Key points are not available for this paper at this time.
Abstract Background Artificial intelligence (AI)-based quantitative computed tomography (AI-QCT) is a novel tool for automated plaque characterisation and quantitation from coronary computed tomography angiography (CCTA). The prognostic value of various AI-QCT plaque types on top of the presence of ischemic coronary artery disease (CAD) is unknown to date. Purpose We aimed to investigate the added prognostic value of the AI-QCT plaque characteristics percent atheroma volume (PAV) %, non-calcified plaque burden (NCPB) %, and calcified plaque burden (CPB) % on top of: 1) A novel AI-based CCTA-derived ischemia algorithm (AI-QCTischemia) which predicts the presence of invasive FFR ≤0.80 using 38 morphological features from CCTA; and 2) ischemic CAD based on hybrid CCTA/15O-H2O positron emission tomography (PET) for the composite of all-cause death, myocardial infarction (MI), or unstable angina pectoris (uAP). Methods This was a cohort of 1772 symptomatic patients having undergone hybrid CCTA/15O-H2O PET perfusion imaging to assess the presence of ischemic CAD. Multivariable Cox regressions adjusted for clinical confounders (age, sex, hypertension, diabetes, smoking, typical angina) were performed and compared with the C-index. AI-QCTischemia models were: 1) AI-QCTischemia, 2) AI-QCTischemia+PAV, 3) AI-QCTischemia+NCPB, 4) AI-QCTischemia+CPB. Hybrid CCTA/PET models were: 5) hybrid CCTA/PET, 6) hybrid CCTA/PET+PAV, 7) hybrid CCTA/PET+NCPB, and 8) hybrid CCTA/PET+CPB. Results Throughout a median follow-up of 7 years, 177/1772 (10%) patients experienced the composite endpoint. The C-indexes (p-value vs. model 1) of the AI-QCTischemia models were 1) 0.729, 2) 0.748 (p=0.008), 3) 0.743 (p=0.012), and 4) 0.739 (p=0.067). The C-indexes (p-value vs. model 5) of the hybrid CCTA/PET models were: 5) 0.733, 6) 0.751 (p=0.014), 7) 0.746 (p=0.018), 8) 0.745 (p=0.058). AI-QCTischemia and hybrid CCTA/PET imaging performed similarly by means of C-statistics. Conclusions PAV and NCPB, but not CPB by AI-QCT plaque analysis improved the prediction of 7-year death, MI, or uAP when added on top of the presence of ischemic CAD, determined either by AI-QCTischemia or hybrid CCTA/PET imaging. Overall, the risk stratification based on AI-QCTischemia performed similarly as hybrid CCTA/PET imaging.C-indexes for death, MI, or uAP
Bär et al. (Thu,) studied this question.