New Fully Automated AI Algorithm More Effective at Predicting Heart Attack Risk

Heart disease remains the leading cause of death in developed countries, with accurate risk prediction being a key challenge in preventing heart attacks. Coronary artery calcium (CAC) scoring has become a popular method for identifying patients at risk of heart disease, but it has its limitations. Current methods, such as the Agatston score, fail to account for the location of calcified plaques along the coronary arteries, which is critical for understanding risk, as plaques near the artery origin carry higher risks. Additionally, highly calcified plaques are paradoxically assigned higher risk despite being stable and associated with a lower risk of cardiac events. Now, a new, fully automated artificial intelligence (AI) algorithm has proven to be more effective than current methods at predicting the risk of a heart attack.

The automated AI algorithm known as the CAC-DAD score has been created by researchers from The University of Western Australia (Perth, Australia) in collaboration with medtech industry partners Artrya Ltd. (Perth, Australia). This algorithm is capable of measuring coronary calcification and the distance of each plaque from the coronary artery origin, providing a more precise risk assessment. By analyzing each individual plaque, the algorithm overcomes the limitations of traditional methods, including the misclassification of stable, highly calcified plaques. The CAC-DAD score can also reclassify plaques as low risk, which was previously not possible with conventional methods. This tool, developed from years of research and technological advancements, is designed to improve the prediction of heart attack risks, particularly in vulnerable populations.

The CAC-DAD score was tested and validated in a study involving a large group of patients at UWA's Medical School and collaborating institutions. The findings, published in The American Journal of Human Genetics, show that the CAC-DAD score is more effective and precise than the Agatston score at predicting heart attack risk, particularly around the time of surgery. Combining the CAC-DAD score with the Agatston score further improved the accuracy of risk predictions, suggesting that the tool has significant potential for clinical use. The researchers plan to validate the predictive power of the CAC-DAD score in larger international cohorts and explore its role in personalized heart disease prevention.

“Your calcium score is the single, greatest predictor for your risk of having a first heart attack and optimizing its accuracy will have significant benefits for the management of your risk,” said Professor Girish Dwivedi, Senior Author, UWA Medical School.

Related Links:
University of Western Australia
Artrya Ltd.