An AI Tool Can Flag Hidden Heart Disease in Routine ECGs
Electrocardiograms have been used for more than a century, and clinicians perform about a billion of them worldwide every year. They can reveal a heart attack or an abnormal rhythm. An ECG alone cannot establish two conditions that often require an echocardiogram to confirm: heart failure and aortic stenosis, a narrowing of a heart valve.
A research team led by Imperial College London trained an artificial-intelligence system to look for risk signals in those ordinary ECG traces. The project, funded by the British Heart Foundation, produced a result in under two seconds and was presented at the European Society of Cardiology Congress in Munich.
The researchers trained the system on 10.6 million ECGs. They then compared its results with known echocardiogram findings in more than 65,000 recordings from US hospital patients. In that testing, it flagged up to 81 percent of people with heart failure and up to 90 percent of people with aortic stenosis. The figures describe the tool’s performance in that research setting, not a completed diagnosis for an individual patient.
Its proposed job is triage. A strong result could identify people who should be considered sooner for an echocardiogram, the ultrasound scan that can establish what is happening in the heart. Professor Fu Siong Ng, the study’s senior investigator, noted that people can wait several months for such a scan after a doctor’s referral. Earlier triage could help patients at greatest risk reach that next test sooner.
Dr Ahmed El-Medany led the analysis at Imperial College London as a British Heart Foundation clinical research fellow. He said routine ECGs may contain information beyond a conventional reading. The next question is whether the result carries over to the NHS and to simpler portable ECG devices.
That work is underway. The team is testing the system with ECGs from 590 NHS patients in London and Bristol, and estimates that routine use may still be about two years away. More than one million people in the UK live with heart failure, according to the foundation.
Dr Sonya Babu-Narayan, the foundation’s clinical director, described the speed of the reading as “the blink of an eye.” The follow-up scan still establishes the diagnosis. The intended benefit is a quicker route to that scan for patients who may need it most. In heart care, she said, earlier diagnosis and treatment can save and improve lives.
