Researchers at the University of Oxford, supported by the British Heart Foundation (BHF), have developed a groundbreaking artificial intelligence (AI) technology that could identify individuals at high risk of a fatal heart attack up to five years before it occurs. The findings were presented at the European Society of Cardiology (ESC) Congress in Paris and published in the European Heart Journal.
The innovation centers around a new biomarker known as the fat radiomic profile (FRP), which was developed using machine learning. This ‘fingerprint’ detects biological warning signs in the perivascular space—the area surrounding blood vessels that supply the heart. By identifying inflammation, scarring, and vascular changes, the FRP can signal a heightened risk of future heart attacks.
Currently, patients presenting with chest pain typically undergo a coronary CT angiogram (CCTA) to assess for arterial blockages. However, in approximately 75% of cases, no significant narrowing is found, and patients are discharged—despite some still being at risk of a heart attack. Until now, there has been no routine method to detect these hidden risks.
To develop the FRP, Professor Charalambos Antoniades and his team analyzed fat biopsies from 167 cardiac surgery patients, examining gene expressions linked to inflammation and vascular changes. They then matched these findings with CCTA images to identify key indicators in perivascular fat. The team further validated the FRP by comparing scans from 101 individuals who experienced heart attacks or cardiovascular deaths within five years to those who did not.
The FRP was tested in 1,575 participants from the SCOT-HEART trial and demonstrated superior predictive power compared to existing clinical tools. The researchers aim to make this technology available to healthcare professionals within a year and hope it will be integrated into routine NHS practice alongside CCTA scans within two years.
Professor Antoniades emphasized the significance of the discovery: “Just because a coronary artery scan shows no narrowing doesn’t mean a patient is safe from a heart attack. By leveraging AI, we’ve developed a fingerprint that identifies harmful characteristics around arteries, enabling early detection and prevention.”
Professor Metin Avkiran, Associate Medical Director at the BHF, added: “This is a major advancement. The FRP extracts critical biological data from routine scans, offering a powerful new tool for personalized cardiovascular care.”
