Synthetic CT‐aided multiorgan segmentation for CBCT‐guided adaptive pancreatic radiotherapy

Abstract

The delineation of organs at risk (OARs) is fundamental to conebeam CT (CBCT)-based adaptive radiotherapy treatment planning, but is time consuming, labor intensive, and subject to interoperator variability. We investigated a deep learning-based rapid multiorgan delineation method for use in CBCT-guided adaptive pancreatic radiotherapy. Methods: To improve the accuracy of OAR delineation, two innovative solutions have been proposed in this study. First, instead of directly segmenting organs on CBCT images, a pretrained cycle-consistent generative adversarial network (cycleGAN) was applied to generating synthetic CT images given CBCT images. Second, an advanced deep learning model called mask-scoring regional convolutional neural network (MS R-CNN) was applied on those synthetic CT to detect the positions and shapes of multiple organs simultaneously for final segmentation. The OAR contours delineated by the proposed method were validated and compared with expert-drawn contours for geometric agreement using the Dice similarity coefficient (DSC), 95th percentile Hausdorff distance (HD95), mean surface distance (MSD), and residual mean square distance (RMS). Results: Across eight abdominal OARs including duodenum, large bowel, small bowel, left and right kidneys, liver, spinal cord, and stomach, the geometric comparisons between automated and expert contours are as follows: 0.92 (0.89-0.97) mean DSC, 2.90 mm (1.63-4.19 mm) mean HD95, 0.89 mm (0.61-1.36 mm) mean MSD, and 1.43 mm (0.90-2.10 mm) mean RMS. Compared to the competing methods, our proposed method had significant improvements (p < 0.05) in all the metrics for all the eight organs. Once the model was trained, the contours of eight OARs can be obtained on the order of seconds. Conclusions: We demonstrated the feasibility of a synthetic CT-aided deep learning framework for automated delineation of multiple OARs on CBCT. The proposed method could be implemented in the setting of pancreatic adaptive radiotherapy to rapidly contour OARs with high accuracy.

Deep learning methods for enhancing cone‐beam CT image quality toward adaptive radiation therapy: A systematic review

“…demonstrated Dice score agreement greater than 0.89 for the pelvic region 46,47 . Recently, segmentation‐specific studies have compared the performance of CBCT‐derived sCT images for abdominal segmentations, reporting Dice scores above 0.8 125,126 . The validation of sCT‐based auto‐segmentation for all anatomic regions is an important step toward online ART.…”

Dosimetric evaluation of magnetic resonance imaging-guided adaptive radiation therapy in pancreatic cancer by extent of re-contouring of organs-at-risk

“…Recently, auto-segmentation for OARs using deep learning for RT has gained attention in the field, and techniques are evolving [24][25][26]. Although challenging, many studies deal with auto-segmentation in intra-abdominal organs, and they report non-inferior results compared with expert-drawn contours [27,28]. These studies are based on cone-beam CT images, but studies from other organs such as the prostate employ MRI-based auto-segmentation [29].…”

Deep learning‐based segmentation in prostate radiation therapy using Monte Carlo simulated cone‐beam computed tomography

“…After reconstruction, voxels outside the field of view were set to 0. The reconstructed voxel values represent the photon attenuation coefficient 𝜇, which was converted to CBCT numbers with CBCT# = 𝜇 × 2 16 − 1024 to mimic the scanner processing indicated in previous works. 26,28…”

“…12 Another approach is to train a generative DL network to synthesize CT images. [13][14][15][16] The training database is then made of delineated planning CT images, which can be obtained from existing clinical databases. The CBCT image is first transformed into a pseudo-CT (pCT; e.g., with cycleGAN) before being processed by the CNN for segmentation.…”

“…One approach is to perform a scatter correction of the CBCT projections, for example, using scatter kernel algorithms, 10 a CT image of the same patient 11 or an anti‐scatter grid 12 . Another approach is to train a generative DL network to synthesize CT images 13–16 . The training database is then made of delineated planning CT images, which can be obtained from existing clinical databases.…”