Researchers from Asan Medical Center, South Korea's largest hospital, have developed an AI-based system for analysing coronary artery images prior to coronary angioplasty.

Coronary angioplasty is a standard treatment to widen narrowed or blocked coronary arteries indicated in coronary artery diseases such as angina and heart attack. 

FINDINGS

The Asan research team's AI system supports cardiovascular angiography, which is done before a coronary angioplasty by an experienced cardiologist to check on a narrowed artery and determine the appropriate stent. Once an angiogram is uploaded to the system during a procedure, the AI evaluates the internal diameter, lesion length, and degree of stenosis of the artery and then calculates the optimal size of the stent. 

In a study, which findings have been published in Cardiovascular Interventions under the Journal of the American College of Cardiology, the researchers compared the outcomes of nearly 400 patients in 13 medical institutions whose treatments were supported with AI and those who underwent 3D-based imaging-guided coronary angioplasty. They found that the AI system achieved comparable safety outcomes as the latter considering the minimum stent area – a key success indicator for the heart procedure. 

The research team also noted the lack of major clinical events, such as death, myocardial infarction, and thrombosis with AI-guided coronary angiography. 

WHY IT MATTERS

Currently, the best and most accurate method to identify coronary artery lesions and support coronary angioplasty is 3D-based high-resolution imaging called intravascular optical coherence tomography. Despite providing an accurate evaluation of artery lesions, the procedure is costly and takes a long time, making it less preferred to use, according to the Asan researchers.

The research team proposed an AI-guided system and proved its comparable safety in what could be the world's first study. 

"Using AI, coronary angiography images can be analysed quickly and objectively without additional time or labour, making coronary intervention possible without error," Jung-min Ahn, AMC professor and study lead, claimed in a statement.

He shared that they plan to expand the application of their AI system to cover complex coronary artery lesions, such as left main spleen lesions and graft vascular lesions. 

MARKET SNAPSHOT

While the AMC researchers' AI-based system might be the first to prove its safety, it was not the first to integrate AI in cardiac angiography. United States-based Cleerly and Caristo Diagnostics from the United Kingdom have developed AI to evaluate CT angiograms for heart disease diagnosis. 

In Australia, publicly-listed company Artrya offers Salix, which also uses AI to detect and assess atherosclerotic plaque in heart CT scans. The product has been approved for sale in Europe and the United Kingdom.