(1) Purpose: Challenges remain in dose accumulation for cervical cancer radiotherapy combined with external beam radiotherapy (EBRT) and brachytherapy (BT) as there are many large and complex organ deformations between different treatments. This study aims to improve deformable image registration (DIR) accuracy with the introduction of multi-metric objectives for dose accumulation of EBRT and BT. (2) Materials and methods: Twenty cervical cancer patients treated with EBRT (45–50 Gy/25 fractions) and high-dose-rate BT (≥20 Gy in 4 fractions) were included for DIR. The multi-metric DIR algorithm included an intensity-based metric, three contour-based metrics, and a penalty term. Nonrigid B-spine transformation was used to transform the planning CT images from EBRT to the first BT, with a six-level resolution registration strategy. To evaluate its performance, the multi-metric DIR was compared with a hybrid DIR provided by commercial software. The DIR accuracy was measured by the Dice similarity coefficient (DSC) and Hausdorff distance (HD) between deformed and reference organ contours. The accumulated maximum dose of 2 cc (D_2cc) of the bladder and rectum was calculated and compared to simply addition of D_2cc from EBRT and BT (ΔD_2cc). (3) Results: The mean DSC of all organ contours for the multi-metric DIR were significantly higher than those for the hybrid DIR (p ≤ 0.011). In total, 70% of patients had DSC > 0.8 using the multi-metric DIR, while 15% of patients had DSC > 0.8 using the commercial hybrid DIR. The mean ΔD_2cc of the bladder and rectum for the multi-metric DIR were 3.25 ± 2.29 and 3.54 ± 2.02 Gy_EQD2, respectively, whereas those for the hybrid DIR were 2.68 ± 2.56 and 2.32 ± 3.25 Gy_EQD2, respectively. The multi-metric DIR resulted in a much lower proportion of unrealistic D_2cc than the hybrid DIR (2.5% vs. 17.5%). (4) Conclusions: Compared with the commercial hybrid DIR, the introduced multi-metric DIR significantly improved the registration accuracy and resulted in a more reasonable accumulated dose distribution.

Tumor bed (TB) is the residual cavity of resected tumor after surgery. Delineating TB from CT is crucial in generating clinical target volume for radiotherapy. Due to multiple surgical effects and low image contrast, segmenting TB from soft tissue is challenging. In clinical practice, titanium clips were used as marks to guide the searching of TB. However, this information is limited and may cause large error. To provide more prior location information, the tumor regions on both pre-operative and post-operative CTs are both used by the deep learning model in segmenting TB from surrounding tissues. For breast cancer patient after surgery and going to be treated by radiotherapy, it is important to delineate the target volume for treatment planning. In clinical practice, the target volume is usually generated from TB by adding a certain margin. Therefore, it is crucial to identify TB from soft tissue. To facilitate this process, a deep learning model is developed to segment TB from CT with the guidance of prior tumor location. Initially, the tumor contour on the pre-operative CT is delineated by physician for surgical planning purpose. Then this contour is transformed to the post-operative CT via the deformable image registration between paired pre-operative and post-operative CTs. The original and transformed tumor regions are both used as inputs for predicting the possible region of TB by the deep-learning model. Compared to the one without prior tumor contour information, the dice similarity coefficient of the deep-learning model with the prior tumor contour information is improved significantly (0.812 vs. 0.520, P = 0.001). Compared to the traditional gray-level thresholding method, the dice similarity coefficient of the deep-learning model with the prior tumor contour information is improved significantly (0.812 vs.0.633, P = 0.0005). The prior tumor contours on both pre-operative and post-operative CTs provide valuable information in searching for the precise location of TB on post-operative CT. The proposed method provided a feasible way to assist auto-segmentation of TB in treatment planning of radiotherapy after breast-conserving surgery.

Background: Accurate delineation of clinical target volume of tumor bed (CTV-TB) is important but it is also challenging due to surgical effects and soft tissue contrast. Recently a few auto-segmentation methods were developed to improve the process. However, those methods had comparatively low segmentation accuracy. In this study the prior information was introduced to aid auto-segmentation of CTV-TB based on a deep-learning model. Methods: To aid the delineation of CTV-TB, the tumor contour on preoperative CT was transformed onto postoperative CT via deformable image registration. Both original and transformed tumor contours were used for prior information in training an auto-segmentation model. Then, the CTV-TB contour on postoperative CT was predicted by the model. 110 pairs of preoperative and postoperative CT images were used with a 5-fold cross-validation strategy. The predicted contour was compared with the clinically approved contour for accuracy evaluation using dice similarity coefficient (DSC) and Hausdorff distance. Results: The average DSC of the deep-learning model with prior information was improved than the one without prior information (0.808 vs. 0.734, P < 0.05). The average DSC of the deep-learning model with prior information was higher than that of the traditional method (0.808 vs. 0.622, P < 0.05). Conclusions: The introduction of prior information in deep-learning model can improve segmentation accuracy of CTV-TB. The proposed method provided an effective way to automatically delineate CTV-TB in postoperative breast cancer radiotherapy.

Background: Deep learning (DL) models have been used for predicting gamma passing rates (GPRs) of patient specific quality assurance (PSQA) in volumetric modulated arc therapy (VMAT). As the model input mostly from planning parameters and dose distribution, the effect of motion precision of the treatment machine has frequently been ignored. In this study,the most recent motion precision of multi-leaf collimator (MLC) leaves was estimated by their records in machine log file and used jointly with the beam intensity map to predict GPRs of VMAT PSQA. Methods: The input of the DL model consists of two parts: the MLC positional error map which was estimated from machine log file and the beam intensity map which was obtained from the treatment plan. A UNet++ model was trained to predict the GPR of VMAT PSQA. The effect of feature maps and time intervals in estimating MLC leaf positional error maps on the prediction accuracy was evaluated. In addition, the proposed model was compared with two competing DL models (ResNet and SE-Net) and a classic regression model. The prediction error was quantified by mean absolute errors (MAEs) under two clinical accepted criteria (3%-3 mm and 3%-2 mm). Pearson’s correlation coefficient (r) between measured and predicted GPR was also assessed. Results: The model with the inputs of beam intensity and MLC leaf positional error maps achieved the best performance among all combinations with various input feature maps. The MAEs predicted by the proposed model with MLC leaf positional error map obtained from 5 days’ records were 0.90 for 3%-3 mm criterion and 1.74 for 3%-2 mm criterion, whereas those predicted by the proposed model with MLC leaf positional error map obtained from 1 day’s record were 1.01 for 3%-3 mm criterion and 1.85 for 3%-2 mm criterion. The proposed model achieved the smaller MAEs (1.01 for 3%-3 mm criterion and 1.85 for 3%-2 mm criterion) than the other two competing models: ResNet (2.85 for 3%-3 mm criterion and 2.83 for 3%-2 mm criterion) and SE-Net (3.53 for 3%-3 mm criterion and 4.55 for 3%-2 mm criterion). Conclusions: The proposed model that combines beam intensity map and MLC leaf positional error map exhibited a better prediction performance for GPRs of VMAT PSQA. It would be advantageous to include more motion precision information of treatment machine into DL models for high-performance prediction of VMAT PSQA.

Background: Initially, cone-beam (CB) projection was developed to serve as a two-dimensional (2D) radiograph for fast target localization in radiotherapy. Later, it was gradually replaced by three-dimensional (3D) cone-beam computed tomography (CBCT), which is reconstructed from multiple CB projections. Since the size of CB projections is relatively large, they are usually discarded to conserve clinical storage. To effectively regenerate these projections, a deep learning (DL) method was developed to synthesize CB projections from CBCT. Methods: CB projection images from 50 patients under image-guided radiotherapy for head and neck cancer was collected. First, the digitally reconstructed radiograph (DRR) was created via a ray-tracing based algorithm. Next, a DL network was built to learn the pixel-to-pixel correspondence between the DRRs and CB projections in the training set. Finally, the CB projections were synthesized from DRRs in the testing set by a DL model. Three DL networks, Attention-UNet, Residual AutoEncoder, and Pix2Pix, were examined. Ablation studies on the effects of scatter correction, CBCT resolution, and training sample size were conducted. Non-DL and DL methods were compared, and a clinically relevant downstream task was conducted. Three similarity metrics, peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and video quality metric (VQM), were used to evaluate model performance. Results: Among three DL models, Attention-UNet achieved the best performance in terms of PSNR (29.71), SSIM (0.97), and VQM (0.20). The other two models showed comparable performance with PSNR, SSIM, and VQM values of 26.55, 0.96, and 0.27, respectively, for Residual AutoEncoder, and of 27.22, 0.94, and 0.25, respectively, for Pix2Pix. With scatter correction, the PSNR, SSIM, and VQM for Attention-UNet improved by 21.3%, 2.1%, and 39.3%, respectively. Under the finest voxel size (0.5×0.5×2.0 mm³), Attention-UNet achieved the highest PSNR (29.71±3.62), SSIM (0.97±0.008), and lowest VQM (0.20±0.091). The scatter correction and higher CBCT resolution could effectively improve the prediction accuracy of these DL models. Conclusions: The DL-based method can effectively synthesize CB projections from CBCT for patients with head and neck cancer. It may thus serve as a means to saving storage space for clinical CB projections.