Agatston-2.0: A next-generation AI-based coronary calcium quantification approach to improve risk stratification among individuals with zero Agatston scores - Part I: An AI-CVD Study within the Multi-Ethnic Study of Atherosclerosis and Framingham Heart Study.
Authors
Affiliations (21)
Affiliations (21)
- HeartLung.AI, Houston, TX 77021, USA.
- Department of Electrical and Computer Engineering, Cornell University, Ithaca, NY 14853, USA.
- Tustin Teleradiology, Tustin, CA 92780, USA.
- Clinic of Radiology and Nuclear Medicine, University Hospital Basel, Basel, Switzerland.
- Department of Radiology, University Medical Center Groningen, Groningen, the Netherlands.
- Department of Diagnostic, Molecular, and Interventional Radiology, Icahn School of Medicine at Mount Sinai, New York, NY, USA.
- Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, USA.
- The Lundquist Institute, Torrance, CA 90502, USA.
- Division of Cardiology, Kaiser Permanente Oakland Medical Center, Oakland, CA, USA.
- BioMedical Engineering and Imaging Institute, Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA.
- The Agatston Center, Miami Beach, FL, USA.
- Departments of Medicine and Radiology, Stanford University, Stanford, CA, USA.
- University of Houston, Houston, TX 77030, USA.
- Cedars-Sinai Medical Center, Los Angeles, CA 90048, USA.
- Department of Cardiac Imaging, Princeton Longevity Center, Princeton, NJ, USA.
- University of Louisville, Louisville, KY 40202, USA.
- Division of Cardiology, Department of Medicine, University of Alberta, Edmonton, Alberta, Canada.
- Stanford Prevention Research Center, Stanford University School of Medicine, Stanford, CA 94305, USA.
- Huntington Medical Research Institutes, Pasadena, CA 91105, USA.
- Keck School of Medicine of the University of Southern California, Los Angeles, CA 90033, USA.
- Heart Disease Prevention Program, Mary and Steve Wen Cardiovascular Division, University of California Irvine, Irvine, CA 92697, USA.
Abstract
A coronary artery calcium (CAC) score of zero using the conventional Agatston scoring method (Agatston-1.0) is associated with very low cardiovascular risk (the "Power of Zero"); however, a small proportion of individuals with CAC=0 still develop coronary heart disease (CHD). Agatston-1.0 relies on thick slices (2.5-3 mm) and a fixed attenuation threshold (≥130 HU), which may miss early, small, low-density, or partially calcified coronary plaques. Agatston-2.0 is an artificial intelligence (AI) framework for automated coronary segmentation and continuous voxel-wise calcium quantification without fixed thresholds, applicable to CT scans with slice thickness ≥0.2 mm and generating an AI-derived CAC score (AI-CAC). To evaluate the prognostic value of Agatston-2.0 for risk stratification among individuals with a baseline CAC=0. We pooled 3,965 participants with CAC=0 from the Multi-Ethnic Study of Atherosclerosis (MESA, n=2,816) and the Framingham Heart Study (FHS, n=1,149). Associations with incident CHD were evaluated using Cox proportional hazards models with up to 20 years of follow-up. An AI-CAC score>0 was detected in 862 participants (21.7%) in population with CAC=0. Participants with AI-CAC>0 had higher 20-year CHD incidence than those with AI-CAC=0 (7.7% vs. 3.8%, p<0.0001). After adjustment for traditional risk factors, AI-CAC>0 remained independently associated with incident CHD (HR 1.71, 95% CI 1.18-2.47). AI-CAC also predicted progression to positive CAC score (adjusted HR 1.95, 95% CI 1.70-2.24). The Agatston-2.0 framework identifies clinically meaningful coronary calcification in individuals classified as CAC=0. If validated in additional cohorts, Agatston-2.0 could become the new standard for coronary calcium scoring.