Evaluation of generalization ability for deep learning‐based auto‐segmentation accuracy in limited field of view CBCT of male pelvic region

Abstract

The aim of this study was to evaluate generalization ability of segmentation accuracy for limited FOV CBCT in the male pelvic region using a full-image CNN. Auto-segmentation accuracy was evaluated using various datasets with different intensity distributions and FOV sizes. Methods: A total of 171 CBCT datasets from patients with prostate cancer were enrolled. There were 151, 10, and 10 CBCT datasets acquired from Vero4DRT, TrueBeam STx, and Clinac-iX, respectively. The FOV for Vero4DRT, TrueBeam STx, and Clinac-iX was 20, 26, and 25 cm, respectively. The ROIs, including the bladder, prostate, rectum, and seminal vesicles, were manually delineated. The U 2 -Net CNN network architecture was used to train the segmentation model. A total of 131 limited FOV CBCT datasets from Vero4DRT were used for training (104 datasets) and validation (27 datasets); thereafter the rest were for testing. The training routine was set to save the best weight values when the DSC in the validation set was maximized. Segmentation accuracy was qualitatively and quantitatively evaluated between the ground truth and predicted ROIs in the different testing datasets. Results:The mean scores ± standard deviation of visual evaluation for bladder, prostate, rectum, and seminal vesicle in all treatment machines were 1.0 ± 0.7, 1.5 ± 0.6, 1.4 ± 0.6, and 2.1 ± 0.8 points, respectively. The median DSC values for all imaging devices were ≥0.94 for the bladder, 0.84-0.87 for the prostate and rectum, and 0.48-0.69 for the seminal vesicles. Although the DSC values for the bladder and seminal vesicles were significantly different among the three imaging devices, the DSC value of the bladder changed by less than 1% point. The median MSD values for all imaging devices were ≤1.2 mm for the bladder and 1.4-2.2 mm for the prostate, rectum, and seminal vesicles. The MSD values for the seminal vesicles were significantly different between the three imaging devices. Conclusion:The proposed method is effective for testing datasets with different intensity distributions and FOV from training datasets.

Organ‐contour‐driven auto‐matching algorithm in image‐guided radiotherapy

“…10,11 Furthermore, deep learningbased auto-segmentation has demonstrated its utility in planning image and daily image segmentation. 12 While the current segmentation accuracy is imperfect, ongoing technological advancements are expected to result in improved accuracy and the eventual achievement of highly precise auto-segmentation capabilities for daily images. 11,13 The primary objective of this study is to demonstrate the potential clinical applicability of an organ-contourdriven auto-matching algorithm in IGRT, laying the groundwork for a future in which daily contouring will become a practical reality.…”

“…Several auto‐segmentation techniques have been introduced in the field of radiotherapy to assist in the delineation of targets and OARs 10,11 . Furthermore, deep learning‐based auto‐segmentation has demonstrated its utility in planning image and daily image segmentation 12 . While the current segmentation accuracy is imperfect, ongoing technological advancements are expected to result in improved accuracy and the eventual achievement of highly precise auto‐segmentation capabilities for daily images 11,13 …”