Automatic segmentation of high‐risk clinical target volume for tandem‐and‐ovoids brachytherapy patients using an asymmetric dual‐path convolutional neural network

Abstract

Purposes Preimplant diagnostic magnetic resonance imaging is the gold standard for image‐guided tandem‐and‐ovoids (T&O) brachytherapy for cervical cancer. However, high dose rate brachytherapy planning is typically done on postimplant CT‐based high‐risk clinical target volume (HR‐CTVCT) because the transfer of preimplant Magnetic resonance (MR)‐based HR‐CTV (HR‐CTVMR) to the postimplant planning CT is difficult due to anatomical changes caused by applicator insertion, vaginal packing, and the filling status of the bladder and rectum. This study aims to train a dual‐path convolutional neural network (CNN) for automatic segmentation of HR‐CTVCT on postimplant planning CT with guidance from preimplant diagnostic MR. Methods Preimplant T2‐weighted MR and postimplant CT images for 65 (48 for training, eight for validation, and nine for testing) patients were retrospectively solicited from our institutional database. MR was aligned to the corresponding CT using rigid registration. HR‐CTVCT and HR‐CTVMR were manually contoured on CT and MR by an experienced radiation oncologist. All images were then resampled to a spatial resolution of 0.5 × 0.5 × 1.25 mm. A dual‐path 3D asymmetric CNN architecture with two encoding paths was built to extract CT and MR image features. The MR was masked by HR‐CTVMR contour while the entire CT volume was included. The network put an asymmetric weighting of 18:6 for CT: MR. Voxel‐based dice similarity coefficient (DSCV), sensitivity, precision, and 95% Hausdorff distance (95‐HD) were used to evaluate model performance. Cross‐validation was performed to assess model stability. The study cohort was divided into a small tumor group (<20 cc), medium tumor group (20–40 cc), and large tumor group (>40 cc) based on the HR‐CTVCT for model evaluation. Single‐path CNN models were trained with the same parameters as those in dual‐path models. Results For this patient cohort, the dual‐path CNN model improved each of our objective findings, including DSCV, sensitivity, and precision, with an average improvement of 8%, 7%, and 12%, respectively. The 95‐HD was improved by an average of 1.65 mm compared to the single‐path model with only CT images as input. In addition, the area under the curve for different networks was 0.86 (dual‐path with CT and MR) and 0.80 (single‐path with CT), respectively. The dual‐path CNN model with asymmetric weighting achieved the best performance with DSCV of 0.65 ± 0.03 (0.61–0.70), 0.79 ± 0.02 (0.74–0.85), and 0.75 ± 0.04 (0.68–0.79) for small, medium, and large group. 95‐HD were 7.34 (5.35–10.45) mm, 5.48 (3.21–8.43) mm, and 6.21 (5.34–9.32) mm for the three size groups, respectively. Conclusions An asymmetric CNN model with two encoding paths from preimplant MR (masked by HR‐CTVMR) and postimplant CT images was successfully developed for automatic segmentation of HR‐CTVCT for T&O brachytherapy patients.

Deep learning in MRI‐guided radiation therapy: A systematic review

“…Since the application of CNNs to MRI-based segmentation in 2017 22 , fully convolutional networks (FCNs) have outperformed competing atlas-based and hand-crafted auto-segmentation methods, often matching the intraobserver variability among physicians. 23 FCNs employ convolutional layers which are trained to detect patterns in either nearby voxels or feature maps output from previous convolutional layers. In contrast with traditional CNNs, FCNs forgo densely connected layers.…”

“…This technique is called Monte Carlo Dropout (MCDO). Cao et al 23 takes pre-implant MRI and post-implant CT as input channels to their network. After preforming intra-observer variability analysis, they achieve performance more similar to a specialist radiation oncologist for cervical tumors in brachytherapy than a non-specialist.…”

Self‐configuring nnU‐Net for automatic delineation of the organs at risk and target in high‐dose rate cervical brachytherapy, a low/middle‐income country's experience

“…38 In IG-HDR cervical brachytherapy, work has been done on automating the reconstruction of applicators using DL, 39,40 with a few studies looking at the automatic segmentation of the OARs and targets. [40][41][42][43][44] These studies however do not look at the clinical acceptability of the generated contours and involve complex deep learning models that require a high level of expertise to reproduce or apply in one's own clinical department. 45 The purpose of this study is to train and implement the self -configuring No New U-Net (nnU-Net), developed by Isensee et al, for the task of automatically delineating the OARs and HR-CTV in IG-HDR cervical brachytherapy.…”

“…DL has also been effectively applied in various areas of brachytherapy from applicator reconstruction, dose calculation, treatment planning as well as organ delineation 38 . In IG‐HDR cervical brachytherapy, work has been done on automating the reconstruction of applicators using DL, 39,40 with a few studies looking at the automatic segmentation of the OARs and targets 40–44 …”

Evaluation of auto-segmentation for brachytherapy of postoperative cervical cancer using deep learning-based workflow

“…Many studies have demonstrated the accuracy and availability about auto-segmentation of target and organs at risk (OARs) in external beam radiation therapy (EBRT) [7][8][9][10] . However, relevant research was rarely mentioned in BT [11][12] . The delineation of planning structures is an important task but also such a labor-intensive part in RT work ow, and always suffers from the inter-and intra-observer variability [13] .…”