Coronary artery disease diagnosis with signal processing and machine learning of heart sound signals: a systematic review - PubMed

5 hours ago

Coronary artery disease (CAD) is a major cause of morbidity and mortality globally.
Heart sound analysis is explored as a noninvasive method for CAD detection, but evidence has been inconsistent.
A systematic review evaluated the diagnostic performance of heart sound analysis for CAD (≥50% stenosis).
40 studies involving 13,814 participants were included in the review.
Signal processing methods (21 studies) mostly reported diagnostic accuracy below 75%, with limited independent validation.
Machine learning-based methods (19 studies) showed accuracy, sensitivity, and specificity consistently above 80%, with better generalizability due to independent validation.
Using the full heart sound signal as input improved sensitivity compared to using only the diastolic component.
Machine learning-based heart sound analysis shows potential for CAD diagnosis, but larger multicenter studies are needed for clinical validation.

Hasty Briefsbeta