Why the study?
Previous computer audition-based methods for CAD detection focused on analyzing and modeling heart sound data while overlooking practical application scenarios.
Does an AI-assisted portable heart sound sensor with a lightweight model accurately diagnose coronary artery disease?
Population
41 CAD patients and 22 non-CAD healthy controls
Comparison
CAD patients vs non-CAD healthy controls
Key result
The TYKDModel deployed on a portable heart sound sensor achieved a classification accuracy of 85.2%, specificity of 88.6%, and sensitivity of 82.8% for detecting coronary artery disease.
Authors
Loading...
May enable pervasive CAD screening via heart sounds; leaves open clinical utility pending prospective validation.
Cross-Sectional (n=63)
Does an AI-assisted portable heart sound sensor with a lightweight model accurately diagnose coronary artery disease?
A novel, lightweight AI-assisted portable heart sound sensor demonstrates promising diagnostic accuracy (85.2%) for detecting coronary artery disease with low computational requirements.
Zhang et al. (2025) conducted a cross-sectional in Coronary Artery Disease (CAD) (n=63). TYKDModel on a portable heart sound collection device vs. Non-CAD healthy controls was evaluated on Classification accuracy for CAD detection. The TYKDModel deployed on a portable heart sound sensor achieved a classification accuracy of 85.2%, specificity of 88.6%, and sensitivity of 82.8% for detecting coronary artery disease.