Dr. Murat Demirci

Name of the study: Artificial-Intelligence Enabled Quantitative CT Assessment of Atherosclerosis and Major Adverse Events: The Multi-Center International CONFIRM2 Registry
Published in Congress:  TCT 2024, Presenter: Alexander van Rosendael

Link: https://www.tctmd.com/slide/artificial-intelligence-enabled-quantitative-ct-assessment-atherosclerosis-and-major-adverse

Introduction: Current chest pain assessment methods primarily rely on symptoms and ischemic status; however, numerous studies have shown that these approaches are not sufficiently effective in improving prognosis. Coronary computed tomography angiography (CTA) stands out as a powerful method for assessing atherosclerotic burden and composition. Artificial intelligence-supported quantitative CT (AI-QCT) offers potential as a reliable tool for prognosis prediction by accurately measuring the total heart plaque burden.

Objective: This study aims to identify the quantitative atherosclerotic CT features most strongly associated with major adverse cardiovascular events (MACE) and to compare these features with traditional clinical risk scores.

Methods: A total of 3,551 patients referred for coronary CTA due to suspected coronary artery disease were included in this study. Patients with known coronary artery disease and those with a life expectancy of less than two years were excluded. Conventional and artificial intelligence-based quantitative CT methods were employed in patient evaluation. Using conventional CT, patients were categorized based on stenosis severity as “none,” “non-obstructive,” and “obstructive.” In the AI-assisted CT analysis, total plaque volume, calcified plaque volume, non-calcified plaque volume, high-risk plaque features, and stenosis percentage were assessed. The primary endpoint of the study was defined as all-cause mortality, myocardial infarction, stroke, congestive heart failure, revascularization performed after 90 days, and hospitalization for unstable angina.

Results: During a mean follow-up period of 4.8 ± 2.2 years, 167 (5%) major events observed. The distribution of these events was as follows: death (n=34), myocardial infarction (n=24), cerebrovascular events (n=12), hospitalizations due to chronic heart failure (n=23), hospitalizations for unstable angina (n=17), and late revascularization (n=84). Based on conventional CT assessment, 86.8% of patients were reported as having “no stenosis” or “non-obstructive stenosis.” Among the 167 patients who experienced a major event, 56.3% were classified under the “no stenosis” or “non-obstructive stenosis” categories. In the AI-assisted CT (AI-QCT) analysis, high-risk plaque features were observed in 5.4% of patients without events and in 15.6% of those who experienced events (P<0.001). Non-calcified plaque volume and stenosis percentage, as measured by AI-QCT, were identified as the strongest predictors for the primary endpoint.

Conclusion: Measurement of lumen stenosis and non-calcified plaque volume via AI-QCT may serve as an effective tool for predicting MACE. The atherosclerotic profile assessed by AI-QCT could guide individualized adjustments, supporting the potential of anti-atherosclerotic therapies or coronary interventional procedures to reduce cardiovascular events.

Comment: This study demonstrates that artificial intelligence-supported quantitative CT (AI-QCT) technology is superior to traditional methods in assessing atherosclerotic plaque burden and composition in patients with coronary artery disease. The high accuracy in predicting cardiovascular events, particularly through parameters such as non-calcified plaque volume and stenosis percentage, suggests that AI-QCT could be a valuable tool in clinical decision-making processes.