Researchers have developed an advanced AI model that can predict the presence of cardiovascular disease (CVD) deaths and coronary heart disease (CHD) deaths in asymptomatic patients better than other state-of-the-art risk prediction methods, sharing their findings in JACC: Cardiovascular Imaging.
The team’s machine learning model was designed to use all available information, including the patient’s coronary artery calcium (CAC) score and non-contrast CT findings, to reach its conclusions. The study’s authors used data from more than 66,000 patients with a mean age of 54 years old. The cohort was 67% male.
Overall, the team found that its AI model achieved an area under the curve for predicting CVD deaths (0.845) higher than two other traditional risk scores, the ASCVD risk calculator (0.821) and the CAC score alone (0.781). The AI model’s area under the curve for predicting CHD deaths (0.860) was also superior to the ASCVD risk calculator (0.835) and CAC score alone (0.816).
“Our findings suggest that future risk prediction models based on all available information can achieve a more accurate and precise model to identify risk that could be implemented clinically to improve the clinical use of CAC scanning in risk assessment and guiding management decisions,” wrote lead author Rine Nakanishi, Toho University Graduate School of Medicine in Tokyo, and colleagues. “The machine learning approach is likely to become a routine tool for risk assessment using CAC scanning with the evolution of the electronic medical record and its integration with imaging data in future clinical practice.”
The full analysis from Nakanishi et al. can be read here.