'Superhuman' AI tool detects heart disease in under two seconds
Researchers have developed an AI tool that spots heart failure and heart valve disease from ECG results in under two seconds, potentially speeding up diagnosis and prioritising patients for scans.
The technology, presented at the European Society of Cardiology annual congress in Munich, extracts more information from a routine electrocardiogram (ECG) than the human eye can typically see. It was trained on millions of patients and tested in a trial involving 67,000 people in the US, where it identified up to 81% of those with heart failure and up to 90% of those with heart valve disease.
A standard ECG records the electrical activity of the heart, including its rate and rhythm, and has been used for a century to diagnose heart attacks and abnormal rhythms. But it cannot detect conditions such as heart failure or heart valve disease. That typically requires an echocardiogram, a type of ultrasound scan that patients may wait months to receive.
The new AI tool is designed to flag signs of these two conditions from ECG data “in the blink of an eye”, according to the research team led by Imperial College London and funded by the British Heart Foundation (BHF). Early diagnosis is considered vital, as it allows patients who need lifesaving treatment to be identified before becoming dangerously unwell.
Dr Sonya Babu-Narayan, consultant cardiologist and clinical director of the BHF, said technology like this “has the potential to identify high-risk patients early”. She cautioned that it will not detect everyone with a heart condition, but added that it “could be a solution to help fast-track the patients who are most likely to have a heart abnormality”.
The AI tool is not intended to definitively diagnose or rule out heart failure or heart valve disease on its own. Instead, patients judged highly likely to have either condition could be sent rapidly for an echocardiogram, rather than remaining on standard waiting lists. That could mean quicker diagnosis and earlier treatment.
Prof Fu Siong Ng, professor of cardiology at Imperial College London, said patients often wait several months for a heart ultrasound scan after being referred. “This makes it exciting that our technology could identify patients most at risk of heart failure and heart valve disease, so they could be prioritised for scans faster and more urgently.”
The technology could also be used opportunistically, running on all ECGs performed in a hospital to flag patients at highest risk of these diseases, even when those conditions are not initially suspected. “The AI model could be run on all ECGs done in a hospital to flag those at highest risk of these diseases, so that they can be diagnosed earlier,” Ng said.
Dr Ahmed El-Medany, a BHF clinical research fellow who led the analysis, described the tool as a “superhuman AI”. He said the next challenge would be to design handheld AI-led ECG readers for healthcare professionals.
The conference also heard about another development in which AI analysis of five-second facial videos could rapidly detect undiagnosed high blood pressure and type 2 diabetes. Researchers at the University of Tokyo and the Institute of Science Tokyo said millions of people with these conditions are unaware they have them.