Purpose: Computational phantoms have important applications in paediatric radiotherapy, in quality assurance of diagnostic/therapeutic protocols, and in reconstructing historical radiation doses. Detailed age-specific computational anatomical models are available, developed from average and/or healthy individuals, which may not be representative of cancer patients due to pathology and/or treatment effects. This study investigated the capability of existing phantoms in representing the paediatric radiotherapy population.
Methods: Computational models evaluated were the International Commission on Radiological Protection (ICRP) paediatric reference computational phantoms (n=8, median age 8y, range: 1–15y) and the default 4D extended cardiac torso (XCAT) (n=75, median age 9y, range: 1–18y). Five key organs (kidneys, lungs, spleen, liver and brain) were automatically segmented on the virtual phantoms similar to clinical organ at risk segmentation protocols. Anatomical similarity was assessed in terms of organ length and mass. These quantities were measured on the phantoms and on a clinical radiotherapy dataset, consisting of planning CT images/contours from craniospinal irradiation patients (n=68, median age 7, range: 2–16y). We also compared clinical measures with published literature on healthy children (9 publications, median age 8y, range 1–16y).
Results: For each dataset (phantom, published and clinical data) we performed a linear fit of the mass and length across the ages. Differences between clinical data and virtual phantoms/published data were calculated as the average relative difference between the linear fits for integer ages across the clinical data. For the phantoms, differences across all organs ranged from 1–22% for lengths and 4–35% for masses. For published data these were 5–23% and 7–39%, respectively. The smallest and largest differences were found for the liver and spleen, respectively.
Conclusion: Quantitative anatomical differences were described between phantoms, literature, and routine radiotherapy data. Our findings will help selecting and tailoring the phantoms most representative of this population.
Funding Support, Disclosures, and Conflict of Interest: RA and JC were supported by National Institute for Health Research University College London Hospitals Biomedical Research Centre and Wellcome Trust Institutional Strategic Support Fund (WISSF). CV is supported by the Royal Academy of Engineering under the Research Fellowship scheme (RF\201718\17140).
Phantoms, CT, Radiation Therapy