AI built on CCTA images can predict heart attacks

Researchers from the University of Oxford have created a new biomarker based off of coronary CT angiography (CCTA) images that can select patients at a high risk of a heart attack five years before they occur.

The development, led by Charalambos Antoniades, MD, PhD, with Oxford’s Division of Cardiovascular Medicine, was published Sept. 3 in the European Heart Journal and presented at the European Society of Cardiology Congress in Paris.

"By harnessing the power of AI, we've developed a fingerprint to find 'bad' characteristics around people's arteries,” Antoniades said in a news release. “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.”

When patients visit the hospital with chest pain, CCTA scans are a normal part of the care routine. If the scan shows no significant narrowing of the coronary artery, patients are typically sent home. However, the authors noted, there isn’t a test or method to help find underlying red flags for future heart attacks.

"Just because someone's scan of their coronary artery shows there's no narrowing, that does not mean they are safe from a heart attack,” Antoniades added.

In order to determine which features most accurately reflect changes to the fat surrounding the heart vessels—perivascular space—the researchers took fat biopsies from 167 patients who underwent heart surgery and matched those images with analyses of gene expressions associated with inflammation, scarring and new blood vessel formation

In order to decipher changes in perivascular space, they then took 101 CCTA scans from patients who had a later heart attack or cardiovascular death within five years of a scan and compared those images with matched controls who did not experience either event. A machine learning technique helped the researchers create the fat radiomic profile (FRP) fingerprint to capture a patient’s level of risk.

After testing its performance in 1,575 people involved in the SCOT-HEART trial, Antoniades et al. found the FRP could accurately predict heart attacks, and performed better than tools currently used by clinicians.

 “This is a significant advance. The new 'fingerprint' extracts additional information about underlying biology from scans used routinely to detect narrowed arteries,” said Metin Avkiran, associate director at the British Heart Foundation, in the same release. “Such AI-based technology to predict an impending heart attack with greater precision could represent a big step forward in personalized care for people with suspected coronary artery disease."

"We genuinely believe this technology could be saving lives within the next year,” Antoniades added.