECG interpretation for arrhythmia detection. Studies were assessed for species, AI model type, diagnostic accuracy, and relevance to ECGs recorded during exercise. Primary outcomes included arrhythmia detection performance and applicability to veterinary medicine. RESULTS: Deep learning models, including convolutional neural networks, achieved accuracies ranging from 79.4% to 98.6% in studies of humans. Research in horses showed encouraging results using restitution analysis and transfer learning approaches. However, small sample sizes and species-specific ECG morphology remain major limitations to broader application in veterinary medicine. CONCLUSION AND CLINICAL IMPORTANCE: Artificial intelligence holds promise for enhancing the accuracy and efficiency of arrhythmia detection in ECGs of equine athletes. Development of species-specific algorithms may facilitate real-time monitoring of cardiac function during exercise, supporting improved cardiovascular assessment in athletic horses.