Pediatric Cardiac Substructure Dose Coefficients for Chest and Abdomen–Pelvis CT Using a Size-Dependent Whole-Body Phantom Library
Abstract
Purpose
Pediatric dose from CT remains a concern, and recent evidence from radiotherapy has increased interest in dose to specific cardiac substructures (e.g., the left anterior descending coronary artery). Pediatric whole-body phantoms that contain detailed cardiac substructures modeling are not available, limiting CT-based substructure dosimetry. We developed detailed pediatric cardiac models and estimated substructure dose coefficients for common CT examinations.
Methods
Cardiologists segmented high-resolution cardiac MR images to delineate detailed cardiac anatomy including heart wall, atria and ventricles, ventricular myocardium, major coronary arteries, and cardiac valves. Segmentations were converted to 3D models and incorporated into the NCI body size-dependent whole-body phantom library (n = 169 with various heights and weights). Chest CT and abdomen–pelvis CT scans were simulated using previously published CT x-ray beam models. Absorbed doses were normalized to scanner output and reported as dose per CTDIvol (Gy/Gy). For reporting, we selected 5-, 10-, and 15-year phantoms whose heights and weights were closest to ICRP reference values from the 169-phantom library.
Results
For chest CT, cardiac substructure dose coefficients were consistently high and clustered across substructures, approximately 1.38 to 1.88 Gy/Gy. For abdomen and pelvis CT, dose coefficients were markedly smaller and showed a broader spread across substructures, approximately 0.07 to 0.43 Gy/Gy, reflecting sensitivity to scan extent when the heart is near the superior edge of the scan range.
Conclusion
We developed the first pediatric detailed cardiac substructure models within whole body phantoms and generated cardiac substructure dose coefficients for major CT examinations. These resources enable anatomically specific pediatric CT dosimetry for epidemiology, risk evaluation, and protocol assessment.