Abstract Body (Do not enter title and authors here): Background: Coronary artery disease is the leading cause of mortality in the United States, affecting approximately 4.9% of adults. Coronary artery calcium scoring (CACS) is used to detect disease and guide preventive therapy, such as statin initiation. Traditional CACS via ECG-gated computed tomography (CT) is accurate but costly, time-consuming, and requiring interpretation by a trained rater. In contrast, non-gated chest CT is widely available and performed for many non-cardiac indications such as infection or malignancy. Recent advances in artificial intelligence (AI) have enabled automated CACS from non-gated CTs. Although promising, the clinical reliability and impact of such models remain insufficiently characterized and warrant further study.
Research Question: Effectiveness of AI-driven CACS in detection of cardiovascular disease
Methods: We retrospectively reviewed patients who had undergone non-contrasted non-gated chest CT imaging at our institution between January 1 and June 30, 2024 for non-cardiac indications. An FDA-cleared AI tool, developed by Bunkerhill Health, performed automated CACS and risk stratification. We only included patient CAC score >400, indicative of severe coronary artery atherosclerosis. Patients with prior coronary artery bypass grafting (CABG), valve replacement, or coronary stenting were excluded. Electronic medical records were reviewed to assess use of statins and antiplatelet agents, cardiology follow-up after imaging, and the occurrence of major adverse cardiovascular events (MACE) and all-cause mortality.
Results: Our analysis included 493 patients aged 37 to 80 years old, with a mean age of 70 ± 7.17. Of those, 53% were on currently antiplatelet agents, and 75% were on a statin medication. As aggregate data, 24% of patients experienced heart failure, 12% myocardial infarction, and 13% angina. Within a 6-month time interval, only 48% of patients were seen by cardiologists. Throughout this period, all-cause mortality was 6% in the study population.
Conclusion: Our study shows the high rate of MACE in patients undergoing chest CT for non-cardiac indications and CAC Score > 400. AI-driven CAC scoring is an effective and accessible tool that can be used for cardiovascular disease detection and prevention. This gives the cardiologist a unique opportunity for truly transformative cardiac care.