By next year, machine learning and AI could warn high-risk patients of a heart attack at least five years ahead.

Artificial intelligence to detect heart attacks may soon be a reality, according to research at the University of Oxford.

Doctors usually scan chest pain patients for narrow or blocked arteries using a coronary CT angiogram (CCTA). Usually, however, most are sent home because there is no visible narrowing of arteries. However, some individuals among these patients do run the risk of a heart attack in the future. Scientists do not have a method to identify these higher-risk patients.

Artificial intelligence to detect heart attacks

But machines can be taught to identify the tell-tale patterns of inflammation, scarring, and changes to the blood vessels that supply blood to the heart, regardless of any signs of narrowing arteries. These patterns up the risk of the patient suffering a heart attack.

Machine learning is the process of training machines to identify early warning clues. This is an application of artificial intelligence to detect heart attacks.

The researchers at the University of Oxford call this AI method the ‘fat radiomic profile (FRP).

This technology could be functional next year

“By harnessing the power of AI, we’ve developed a fingerprint to find ‘bad’ characteristics around people’s arteries,” said Professor Charalambos Antoniades, BHF Senior Clinical Fellow at the University of Oxford. “This has huge potential to detect the early signs of disease, and to be able to take all preventative steps before a heart attack strikes, ultimately saving lives.”

Artificial intelligence to detect heart attacks also finds application in atrial fibrillation

Computer modeling at the Mayo clinic identified patients who had previous episodes of atrial fibrillation, a condition that causes the heart to beat irregularly, even though the heart is now beating normally.

The computer modeling analyzed tests carried out on nearly 181,000 patients between 1993 and 2017. These patients had normal electrocardiograms at first but later suffered from irregular heart rhythms.

The AI-based training enabled the computer to correctly identify the subsequent diagnosis from the normal test results in 83% of cases.