Room: AAPM ePoster Library
Purpose: in-house computational phantom uses age-scaling functions (ASF) that account for non-uniform growth by uniquely scaling each body region to any age (newborn-to-adult) in 3D. We aimed to determine if these ASFs can be used to scale human anatomy computed tomography (CT)-images from one age to another and evaluate if age-scaled anatomy is representative of growth between ages. The scaled phantom can be used for late effects studies in which dose reconstructions varies with patient’s age and anatomy.
Methods: applied our ASFs to 6 whole-body CT-based phantoms (ages: 0.1, 1, 5, 10, 15 and 30 years) from the National Cancer Institute (NCI) humanoid phantom series. Each phantom was separated into body regions and each region was scaled to each of the six ages in the NCI phantom series; creating a total of 36 age-scaled phantoms. To evaluate scaling ability, we calculated the Dice Similarity Coefficient (DSC) and Hausdorff distance (HD) between each age-scaled phantom and the original NCI phantom for the same age.
Results: average DSCs for the head, neck and trunk regions were 0.86 (range 0.63-0.97), 0.52 (range 0.18-0.86) and 0.89 (range 0.83-0.96), respectively. The average HD (cm) for the head, neck, and trunk regions were 2.22 (range 1.16-4.31), 2.70 (range 0.50-10.91) and 3.11 (range 0.74-7.04), respectively. Overall, all regions were reproduced with reasonable accuracy except for the neck for ages 0.1 and 1 years and for the head of the infant phantom. Head and neck discrepancies were attributed to inadequate separation/delineation of those regions from the adjacent regions.
Conclusion: have demonstrated that our ASFs can be used to scale human anatomy CT-images from one age to another and that age-scaled anatomy is representative of growth that occurs between ages. In the future, we will expand our analysis to include longitudinal whole-body CT scans of actual patients.