New AI tool predicts heart attacks years in advance

AI can spot early heart problems that are missed on CT scans reviewed by the human eye, saving thousands of lives every year, a study suggests.

Experts hope the AI technology will help prevent thousands of avoidable deaths from heart attacks. Picture: istock

Thousands of lives could be saved every year by using an artificial intelligence tool that performs better than doctors in predicting whether patients will have a heart attack, a study suggests.

The technology from the University of Oxford, based on 40,000 patients in British hospitals, can spot people with early heart problems that are missed on CT scans reviewed by the human eye.

Those people could start preventive treatment, such as statins, cutting their risk of suffering or dying from a heart attack or stroke in the next decade. The AI tool is being piloted in five hospitals, and researchers hope it can be introduced more widely.

About 350,000 people a year in Britain have a cardiac CT scan, the standard test to identify narrowings or blockages in the coronary arteries. The research, presented at the American Heart Association’s Scientific Sessions, in Philadelphia, found the AI tool could improve treatment for almost half.

Three quarters of people having a CT scan, typically after chest pain, will show no clear sign of narrowing, and patients will often be reassured and discharged. But many will go on to die from a heart attack, because undetectable deposits may break up if the arteries are inflamed and block them.

Researchers trained the AI tool to look at changes in the fat around inflamed arteries, and information on narrowings and other clinical risk factors, to predict a person’s risk of heart attacks or strokes.

CT scans can show no clear sign of significant narrowing of arteries even in patients who are at risk from a heart attack. Picture: Supplied

Among those with no obstructions to their arteries, those with the most inflammation in their blood vessels had a more than tenfold greater risk of dying from heart-related problems than those with the least inflammation.

The AI tool was tested on a further 3,393 patients over almost eight years and was found to predict the heart attack risk accurately. AI-generated risk scores were then sent to doctors for 744 patients, and 45 per cent had their treatment plans altered as a result.

The researchers found the tool was cost-effective compared with standard NHS care, and estimated it could lead to 20 per cent fewer heart attacks and 8 per cent fewer cardiac deaths and strokes among those having the test.

Professor Charalambos Antoniades, who led the research, said: “Providing an accurate picture of risk to clinicians can alter and potentially improve the course of treatment for many heart patients. We hope this AI tool will soon be implemented across the NHS, helping prevent thousands of avoidable deaths from heart attacks every year.”

Professor Sir Nilesh Samani, medical director of the British Heart Foundation, which partly funded the research, said: “Too many people are needlessly dying from heart attacks each year. It is vital we harness the potential of AI to guide patient treatment.”

The NHS is using AI in trial schemes across the country that officials hope will ease pressures this winter. In Buckinghamshire trackers attached to kettles and fridges will spot changes in use that could mean someone is deteriorating, so care teams can step in early.

In Somerset an AI system will look at GP records to identify those with complex needs at risk of hospital admission or who rarely contact their GP. Clinicians will offer them help ranging from food parcels to support to avoid falls.

Amanda Pritchard, chief executive of NHS England, said: “Not only are these measures better for patients who can be cared for at home where they feel most comfortable but they are good for the NHS too, particularly when we know that this winter is likely to be incredibly challenging.”

The Times