DiagnosticPrevention AI-enabled ECGs for early detection of health risks

A group of investigators from Imperial College London and Imperial College Healthcare NHS Trust, including a number of NIHR Imperial BRC-supported researchers, have developed an innovative artificial intelligence (AI) system that has the potential to revolutionise how we predict health risks using electrocardiograms (ECGs).

The researchers believe their work could be used in the NHS within five years. It would enable doctors to find the disease earlier and prioritise the most urgent cases for treatment.

An electrocardiogram (ECG) records the electrical activity of the heart and is one of the most common medical tests in the world.

The team used very large sets of data from international sources, consisting of millions of ECGs previously taken as part of routine care, to train their AI model to analyse an ECG and accurately predict which patients went on to experience new or worse diseases, or who subsequently died.

ECGs depict the flow of electrical signals within and between the heart’s chambers—the atria and ventricles. The AI model was trained to ‘read’ that information and see patterns in the electrical signals. Researchers say the model can see and understand ECG patterns with more complexity and subtlety than a cardiologist can.

Dr Arunashis Sau, a British Heart Foundation (BHF) Clinical Research Fellow at Imperial College London’s National Heart and Lung Institute (NHLI), and cardiology registrar at Imperial College Healthcare NHS Trust, led the research. He explained: “We cardiologists use our experience and standard guidelines when we look at ECGs, sorting them into ‘normal’ and ‘abnormal’ patterns to help us diagnose disease. However, the AI model detects much more subtle detail, so it can ‘spot’ problems in ECGs that would appear normal to us, and potentially long before the disease develops fully.”

The AI model – known as AI-ECG risk estimation, or AIRE – was able to correctly identify the risk of death in the ten years following the ECG (from high to low) in 78% of cases. For the remainder of cases where the model was incorrect, researchers believe that was influenced by unknowable factors, such as subsequent treatment of the patient or an unpredicted cause of death.

The system can predict future health risks such as heart rhythm problems, heart attacks and heart failure, as well as when someone would die from a non-heart related cause. The researchers found it could predict these risks with a high level of accuracy.

Dr Sau further explained: “ECGs capture a great deal of information from around the body because diseases such as diabetes, that affect organs like the kidneys or liver, will also affect the heart in some way. Our analysis shows that the AI can tell us a lot about not only the heart but also what is going on elsewhere in the body and may be able to detect accelerated ageing.”

AI-enhanced ECGs are already known to be highly accurate in diagnosing heart diseases but have not previously been used to tell clinicians about an individual patient’s risk of developing a range of specific, treatable diseases in the future. AI-enhanced ECGs are not currently part of routine hospital care or diagnosis.

The researchers also analysed imaging and genetic information, which helped them confirm that the AI predictions were linked to real biological factors in the heart’s structure and function. This is something they say is crucial for the credibility of the model with clinicians, showing it can pick up subtle changes in the heart’s structure over time, which are early signs of risk of disease or death.

The senior author of this study is Dr Fu Siong Ng, Reader in Cardiac Electrophysiology at the NHLI at Imperial College London and a consultant cardiologist at Imperial College Healthcare NHS Trust and Chelsea and Westminster Hospital NHS Foundation Trust.

He said: “Our work has shown that this AI model is a credible and reliable tool that could, in future, be programmed for use in different areas of the NHS to provide doctors with relevant risk information. This could have a positive impact on how patients are treated, and ultimately improve patient longevity and quality of life. It could also reduce waiting lists and allow more efficient allocation of resources. We believe this could have major benefits for the NHS, and globally.”

He added: “The important next step is to test whether using these models can improve patient outcomes in clinical studies.”

Trials of AIRE in the NHS are already planned in hospitals across Imperial College Healthcare NHS Trust and Chelsea and Westminster Hospital NHS Foundation Trust. These clinical trials will focus on evaluating the benefits of implementing the model with real patients and will start by mid-2025. Patients will be recruited from outpatient clinics and also from the inpatient medical wards.

Professor Bryan Williams, Chief Scientific and Medical Officer at the British Heart Foundation, which funded the research, said: “This large, exciting study offers a glimpse into how AI could be used to improve diagnosis of heart disease. ECGs have been used to assess the heart for over a century, and this research has demonstrated the extraordinary power of AI to gain important health insights from a routine test. This could take the use of ECGs beyond what has previously been possible, by helping assess the risk of future heart and health problems, as well as risk of death.

“We look forward to seeing how AI can be piloted in routine clinical practice and how this will help accelerate and inform clinical decision-making, ensuring patients receive the most timely and effective treatment and support.”

Dr Sau added: “We now need to see how the model performs in a real healthcare system. But in the future, it’s a possibility that we could see patients provided with wearable technology that provides doctors with continuous remote monitoring and a potential alert system.”