Heart failure and atrial fibrillation prediction from non-gated chest CT scans using AI-based cardiac chamber volumetry: an AI-CVD study within the multi-ethnic study of atherosclerosis (MESA).
Authors
Affiliations (19)
Affiliations (19)
- HeartLung.AI, Houston, TX, 77021, USA.
- Department of Electrical and Computer Engineering, Cornell University, Ithaca, NY. 14853, USA.
- Clinic of Radiology and Nuclear Medicine, University Hospital Basel, Petersgraben 4, CH-4031 Basel, Switzerland.
- Tustin Teleradiology, Tustin, CA 92780, USA.
- Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA.
- Department of Epidemiology and Biostatistics, Wen School of Public Health, University of California, Irvine, USA.
- Kravis Center for Clinical Cardiovascular Health, Mount Sinai Fuster Heart Hospital, New York, NY 10029 USA.
- Houston Methodist Hospital, Houston, TX 77030, USA.
- BioMedical Engineering and Imaging Institute. Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA.
- Stanford Prevention Research Center, Stanford University School of Medicine, Stanford, CA, 94305, USA.
- Division of Cardiology, Kaiser Permanente Oakland Medical Center, California, USA.
- Department of Radiology, University Medical Center Groningen, GZ 9713, Netherlands.
- University of Houston, Houston, TX, 77030, USA.
- The Lundquist Institute, CA 90502, USA.
- University of Louisville, Louisville, KY, 40202, USA.
- Cedars-Sinai Medical Center, Los Angeles, CA 90048, USA.
- The Agatston Center, Miami Beach, Florida, USA.
- Huntington Medical Research Institutes, Pasadena, CA, 91105, USA and Keck School of Medicine of University of Southern California, Los Angeles, CA 90033, USA.
- Heart Disease Prevention Program, Mary and Steve Wen Cardiovascular Division, University of California Irvine School of Medicine, CA 92697, USA.
Abstract
We previously demonstrated that AI-derived cardiac chamber volumetry in ECG-gated CT scans predicts atrial fibrillation (AF) and heart failure (HF). Since most chest CTs are non-gated, and the American Heart Association recently highlighted their value for opportunistic cardiovascular risk assessment using AI-based coronary artery calcium scoring, we extended this concept to evaluate whether AI-derived chamber volumetry from non-gated CTs can reliably predict HF and AF.s. We compared AI-derived chamber volumetry from non-gated and ECG-gated chest CTs in 2053 Multi-Ethnic Study of Atherosclerosis (MESA) Exam 5 participants using mean and standard-deviation of absolute differences, intraclass correlation coefficients (ICC), and Cohen's κ. Variability between paired ECG-gated scans was determined by analyzing scans acquired 2 min apart in 5749 MESA Exam 1 participants. Predictive performance for incident HF and AF was compared using cumulative incidence curves, age- and sex-adjusted C-indices, and non-inferiority testing (δ=0.05). Non-gated scans demonstrated excellent overall agreement with ECG-gated scans, with comparable performance for detecting cardiac chamber enlargement (98.8-99.2%, κ=0.76-0.84), relative to agreement observed between paired ECG-gated scans (98.6-98.9%, κ=0.72-0.77). Cumulative incidence curves for HF and AF stratified by quartiles showed substantial overlap. Age- and sex-adjusted C-indices were comparable for left atrial volume (0.75 vs 0.74, <i>p</i> = 0.001, non-inferiority <i>p</i> < 0.001), ventricular volume (both 0.77, <i>p</i> = 0.97, non-inferiority <i>p</i> < 0.001), and ventricular ratio (both 0.74, <i>p</i> = 0.57, non-inferiority <i>p</i> < 0.001) in predicting AF and HF. AI-based evaluation of non-gated chest CT may provide reliable cardiac chamber volumetry and cardiovascular outcome prediction, enabling opportunistic screening for HF and AF risk.