Background Routinely delineating of important skeletal growth centers is imperative to mitigate radiation‐induced growth abnormalities for pediatric cancer patients treated with radiotherapy. However, it is hindered by several practical problems, including difficult identification, time consumption, and inter‐practitioner variability. Purpose The goal of this study was to construct and evaluate a novel Triplet‐Attention U‐Net (TAU‐Net)‐based auto‐segmentation model for important skeletal growth centers in childhood cancer radiotherapy, concentrating on the accuracy and time efficiency. Methods A total of 107 childhood cancer patients fulfilled the eligibility criteria were enrolled in the training cohort (N = 80) and test cohort (N = 27). The craniofacial growth plates, shoulder growth centers, and pelvic ossification centers, with a total of 19 structures in the three groups, were manually delineated by two experienced radiation oncologists on axial, coronal, and sagittal computed tomography images. Modified from U‐Net, the proposed TAU‐Net has one main branch and two bypass branches, receiving semantic information of three adjacent slices to predict the target structure. With supervised deep learning, the skeletal growth centers contouring of each group was generated by three different auto‐segmentation models: U‐Net, V‐Net, and the proposed TAU‐Net. Dice similarity coefficient (DSC) and Hausdorff distance 95% (HD95) were used to evaluate the accuracy of three auto‐segmentation models. The time spent on performing manual tasks and manually correcting auto‐contouring generated by TAU‐Net was recorded. The paired t‐test was used to compare the statistical differences in delineation quality and time efficiency. Results Among the three groups, including craniofacial growth plates, shoulder growth centers, and pelvic ossification centers groups, TAU‐Net had demonstrated highly acceptable performance (the average DSC = 0.77, 0.87, and 0.83 for each group; the average HD95 = 2.28, 2.07, and 2.86 mm for each group). In the overall evaluation of 19 regions of interest (ROIs) in the test cohort, TAU‐Net had an overwhelming advantage over U‐Net (63.2% ROIs in DSC and 31.6% ROIs in HD95, p = 0.001–0.042) and V‐Net (94.7% ROIs in DSC and 36.8% ROIs in HD95, p = 0.001–0.040). With an average time of 52.2 min for manual delineation, the average time saved to adjust TAU‐Net‐generated contours was 37.6 min (p < 0.001), a 72% reduction. Conclusions Deep learning–based models have presented enormous potential for the auto‐segmentation of important growth centers in pediatric skeleton, where the proposed TAU‐Net outperformed the U‐Net and V‐Net in geometrical precision for the majority status.

Objective: To compare the 6-dimensional errors of different immobilization devices and body regions based on 3-dimensional cone beam computed tomography for image-guided radiotherapy and to further quantitatively evaluate the impact of rotational corrections on translational shifts and dose distribution based on anthropomorphic phantoms. Materials and Methods: Two hundred ninety patients with cone beam computed tomographies from 3835 fractions were retrospectively analyzed for brain, head & neck, chest, abdomen, pelvis, and breast cases. A phantom experiment was conducted to investigate the impact of rotational errors on translational shifts using cone beam computed tomography and the registration system. For the dosimetry study, pitch rotations were simulated by adjusting the breast bracket by ±2.5°. Roll and yaw rotations were simulated by rotating the gantry and couch in the planning system by ±3.0°, respectively. The original plan for the breast region was designed in the computed tomography image space without rotation. With the same planning parameters, the original plan was transplanted into the image space with different rotations for dose recalculation. The effect of these errors on the breast target and organs at risk was assessed by dose-volume histograms. Results: Most of the mean rotational errors in the breast region were >1°. A single uncorrected yaw of 3° caused a change of 2.9 mm in longitudinal translation. A phantom study for the breast region demonstrated that when the pitch rotations were −2.5° and 2.5° and roll and yaw were both 3°, the reductions in the planning target volumes-V50 Gy were 20.07% and 29.58% of the original values, respectively. When the pitch rotation was +2.5°, the left lung V5 Gy and heart Dmean were 7.49% and 165.76 Gy larger, respectively, than the original values. Conclusions: Uncorrected rotations may cause changes in the values and directions of translational shifts. Rotational corrections may improve the patient setup and dose distribution accuracy.