Are age and gender suitable matching criteria in organ dose reconstruction using surrogate childhood cancer patients’ CT scans?

摘要: PURPOSE The purpose of this work was to assess the feasibility using surrogate CT scans matched patients for organ dose reconstructions childhood cancer (CC) survivors, treated in past with only 2D imaging data available instead 3D data, and particular current literature standard matching based on similarity age gender. METHODS Thirty-one recently CC abdominal were divided into six age- gender-matched groups. From each group, two radiotherapy plans Wilms' tumor selected as reference applied patients' CTs respective group. Two reconstruction strategies investigated: S1) without field adjustments; S2) manual adjustments according anatomical information, a visual check digitally reconstructed radiographs. To level agreement between distributions, we computed (using collapsed cone algorithm) compared absolute deviation mean maximum normalized by prescribed (i.e., errors |NEmean | |NE2cc |) eight organs at risk (OARs): heart, lungs, liver, spleen, kidneys, spinal cord. Furthermore, assessed quality case varying acceptance thresholds |. A accepted considered pass) if all OARs are smaller than threshold. pass fraction given threshold then defined percentage cases that classified pass. consider impact allowing use different scan OAR. RESULTS Slightly achieved S2 multiple S1 (P < 0.05). Among OARs, best found cord (average ≤ 4%). largest average spleen (18%). left lung (26%). Less 30% fraction) meet criteria 20% when gender single do reconstructions. Allowing other matchings combining from patients, increases substantially more 60%. CONCLUSIONS conclude, small deviations can be obtained scans, making general approach promising. However, select is not sufficient guarantee sufficiently low errors. It therefore suggested include features (e.g., height, extracted radiographs) survivors pre-3D planning era ways combine focused OARs.