Artificial intelligence for heart attack prediction

When someone has chest pain that may be due to coronary heart disease, they usually have a CT (computerised tomography) scan of their coronary arteries to check for any narrowed or blocked areas. About three quarters of the time, the scan finds no significant narrowing of the arteries. But, despite this, some of these people do have coronary heart disease, which could lead to a heart attack at some point in the future. At the moment, there are no tools used routinely by doctors that can spot which of these patients will go on to have a heart attack and who could benefit from life-saving preventive treatments.

BHF Senior Fellow Professor Charalambos Antoniades and his team at the University of Oxford have recently developed a new technology by using machine learning (a type of artificial intelligence) and applying it to routine CT scans. They have found a specific combination of changes that can be seen on the CT scan, which reveal areas of inflammation and scarring in the fatty tissue immediately around the coronary arteries. They have shown that detecting inflammation and scarring significantly increases our ability to predict a future heart attack up to five years before it strikes.

This new technology has huge potential to detect the early warning signs of future heart attack from a routine CT scan, so that people can receive the right preventive treatment, saving lives. The technology has attracted outside investment, enabling Professor Antoniades to set up a CT image analysis company, called Caristo Diagnostics, to make his findings widely available. CaRi-Heart®, developed by Caristo Diagnostics, has received it’s CE mark accreditation in 2021 and the technology is already available in 15 NHS trusts, including hospitals in Oxford, Leicester and London.

Updated 2nd September 2021