An AI tool accurately detected structural heart disease from smartwatch single-lead ECGs in a 600-person prospective trial.
Key Details
- 1AI model was trained using 266,000 12-lead ECGs from 110,000 patients and validated on large external cohorts.
- 2Prospective real-world testing was performed with 600 adults using smartwatch-recorded single-lead ECGs.
- 3Algorithm achieved 88% accuracy (AUC) in smartwatch data; sensitivity was 86% and negative predictive value 99%.
- 4Diseases detected included weakened pumping, valve damage, or thickened heart muscle—most commonly found with cardiac ultrasound (echo).
- 5Preliminary results were presented at AHA Scientific Sessions 2025 and have not yet been peer-reviewed.
Why It Matters
Wearable devices leveraging AI may democratize early detection of serious heart conditions, traditionally requiring advanced imaging. This signals a potential paradigm shift in accessible screening workflows, informing radiology and cardiology practice.
Source
EurekAlert
Related News
•EurekAlert
AI Repurposes Routine CT Scans for Osteoporosis Detection
AI algorithms can extract bone density data from routine CT scans to identify osteoporosis, enabling opportunistic screening.
•EurekAlert
AI Outperforms Radiologists in Detecting Hidden Airway Objects on Chest CT
Southampton researchers developed an AI that surpassed radiologists in detecting hard-to-see airway obstructions on chest CT scans.
•EurekAlert
AI Method Automates X-ray Absorption Spectroscopy for Material Analysis
Researchers have developed an AI-based approach to automate and enhance the analysis of X-ray absorption spectroscopy (XAS) data for materials science.