AI model gets helpful heart assessments out of nongated CT scans
A new artificial intelligence (AI) algorithm is capable of reading noncontrast, nongated CT scans performed for noncardiac purposes and calculating a patient’s coronary artery calcium (CAC) score. The group behind the new AI model shared its efforts in NEJM AI, an AI-focused journal from The New England Journal of Medicine.[1]
“Millions of chest CT scans are taken each year, often in healthy people, for example to screen for lung cancer,” senior author Hugo Aerts, PhD, director of the AI in Medicine program at Mass General Brigham, and colleagues. “Our study shows that important information about cardiovascular risk is going unnoticed in these scans. Our study shows that AI has the potential to change how clinicians practice medicine and enable physicians to engage with patients earlier, before their heart disease advances to a cardiac event.”
Aerts et al. developed their deep learning algorithm, AI-CAC, using data from 98 medical centers within the Veterans Affairs healthcare system. They then compared AI-CAC to electrocardiogram-gated CAC scoring by testing the algorithm on more than 8,000 patients.
Overall, the group linked its algorithm to an accuracy of 89.4% when determining whether a patient’s imaging results contained signs of CAC. Taking that one step further, the algorithm’s accuracy was 87.3% when determining if patients with CAC would have a score higher or lower than 100.
In addition, cardiologists confirmed that 99.2% of patients identified by the algorithm as having a very high CAC score would, in fact, benefit from lipid-lowering therapy.
“At present, VA imaging systems contain millions of nongated chest CT scans that may have been taken for another purpose, around 50,000 gated studies. This presents an opportunity for AI-CAC to leverage routinely collected nongated scans for purposes of cardiovascular risk evaluation and to enhance care,” said first author Raffi Hagopian, MD, a cardiologist and researcher in the Applied Innovations and Medical Informatics group at the VA Long Beach Healthcare System. “Using AI for tasks like CAC detection can help shift medicine from a reactive approach to the proactive prevention of disease, reducing long-term morbidity, mortality and healthcare costs.”
Because AI-CAC was trained and tested on U.S. veterans, the group did emphasize that it may not be as effective with a more general patient population. It hopes to run similar studies on a general population in the future.
Click here to read the full study in NEJM AI.
Michael has more than 18 years of experience as a professional writer and editor. He has written at length about cardiology, radiology, artificial intelligence and other key healthcare topics.