A study by the European Society of Cardiology (ESC) has shown that an AI-enabled digital stethoscope can more than double the detection of moderate to severe valvular heart disease during routine clinical examinations compared to a traditional stethoscope.
The findings, published in the European Heart Journal – Digital Health, are based on a US prospective, single-arm, single-blinded study evaluating the use of artificial intelligence–assisted auscultation in real-world primary care settings.
The study assessed 357 patients aged 50 years and above with risk factors for heart disease, who were examined using both a conventional stethoscope and an AI-enabled digital stethoscope. Participants were recruited from three primary care centres within the same geographic area. The median age of the cohort was 70 years, and 61.9% of participants were women.
Results showed that the AI-enabled stethoscope achieved a sensitivity of 92.3% in detecting heart sound patterns indicative of moderate to severe valvular heart disease, compared with 46.2% using a traditional stethoscope.
Valvular heart disease affects more than one in two adults over the age of 65 and frequently remains undiagnosed in general practice due to absent, vague, or non-specific symptoms. Undetected disease can impair heart function, limit physical activity, and increase the risk of arrhythmias, heart failure, hospitalisation, and mortality.
The AI-enabled digital stethoscope works by recording high-fidelity heart sounds and analysing them using machine-learning algorithms trained to identify acoustic patterns associated with valvular abnormalities. Traditional auscultation depends largely on clinician experience and hearing and may be influenced by background noise and time constraints. Patients identified as being at risk through AI-assisted screening were referred for confirmatory diagnosis using echocardiography.
The study noted a slight reduction in specificity with the AI-enabled device, which could increase false-positive findings, but researchers indicated this was offset by the benefit of earlier detection and intervention. Increased patient engagement during consultations was also observed when the AI-enabled device was used.
Also Read