A new artificial intelligence (AI) tool has been developed that can detect heart disease from routine electrocardiograms (ECGs) in under two seconds, offering potential for faster diagnosis and treatment of conditions such as heart failure and heart valve disease. The technology, created by researchers at Imperial College London and funded by the British Heart Foundation (BHF), was unveiled recently at the European Society of Cardiology congress in Munich.
ECGs, which record the electrical activity of the heart, are among the most commonly used medical tests globally, with approximately one billion performed annually. While traditionally effective in identifying heart attacks and arrhythmias, they have not been able to diagnose heart failure or valve disease without additional imaging tests such as echocardiograms—ultrasound scans that patients often wait months to receive. The AI model leverages subtle patterns in ECG data invisible to the human eye to identify signs of these conditions rapidly.
Trained on data from 10.6 million ECGs, the AI system was tested on more than 65,000 U.S. patients, correctly identifying up to 81% of those with heart failure and 90% of those with valve disease. While not intended to replace echocardiograms or provide definitive diagnoses, this technology is designed to serve as a fast, initial screening tool, flagging high-risk patients for further testing and prioritising those in urgent need of care.
The AI tool is currently undergoing trials involving 590 NHS patients in London and Bristol, with the goal of nationwide implementation within two years. Experts suggest the adoption of this technology could reduce long waiting times for heart scans, which currently leave nearly 400,000 people in England on cardiology waiting lists.
Dr Sonya Babu-Narayan, clinical director at the BHF and a consultant cardiologist, said the AI’s ability to rapidly identify patients at high risk of heart disease offers a promising way to fast-track diagnosis and treatment, though it will not detect every case. Similarly, Prof Fu Siong Ng, a cardiology professor at Imperial College London, emphasised that the tool could prioritise patients for ultrasound scans more efficiently, potentially saving lives by enabling earlier interventions.
Dr Roy Jogiya, chief medical adviser at Heart Research UK, described the AI as an “extra pair of eyes” supporting clinicians rather than replacing them. He highlighted the importance of larger-scale research to validate the approach further.
In addition to heart disease detection, related research presented at the Munich congress showcased AI's capability to diagnose high blood pressure and type 2 diabetes within seconds by analysing brief facial video recordings. Developed by teams from the University of Tokyo and the Institute of Science Tokyo, this technology demonstrated accuracy rates above 90% for hypertension and 80% for diabetes, pointing to a broader future role for AI in rapid, non-invasive health screening.
Overall, these advancements reflect a growing interest in AI’s potential to enhance early diagnosis in cardiovascular and metabolic diseases, potentially transforming patient care through faster, more accessible testing methods.