Artificial Intelligence-Guided Quantitative Coronary CT Assessment to Rule-In or Rule-Out Myocardial Ischemia.
Authors
Affiliations (13)
Affiliations (13)
- Department of Cardiology, Leiden University Medical Center, Leiden, The Netherlands.
- Faculty of Medicine, Universitas Brawijaya, Malang, Indonesia.
- Department of Cardiology, Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands.
- Department of Vascular Medicine, Amsterdam UMC, University of Amsterdam, Amsterdam, The Netherlands.
- Department of Cardiology, Radboud University Medical Center, Nijmegen, The Netherlands.
- Department of Radiology and Nuclear Medicine, Amsterdam UMC, Vrije Universiteit Amsterdam, the Netherlands.
- Department of Perioperative Cardiology and Cardiovascular Imaging, Centro Cardiologico Monzino IRCCS, Milan.
- Department of Biomedical, Surgical and Dental Sciences, University of Milan, Milan, Italy.
- Department of University Cardiology and Cardiac Imaging, IRCCS Ospedale Galeazzi Sant'Ambrogio, Milan, Italy.
- Division of Cardiology, Severance Cardiovascular Hospital and Severance Biomedical Science Institute, Yonsei University College of Medicine, Yonsei University Health System, Seoul, South Korea.
- Division of Cardiology, The George Washington University School of Medicine, Washington, DC, USA.
- Turku University Hospital and University of Turku, Turku, Finland.
- Department of Cardiology, Division of Heart and Lungs, Utrecht University, Utrecht University Medical Center, Utrecht, The Netherlands.
Abstract
To evaluate the ability of AI-based quantitative CT (AI-QCT) parameters, diameter stenosis, percent atheroma volume (PAV) and average lumen area (ALA) to rule-in or rule-out ischemia. This post-hoc, vessel-level analysis included patients with suspected CAD from the CREDENCE (612 patients; 1727 vessels) and PACIFIC-1 (208 patients; 612 vessels) studies who underwent CCTA and invasive FFR. In addition to diameter stenosis, PAV and ALA were evaluated as key predictors of ischemia. We report abnormal FFR prevalence based on these variables and define rule-out (<15% ischemia prevalence, defer further testing), rule-in (>75% prevalence, ischemia highly likely; further testing typically unnecessary) and intermediate risk (15-75%, consider additional functional assessment). PAV and ALA were dichotomized using median values derived from the CREDENCE cohort (14.7% and 3.9 mm2) and validated in PACIFIC-1. In CREDENCE, all vessels with 1-24% stenosis were ruled-out. Among vessels with 25-49% stenosis, 74% met rule-out criteria, while 26%, characterized by large PAV and small ALA, were intermediate risk. Within the proposed framework vessels with 50-69% stenosis were classified as intermediate risk. For 70-99% stenosis, 93% met rule-in criteria, except a small subset with small PAV and large ALA. In PACIFIC-1, 86% of vessels with <50% stenosis were ruled-out, and 61% of those with 50-99% stenosis were ruled-in. A simplified framework incorporating AI-QCT parameters including diameter stenosis, PAV (>14.7%), and ALA (<3.9 mm2), stratifies myocardial ischemia risk. Most non-obstructive lesions can be ruled-out, while most stenoses >70% are reliably ruled-in. This practical approach enhances the diagnostic utility of CCTA and streamline clinical decision-making.