Key result
AI-enhanced auscultation detects significant VHD with an AUROC of 0.83, outperforming GPs.
Why the study?
Over half of valvular heart disease cases remain undiagnosed with traditional stethoscope screening, while current AI tools rely on murmur detection and lack sensitivity for common subtypes like mitral regurgitation.
Does an AI-enhanced auscultation algorithm improve the detection of clinically significant valvular heart disease compared to general practitioners in adult patients?
RCT (n=1,767)
stratified minimisation algorithm
Yes
Does an AI-enhanced auscultation algorithm improve the detection of clinically significant valvular heart disease compared to general practitioners in adult patients?
Effect estimate: AUROC 0.83; sensitivity 72% vs 62% for GPs; specificity 82% vs 64% for GPs (95% CI AUROC 0.79–0.88; sensitivity 65–79%; specificity 74–89%; GP sensitivity 55–70%, GP specificity 54–73%)
Absolute Event Rate: 72% vs 62%
p-value: p=sensitivity p=0.01; specificity p=0.002
An AI-enhanced stethoscope algorithm trained directly on echocardiographic data accurately detects clinically significant valvular heart disease, outperforming general practitioners and offering a scalable screening tool.
May support AI stethoscope screening in primary care; leaves open prospective validation before practice change.
Valvular heart disease (VHD) is a growing public health concern, yet over half of cases remain undiagnosed due to late symptom onset, limited public awareness, and low sensitivity of traditional stethoscope-based screening. Current AI-enabled tools rely on murmur detection as a proxy for VHD but lack sensitivity for common subtypes like mitral regurgitation and are limited by small datasets. This study presents a novel neural network that directly predicts clinically significant VHD from stethoscope recordings, trained using echocardiographic targets rather than heart murmur labels. A diverse dataset of 1767 patients across UK primary care and hospital settings was developed, combining stethoscope recordings with echocardiographic labels. The trained recurrent neural network achieved an AUROC of 0.83, outperforming general practitioners and demonstrating exceptional sensitivity for severe aortic stenosis (98%) and severe mitral regurgitation (94%). This algorithm shows promise as a scalable, low-cost screening tool, enabling earlier diagnosis and timely referral for intervention. This research was registered with ClinicalTrials.gov (CAIS: NCT04445012 registered on 2020-06-21, DUO-EF: NCT04601415 registered on 2020-10-19).
No takes yet. Share an insight, caveat, or question.
McDonald et al. (2026) conducted an RCT in Adults aged >16 years with and without clinically significant degenerative valvular heart disease attending UK primary care and hospital settings (n=1,767). AI-enhanced auscultation using a recurrent neural network algorithm ('VHD Detector') analyzing electronic stethoscope recordings vs. General practitioners auscultation assessment using electronic stethoscope recordings was evaluated on Detection of clinically significant valvular heart disease (moderate or severe VHD) confirmed by echocardiography (AUROC 0.83; sensitivity 72% vs 62% for GPs; specificity 82% vs 64% for GPs, 95% CI AUROC 0.79–0.88; sensitivity 65–79%; specificity 74–89%; GP sensitivity 55–70%, GP specificity 54–73%, p=sensitivity p=0.01; specificity p=0.002). The AI-enhanced auscultation algorithm detected clinically significant valvular heart disease with an AUROC of 0.83, sensitivity of 72%, and specificity of 82%, significantly outperforming GPs who had 62% sensitivity and 64% specificity.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: