This revolutionary AI technology can identify invisible heart disease threats

Oxford University spin-out company Caristo Diagnostics has developed an AI model that can identify individuals at risk of having a heart attack within the next ten years. BBC report.

Scientists call the model revolutionary because it can identify heart inflammation that computed tomography, or CT scans (a combination of X-rays and computer technology), cannot distinguish.

A pilot scheme is currently being run at five hospital trusts in Oxford, Milton Keynes, Leicester, Liverpool and Wolverhampton with support from NHS England. The scheme has been put forward and a decision will be made in the coming months on whether it is suitable for the NHS.

They are also using the technology to prevent diabetes and stroke. Professor Keith Channon from the University of Oxford said: “This is a transformative, game-changing technology because for the first time we can monitor biological activity that is invisible to the naked eye, which occurs before these narrowings and blockages (in the heart) form.”

In the pilot, patients with chest pain who were referred for a standard CT scan had their scans run through Caristo Diagnostics’ CaRi-Heart AI. Operators then reviewed and confirmed an algorithm that indicated the presence of inflammation and plaque in the coronary arteries.

They found that increased inflammation also meant a higher chance of cardiovascular disease, or a fatal heart attack. The British Heart Foundation (BHF) says about 76,000 people in the UK suffer from heart disease, and the NHS spends £750 million a year treating it. About 350,000 people in the UK have heart CT scans each year.

The Orfan study published in The Lancet showed that 80% of 40,000 patients studied by the Oxford Centre for Risk Factors and Non-Invasive Imaging were referred back to primary care without any specific preventive management or course of treatment.

Studies have shown that patients with inflamed coronary arteries are 20 to 30 times more likely to die from heart disease within the next 10 years. With the help of this AI technology, 45% of these patients were either prescribed medication or advised to make changes to avoid future heart attacks.