Cardiologists use video-based AI model to ID coronary artery disease
Researchers out of Cedars-Sinai in Los Angeles have developed a new video-based artificial intelligence (AI) model that can identify signs of coronary artery disease in a patient’s echocardiography results, sharing their findings in the Journal of the American Society of Echocardiography.[1]
The group trained its deep learning model using transthoracic echocardiograms (TTEs) from nearly 3,000 patients treated at a high-volume health system from 2015 to 2020. All patients underwent a coronary artery calcium (CAC) test and a TTE within one year of one another.
Overall, the video-based deep learning model was able to “read” echocardiograph results and label patients with zero CAC or high CAC with considerable accuracy. The area under the receiver operating curve (ROC) for identifying patients with zero CAC was 0.81. The area under the ROC curve for identifying patients with high CAC was 0.74. An external dataset of 92 TTEs from a separate high-volume health system was used to confirm the AI model’s effectiveness.
The authors noted that their research could potentially be used to determine when a patient faces an above-average risk of cardiovascular disease or mortality within the next 15 years. This could help clinicians determine when to consider certain medical therapies right away for a patient as opposed to waiting and tracking their health over time.
“We show that echocardiograms, when interpreted with our AI software, can predict coronary artery calcium and predict heart attack risk nearly as well as CT scans,” senior author David Ouyang, MD, a cardiologist in the department of cardiology in the Smidt Heart Institute at Cedars-Sinai, said in a prepared statement. “This proved true even in cases where the naked eye of an expert reader sees the ultrasound image of the heart as appearing fairly normal.”
As Ouyang et al. emphasized in their analysis, echocardiography provides clinicians with many benefits over other imaging modalities such as CT. It’s widely available, for instance, it does not require any radiation and it is associated with lower costs.
Read the full study here.
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.