Prime Highlights:
- An AI system developed by the University of Cambridge correctly identified 98% of severe aortic stenosis cases and 94% of severe mitral regurgitation cases, outperforming general practitioners.
- The AI could enable faster, more reliable screening for heart valve disease, potentially saving thousands of lives by flagging patients who need further testing.
Key Facts:
- The study tested nearly 1,800 patients across five NHS Trusts, using heart sounds recorded with digital stethoscopes and echocardiograms as the diagnostic reference.
- Heart valve disease affects over half of people aged 65 and above, but many cases remain undiagnosed until symptoms become severe.
Background:
Artificial intelligence could help doctors spot serious heart valve disease much earlier, potentially saving thousands of lives, according to a new study led by researchers at the University of Cambridge.
The research team tested an AI system that analyzes heart sounds recorded using digital stethoscopes. The study included nearly 1,800 patients across five NHS Trusts. Each patient also underwent an echocardiogram, which doctors used as the reference standard for diagnosis.
The AI showed strong results. It correctly identified 98% of patients with severe aortic stenosis, the most common valve disease that requires surgery. It found 94% of severe mitral regurgitation cases, a condition where a heart valve doesn’t close properly and lets blood flow backward.
When compared with 14 general practitioners who listened to the same heart sound recordings, the AI performed better in every case. While GP assessments varied widely, the AI delivered consistent and reliable results, especially for severe disease.
Heart valve disease is common among older adults and often develops without clear symptoms. Experts say more than half of people over 65 are affected, and many remain undiagnosed until the condition becomes critical. Once symptoms appear, the risk of death can rise sharply if treatment is delayed.
Doctors currently rely on echocardiography to diagnose the disease, but it costs a lot, takes time and is not available for routine screening. Short GP appointments and less use of stethoscopes cause many cases to be missed.
Unlike traditional tools, the AI learned directly from echocardiogram results instead of heart murmurs. This lets it find subtle sound patterns that doctors might miss, even when there is no clear murmur.
The researchers stress that the technology is not meant to replace doctors. Instead, it could help GPs quickly decide which patients need further testing. The test takes only seconds and requires minimal training.
Further real-world trials are planned before wider rollout.
