AI-defined cardiac anatomy Improves Risk Stratification of Hybrid Perfusion Imaging

作者：Robert J. H. Miller, Aakash D. Shanbhag, A. Killekar, M. Lemley, B. Bednarski, P. Kavanagh, A. Feher, E. J. Miller, T. Bateman, V. Builoff, Joanna X. Liang, D. Newby, D. Dey, Daniel S. Berman, Piotr J. Slomka · 发表于：JACC Cardiovascular Imaging · 年份：2024 · DOI：10.1016/j.jcmg.2024.01.006 · 研究领域：Medicine

Background: CTAC improves perfusion quantification of hybrid myocardial perfusion imaging (MPI) by correcting for attenuation artifacts. AI can automatically measure coronary artery calcium (CAC) from CTAC to improve risk prediction but could potentially derive additional anatomic features. Objectives: We evaluated artificial intelligence (AI)-based derivation of cardiac anatomy from CT attenuation correction (CTAC) and assessed its added prognostic utility. Methods: We considered consecutive patients without known coronary artery disease who underwent SPECT/CT MPI at 3 separate centers. Previously validated AI models were used to segment CAC and cardiac structures (left atrium[LA], left ventricle[LV], right atrium[RA], and right ventricle[RV] volume and LV mass) from CTAC. We evaluated associations with major adverse cardiovascular events (MACE), which included death, myocardial infarction, unstable angina, or revascularization. Results: In total, 7,613 patients were included with median age 64. During median follow-up of 2.4 (interquartile range 1.3 – 3.4) years, MACE occurred in 1,045 (13.7%) patients. Fully automated AI processing took an average of 6.2 ± 0.2 seconds for CAC and 15.8 ± 3.2 seconds for cardiac volumes and LV mass. Patients in the highest quartile of LV mass and LA, LV, RA, and RV volume were at significantly increased risk of MACE compared to patients in the lowest quartile, with a hazard ratio ranging from 1.46 to 3.31. The addition of all CT-based volume...