Deepak Dev Vivekanandan, Nikita Singh, Marshall Robaczewski, Abigayle Wyer, Lucas N. Canaan, Daniel Whitson, Nathaniel Grabill, Mena Louis
Coronary artery disease (CAD) remains a leading global cause of morbidity and mortality, underscoring the need for effective cardiovascular risk stratification and preventive strategies. Coronary artery calcium (CAC) scoring, traditionally performed using electrocardiogram (ECG)-gated cardiac computed tomography (CT) scans, has been widely validated as a robust tool for assessing cardiovascular risk. However, its application has been largely limited to high-risk populations due to the costs, technical requirements, and limited accessibility of cardiac CT scans. Recent advancements in artificial intelligence (AI) have introduced transformative opportunities to extend CAC detection to noncardiac CT scans, such as those performed for lung cancer screening, enabling broader and more accessible cardiovascular screening.
This review provides a comprehensive analysis of AI-driven CAC detection, examining various types of AI models for CAC detection, like convolutional neural networks (CNNs) and U-Net architectures, and exploring the clinical, operational, and ethical implications of incorporating these technologies into routine practice. Technical challenges, including imaging variability, data privacy, and model bias, are discussed alongside essential areas for further research, such as standardization and validation across diverse populations. By leveraging widely available imaging data, AI-enabled CAC detection has the potential to advance preventive cardiology, supporting earlier risk identification, optimizing healthcare resources, and improving patient outcomes.