Opportunistic Detection of Coronary Artery Calcium on Noncardiac Chest Computed Tomography: An Emerging Tool for Cardiovascular Disease Prevention: A Scientific Statement From the American Heart Association

Comprehensive Summary

This review article discusses the current research on opportunistically identifying coronary artery calcium (CAC) during noncardiac chest CT scans. Although CAC is strongly associated with atherosclerotic cardiovascular disease and related negative cardiac outcomes, only about 1 million CAC scans are performed in the U.S. annually. The lack of CAC scan utilization is likely due to a combination of clinician unawareness, out-of-pocket cost, and lack of proper equipment. Conversely, noncardiac CT scans are performed almost twenty times more frequently than dedicated CAC scans. Manual CAC scoring techniques for noncardiac CT scans exist, but vary in objectivity and standardization, and are subject to reader fatigue and pose a time burden. Deep learning models have been developed with high accuracy to Agaston scores, the gold standard of CAC reporting. A 2016 study developed convolutional neural network models that examined low-dose chest CT scans and identified 97% of calcifications with 84% accuracy compared with ECG-gated CAC (Lessmann, et al). A separate study performed in 2021 developed convolutional neural network models with a sensitivity of 91% and specificity of 92% (Xu, et al). These results indicate a high potential for future AI-algorithms to assist in CAC detection without the need for a dedicated CAC scan.

Outcomes and Implications

Although CAC presence and severity has been associated with adverse cardiac events, multiple barriers currently stand in the way of implementing routine CAC assessment, including the time and financial burden of performing an ECG-gated CAC scan, and even lack of equipment to do so. These barriers are especially prevalent in areas with greater financial burden and fewer medical resources. The combination of mobile scanning units with automated CAC detection with AI may eventually allow clinicians of underrepresented populations to detect risk of atherosclerotic cardiovascular disease earlier and discuss preventive strategies with their patients, such as weight management, LDL goals, and antithrombotic therapy.