Uncertainty estimation- and attention-based semi-supervised models for automatically delineate clinical target volume in CBCT images of breast cancer.

OBJECTIVES: Accurate segmentation of the clinical target volume (CTV) of CBCT images can observe the changes of CTV during patients' radiotherapy, and lay a foundation for the subsequent implementation of adaptive radiotherapy (ART). However, segmentation is challenging due to the poor quality of CBCT images and difficulty in obtaining target volumes. An uncertainty estimation- and attention-based semi-supervised model called residual convolutional block attention-uncertainty aware mean teacher (RCBA-UAMT) was proposed to delineate the CTV in cone-beam computed tomography (CBCT) images of breast cancer automatically. METHODS: A total of 60 patients who undergone radiotherapy after breast-conserving surgery were enrolled in this study, which involved 60 planning CTs and 380 CBCTs. RCBA-UAMT was proposed by integrating residual and attention modules in the backbone network 3D UNet. The attention module can adjust channel and spatial weights of the extracted image features. The proposed design can train the model and segment CBCT images with a small amount of labeled data (5%, 10%, and 20%) and a large amount of unlabeled data. Four types of evaluation metrics, namely, dice similarity coefficient (DSC), Jaccard, average surface distance (ASD), and 95% Hausdorff distance (95HD), are used to assess the model segmentation performance quantitatively. RESULTS: The proposed method achieved average DSC, Jaccard, 95HD, and ASD of 82%, 70%, 8.93, and 1.49 mm for CTV delineation on CBCT images of breast cancer, respectively. Compared with the three classical methods of mean teacher, uncertainty-aware mean-teacher and uncertainty rectified pyramid consistency, DSC and Jaccard increased by 7.89-9.33% and 14.75-16.67%, respectively, while 95HD and ASD decreased by 33.16-67.81% and 36.05-75.57%, respectively. The comparative experiment results of the labeled data with different proportions (5%, 10% and 20%) showed significant differences in the DSC, Jaccard, and 95HD evaluation indexes in the labeled data with 5% versus 10% and 5% versus 20%. Moreover, no significant differences were observed in the labeled data with 10% versus 20% among all evaluation indexes. Therefore, we can use only 10% labeled data to achieve the experimental objective. CONCLUSIONS: Using the proposed RCBA-UAMT, the CTV of breast cancer CBCT images can be delineated reliably with a small amount of labeled data. These delineated images can be used to observe the changes in CTV and lay the foundation for the follow-up implementation of ART.

Semi-automatic fine delineation scheme for pancreatic cancer.

BACKGROUND: Pancreatic cancer fine delineation in medical images by physicians is a major challenge due to the vast volume of medical images and the variability of patients. PURPOSE: A semi-automatic fine delineation scheme was designed to assist doctors in accurately and quickly delineating the cancer target region to improve the delineation accuracy of pancreatic cancer in computed tomography (CT) images and effectively reduce the workload of doctors. METHODS: A target delineation scheme in image blocks was also designed to provide more information for the deep learning delineation model. The start and end slices of the image block were manually delineated by physicians, and the cancer in the middle slices were accurately segmented using a three-dimensional Res U-Net model. Specifically, the input of the network is the CT image of the image block and the delineation of the cancer in the start and end slices, while the output of the network is the cancer area in the middle slices of the image block. Meanwhile, the model performance of pancreatic cancer delineation and the workload of doctors in different image block sizes were studied. RESULTS: We used 37 3D CT volumes for training, 11 volumes for validating and 11 volumes for testing. The influence of different image block sizes on doctors' workload was compared quantitatively. Experimental results showed that the physician's workload was minimal when the image block size was 5, and all cancer could be accurately delineated. The Dice similarity coefficient was 0.894 ± 0.029, the 95% Hausdorff distance was 3.465 ± 0.710 mm, the normalized surface Dice was 0.969 ± 0.019. By completing the accurate delineation of all the CT images, the speed of the new method is 2.16 times faster than that of manual sketching. CONCLUSION: Our proposed 3D semi-automatic delineative method based on the idea of block prediction could accurately delineate CT images of pancreatic cancer and effectively deal with the challenges of class imbalance, background distractions, and non-rigid geometrical features. This study had a significant advantage in reducing doctors' workload, and was expected to help doctors improve their work efficiency in clinical application.

uRP: An integrated research platform for one-stop analysis of medical images.

Introduction: Medical image analysis is of tremendous importance in serving clinical diagnosis, treatment planning, as well as prognosis assessment. However, the image analysis process usually involves multiple modality-specific software and relies on rigorous manual operations, which is time-consuming and potentially low reproducible. Methods: We present an integrated platform - uAI Research Portal (uRP), to achieve one-stop analyses of multimodal images such as CT, MRI, and PET for clinical research applications. The proposed uRP adopts a modularized architecture to be multifunctional, extensible, and customizable. Results and Discussion: The uRP shows 3 advantages, as it 1) spans a wealth of algorithms for image processing including semi-automatic delineation, automatic segmentation, registration, classification, quantitative analysis, and image visualization, to realize a one-stop analytic pipeline, 2) integrates a variety of functional modules, which can be directly applied, combined, or customized for specific application domains, such as brain, pneumonia, and knee joint analyses, 3) enables full-stack analysis of one disease, including diagnosis, treatment planning, and prognosis assessment, as well as full-spectrum coverage for multiple disease applications. With the continuous development and inclusion of advanced algorithms, we expect this platform to largely simplify the clinical scientific research process and promote more and better discoveries.

Validation of an automatic method for reconstruction, delineation, and measurement of levator hiatus in clinical practice.

OBJECTIVES: To evaluate the concordance between an automatic software program and manual evaluation in reconstructing, delineating, and measuring the levator hiatus (LH) on maximal Valsalva maneuver. METHODS: This was a retrospective study analyzing archived raw ultrasound imaging data of 100 patients underwent transperineal ultrasound (TPUS) examination. Each data were assessed by the automatic Smart Pelvic System software program and manual evaluation. The Dice similarity index (DSI), mean absolute distance (MAD), and Hausdorff distance (HDD) were calculated to quantify delineation accuracy of LH. Agreement between automatic and manual measurement of levator hiatus area was assessed by intraclass correlation coefficient (ICC) and Bland-Altman method. RESULTS: The satisfaction rate of automatic reconstruction was 94%. Six images were recognized as unsatisfactory reconstructed images for some gas in the rectum and anal canal. Compared with satisfactory reconstructed images, DSI of unsatisfactory reconstructed images was lower, MAD and HDD were larger (p = 0.001, p = 0.001, p = 0.006, respectively). The ICC was up to 0.987 in 94 satisfactory reconstructed images. CONCLUSIONS: The Smart Pelvic System software program had good performance in reconstruction, delineation, and measurement of LH on maximal Valsalva maneuver in clinical practice, despite misidentification of the border of posterior aspect of LH due to the influence of gas in the rectum.

The repeatability and consistency of different methods for measuring the volume parameters of the primary rectal cancer on diffusion weighted images.

Background: To determine the reproducibility of measuring the gross total volume (GTV) of primary rectal tumor with manual and semi-automatic delineation on the diffusion-weighted image (DWI), examine the consistency of using the same delineation method on DWI images with different high b-values, and find the optimal delineation method to measure the GTV of rectal cancer. Methods: 41 patients who completed rectal MR examinations in our hospital from January 2020 to June 2020 were prospectively enrolled in this study. The post-operative pathology confirmed the lesions were rectal adenocarcinoma. The patients included 28 males and 13 females, with an average age of (63.3 ± 10.6) years old. Two radiologists used LIFEx software to manually delineate the lesion layer by layer on the DWI images (b=1000 s/mm2 and 1500 s/mm2) and used 10% to 90% of the highest signal intensity as thresholds to semi-automatically delineate the lesion and measure the GTV. After one month, Radiologist 1 performed the same delineation work again to obtain the corresponding GTV. Results: The inter- and intra-observer interclass correlation coefficients (ICC) of measuring GTV using semi-automatic delineation with 30% to 90% as thresholds were all >0.900. There was a positive correlation between manual delineation and semi-automatic delineation with 10% to 50% thresholds (P < 0.05). However, the manual delineation was not correlated with the semi-automatic delineation with 60%, 70%, 80%, and 90% thresholds. On the DWI images with b=1000 s/mm2 and 1500 s/mm2, the 95% limit of agreement (LOA%) of measuring GTV using semi-automatic delineation with 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, and 90% thresholds were -41.2~67.4, -17.8~51.5, -16.1~49.3, -26.2~50.1, -42.3~57.6, -57.1~65.4, -67.3~66.5, -101.6~91.1, -129.4~136.0, and -15.3~33.0, respectively. The time required for GTV measurement by semi-automatic delineation was significantly shorter than that of manual delineation (12.9 ± 3.6s vs 40.2 ± 13.1s). Conclusions: The semi-automatic delineation of rectal cancer GTV with 30% threshold had high repeatability and consistency, and it was positively correlated with the GTV measured by manual delineation. Therefore, the semi-automatic delineation with 30% threshold could be a simple and feasible method for measuring rectal cancer GTV.

Deep learning-based accurate delineation of primary gross tumor volume of nasopharyngeal carcinoma on heterogeneous magnetic resonance imaging: A large-scale and multi-center study.

BACKGROUND AND PURPOSE: The problem of obtaining accurate primary gross tumor volume (GTVp) segmentation for nasopharyngeal carcinoma (NPC) on heterogeneous magnetic resonance imaging (MRI) images with deep learning remains unsolved. Herein, we reported a new deep-learning method than can accurately delineate GTVp for NPC on multi-center MRI scans. MATERIAL AND METHODS: We collected 1057 patients with MRI images from five hospitals and randomly selected 600 patients from three hospitals to constitute a mixed training cohort for model development. The resting patients were used as internal (n = 259) and external (n = 198) testing cohorts for model evaluation. An augmentation-invariant strategy was proposed to delineate GTVp from multi-center MRI images, which encouraged networks to produce similar predictions for inputs with different augmentations to learn invariant anatomical structure features. The Dice similarity coefficient (DSC), 95 % Hausdorff distance (HD95), average surface distance (ASD), and relative absolute volume difference (RAVD) were used to measure segmentation performance. RESULTS: The model-generated predictions had a high overlap ratio with the ground truth. For the internal testing cohorts, the average DSC, HD95, ASD, and RAVD were 0.88, 4.99 mm, 1.03 mm, and 0.13, respectively. For external testing cohorts, the average DSC, HD95, ASD, and RAVD were 0.88, 3.97 mm, 0.97 mm, and 0.10, respectively. No significant differences were found in DSC, HD95, and ASD for patients with different T categories, MRI thickness, or in-plane spacings. Moreover, the proposed augmentation-invariant strategy outperformed the widely-used nnUNet, which uses conventional data augmentation approaches. CONCLUSION: Our proposed method showed a highly accurate GTVp segmentation for NPC on multi-center MRI images, suggesting that it has the potential to act as a generalized delineation solution for heterogeneous MRI images.

Educative Impact of Automatic Delineation Applied to Head and Neck Cancer Patients on Radiation Oncology Residents.

To evaluate the educational impact on radiation oncology residents in training when introducing an automatic segmentation software in head and neck cancer patients regarding organs at risk (OARs) and prophylactic cervical lymph node level (LNL) volumes. Two cases treated by exclusive intensity-modulated radiotherapy were delineated by an expert radiation oncologist and were considered as reference. Then, these cases were delineated by residents divided into two groups: group 1 (control group), experienced residents delineating manually, group 2 (experimental group), young residents on their first rotation trained with automatic delineation, delineating manually first (M -) and then after using the automatic system (M +). The delineation accuracy was assessed using the Overlap Volume (OV). Regarding the OARs, mean OV was 0.62 (SD = 0.05) for group 1, 0.56 (SD = 0.04) for group 2 M - , and 0.61 (SD = 0.03) for group 2 M + . Mean OV was higher in group 1 compared to group 2 M - (p = 0.01). There was no OV difference between group 1 and group 2 M + (p = 0.67). Mean OV was higher in the group 2 M + compared to group 2 M - (p < 0.003). Regarding LNL, mean OV was 0.53 (SD = 0.06) in group 1, 0.54 (SD = 0.03) in group 2 M - , and 0.58 (SD = 0.04) in group 2 M + . Mean OV was higher in group 2 M + for 11 of the 12 analysed structures compared to group 2 M - (p = 0.016). Prior use of the automatic delineation software reduced the average contouring time per case by 34 to 40%. Prior use of atlas-based automatic segmentation reduces the delineation duration, and provides reliable OARs and LNL delineations.

Research on multi-Atlas segmentation based on feature clustering / 中华放射肿瘤学杂志

Objective:To study the improvement of normal tissue region of interest (ROI) segmentation based on clustering-based multi-Atlas segmentation method, thereby achieving better delineation of organs at risk.Methods:CT images of 100 patients with cervical cancer who had completed treatment in Zhejiang Cancer Hospital during 2019-2020 were selected as the Atlas database. According to the volume characteristic parameters of the organs at risk (bladder, rectum and outer contour), the Atlas database was divided into several subsets by k-means clustering algorithm. The image to be segmented was matched to the corresponding Atlas library for multi-Atlas segmentation. The dice similarity coefficient (DSC) was used to evaluate the segmentation results.Results:Using 30 patients as the test set, the sub-Atlas generated by different clustering methods were compared for the improvement of image segmentation results. Compared with general multi-Atlas segmentation methods, clustering-based multi-Atlas segmentation method significantly improve the segmentation accuracy for the bladder (DSC=0.83±0.09 vs. 0.69±0.15, P<0.001) and the rectum (0.7±0.07 vs. 0.56±0.16, P<0.001), but no statistical significance was observed for left and right femoral head (0.92±0.04, 0.91±0.02) and bone marrow (0.91±0.06). The average segmentation time of clustering-based multi-Atlas segmentation method was shorter than that of the general multi-Atlas segmentation method (2.7 min vs. 6.3 min). Conclusion:The clustering-based multi-Atlas segmentation method can not only reduce the number of Atlas images registered with the image to be segmented, but also can be expected to improve the segmentation effect and obtain higher accuracy.

Automatic-delineation model construction for prostate cancer target volume of postoperative radiotherapy based on artificial intelligence / 中华放射肿瘤学杂志

Objective:To explore the method of constructing automatic delineation model for clinical target volume (CTV) and partially organs at risk (OAR) of postoperative radiotherapy for prostate cancer based on convolutional neural network, aiming to improve the clinical work efficiency and the unity of target area delineation.Methods:Postoperative CT data of 117 prostate cancer patients manually delineated by one experienced clinician were retrospectively analyzed. A multi-class auto-delineation model was designed based on 3D UNet. Dice similarity coefficient (DSC), 95% Hausdorf distance (95%HD), and average surface distance (ASD) were used to evaluate the segmentation ability of the model. In addition, the segmentation results in the test set were evaluated by two senior physicians. And the CT data of 78 patients treated by other physicians were also collected for external validation of the model. The automatic segmentation of these 78 patients by CTV-UNet model was also evaluated by two physicians.Results:The mean DSC for tumor bed area (CTV1), pelvic lymph node drainage area (CTV2), bladder and rectum of CVT-UNet auto-segmentation model in the test set were 0.74, 0.82, 0.94 and 0.79, respectively. Both physicians' scoring results of the test set and the external validation showed more consensus on the delineation of CTV2 and OAR. However, the consensus of CTV1 delineation was less.Conclusions:The automatic delineation model based on convolutional neural network is feasible for CTV and related OAR of postoperative radiotherapy for prostate cancer. The automatic segmentation ability of tumor bed area still needs to be improved.

Automatic delineation of organ at risk in cervical cancer radiotherapy based on ensemble learning. / åŸºäºŽé›†æˆå­¦ä¹ çš„å®«é¢ˆç™Œæ”¾å°„æ²»ç–—å±åŠå™¨å®˜çš„è‡ªåŠ¨å‹¾ç”».

OBJECTIVES: The automatic delineation of organs at risk (OARs) can help doctors make radiotherapy plans efficiently and accurately, and effectively improve the accuracy of radiotherapy and the therapeutic effect. Therefore, this study aims to propose an automatic delineation method for OARs in cervical cancer scenarios of both after-loading and external irradiation. At the same time, the similarity of OARs structure between different scenes is used to improve the segmentation accuracy of OARs in difficult segmentations. METHODS: Our ensemble model adopted the strategy of ensemble learning. The model obtained from the pre-training based on the after-loading and external irradiation was introduced into the integrated model as a feature extraction module. The data in different scenes were trained alternately, and the personalized features of the OARs within the model and the common features of the OARs between scenes were introduced. Computer tomography (CT) images for 84 cases of after-loading and 46 cases of external irradiation were collected as the train data set. Five-fold cross-validation was adopted to split training sets and test sets. The five-fold average dice similarity coefficient (DSC) served as the figure-of-merit in evaluating the segmentation model. RESULTS: The DSCs of the OARs (the rectum and bladder in the after-loading images and the bladder in the external irradiation images) were higher than 0.7. Compared with using an independent residual U-net (convolutional networks for biomedical image segmentation) model [residual U-net (Res-Unet)] delineate OARs, the proposed model can effectively improve the segmentation performance of difficult OARs (the sigmoid in the after-loading CT images and the rectum in the external irradiation images), and the DSCs were increased by more than 3%. CONCLUSIONS: Comparing to the dedicated models, our ensemble model achieves the comparable result in segmentation of OARs for different treatment options in cervical cancer radiotherapy, which may be shorten time for doctors to sketch OARs and improve doctor's work efficiency.

Robust extraction of biological information from diffusion-weighted magnetic resonance imaging during radiotherapy using semi-automatic delineation.

Background and purpose: Diffusion-Weighted Magnetic Resonance imaging (DWI) quantifies water mobility through the Apparent Diffusion Coefficient (ADC), a promising radiotherapy response biomarker. ADC measurements depend on manual delineation of a region of interest, a time-consuming and observer-dependent process. Here, the aim was to introduce and test the performance of a new, semi-automatic delineation tool (SADT) for ADC calculation within the viable region of the tumour. Materials and methods: Thirty patients with rectal cancer were scanned with DWI before radiotherapy (RT) (baseline) and two weeks into RT (week 2). The SADT was based on intensities in b=1100 s mm-2 DWI and derived ADC maps. ADC values measured using the SADT and manual delineations were compared using Bland-Altman- and correlation analyses. Delineations were repeated to assess intra-observer variation, and repeatability was estimated using repeated DWI scans. Results: ADC measured using the SADT and manual delineation showed strong and moderate correlation at baseline and week 2, respectively, with the SADT measuring systematically smaller values. Intra-observer ADC variation was slightly smaller for the SADT compared to manual delineation both at baseline, [-0.00; 0.03] vs. [-0.02; 0.04] 10-3 mm2 s-1, and week 2, [-0.01; 0.00] vs. [-0.04; 0.07] 10-3 mm2 s-1 (68.3% limits of agreement). The ADC change between baseline and week 2 was larger than the ADC uncertainty ( ± 0.04 · 10-3 mm2 s-1) in all cases except one. Conclusion: The presented SADT showed performance comparable to manual expert delineation, and with sufficient consistency to allow extraction of potential biological information from the viable tumour.

Three-dimensional deep neural network for automatic delineation of cervical cancer in planning computed tomography images.

PURPOSE: Radiation therapy is an essential treatment modality for cervical cancer, while accurate and efficient segmentation methods are needed to improve the workflow. In this study, a three-dimensional V-net model is proposed to automatically segment clinical target volume (CTV) and organs at risk (OARs), and to provide prospective guidance for low lose area. MATERIAL AND METHODS: A total of 130 CT datasets were included. Ninety cases were randomly selected as the training data, with 10 cases used as the validation data, and the remaining 30 cases as testing data. The V-net model was implemented with Tensorflow package to segment the CTV and OARs, as well as regions of 5 Gy, 10 Gy, 15 Gy, and 20 Gy isodose lines covered. The auto-segmentation by V-net was compared to auto-segmentation by U-net. Four representative parameters were calculated to evaluate the accuracy of the delineation, including Dice similarity coefficients (DSC), Jaccard index (JI), average surface distance (ASD), and Hausdorff distance (HD). RESULTS: The V-net and U-net achieved the average DSC value for CTV of 0.85 and 0.83, average JI values of 0.77 and 0.75, average ASD values of 2.58 and 2.26, average HD of 11.2 and 10.08, respectively. As for the OARs, the performance of the V-net model in the colon was significantly better than the U-net model (p = 0.046), and the performance in the kidney, bladder, femoral head, and pelvic bones were comparable to the U-net model. For prediction of low-dose areas, the average DSC of the patients' 5 Gy dose area in the test set were 0.88 and 0.83, for V-net and U-net, respectively. CONCLUSIONS: It is feasible to use the V-Net model to automatically segment cervical cancer CTV and OARs to achieve a more efficient radiotherapy workflow. In the delineation of most target areas and OARs, the performance of V-net is better than U-net. It also offers advantages with its feature of predicting the low-dose area prospectively before radiation therapy (RT).

Deep learning-based automatic delineation of anal cancer gross tumour volume: a multimodality comparison of CT, PET and MRI.

BACKGROUND: Accurate target volume delineation is a prerequisite for high-precision radiotherapy. However, manual delineation is resource-demanding and prone to interobserver variation. An automatic delineation approach could potentially save time and increase delineation consistency. In this study, the applicability of deep learning for fully automatic delineation of the gross tumour volume (GTV) in patients with anal squamous cell carcinoma (ASCC) was evaluated for the first time. An extensive comparison of the effects single modality and multimodality combinations of computed tomography (CT), positron emission tomography (PET), and magnetic resonance imaging (MRI) have on automatic delineation quality was conducted. MATERIAL AND METHODS: 18F-fluorodeoxyglucose PET/CT and contrast-enhanced CT (ceCT) images were collected for 86 patients with ASCC. A subset of 36 patients also underwent a study-specific 3T MRI examination including T2- and diffusion-weighted imaging. The resulting two datasets were analysed separately. A two-dimensional U-Net convolutional neural network (CNN) was trained to delineate the GTV in axial image slices based on single or multimodality image input. Manual GTV delineations constituted the ground truth for CNN model training and evaluation. Models were evaluated using the Dice similarity coefficient (Dice) and surface distance metrics computed from five-fold cross-validation. RESULTS: CNN-generated automatic delineations demonstrated good agreement with the ground truth, resulting in mean Dice scores of 0.65-0.76 and 0.74-0.83 for the 86 and 36-patient datasets, respectively. For both datasets, the highest mean Dice scores were obtained using a multimodal combination of PET and ceCT (0.76-0.83). However, models based on single modality ceCT performed comparably well (0.74-0.81). T2W-only models performed acceptably but were somewhat inferior to the PET/ceCT and ceCT-based models. CONCLUSION: CNNs provided high-quality automatic GTV delineations for both single and multimodality image input, indicating that deep learning may prove a versatile tool for target volume delineation in future patients with ASCC.

Research on automatic delineation of nasopharyngeal carcinoma target area based on generative adversarial network / 中华放射肿瘤学杂志

Objective:To propose a deep learning network model 2D-PE-GAN to automatically delineate the target area of nasopharyngeal carcinoma and improve the efficiency of target area delineation.Methods:The model adopted the architecture of generative adversarial networks which used a UNet similar structure as the generator, and 2D-PE-block was added after each layer of convolution operation of the generator to improve the accuracy of delineation. The experimental data included CT images from 130 cases of nasopharyngeal carcinoma. The images were preprocessed before model training. In addition, three models of UNet, GAN, and GAN with an attention mechanism were compared, and Dice similarity coefficient, Hausdorff distance, accuracy, Matthews correlation coefficient, Jaccard distance were employed to evaluate network performance.Results:Compared with UNet, GAN and GAN with the attention mechanism, the average Dice similarity coefficient of 2D-PE-GAN network segmentation of CTV was increased by 26%, 4% and 2%. The average Dice similarity coefficient of GTV segmentation was increased by 21%, 4%, 2%, respectively. Compared with the GAN network with the attention mechanism, the parameters and time of 2D-PE-GAN were reduced by 0.16% and 18%, respectively.Conclusions:Compared with the above three networks, 2D-PE-GAN network can increase the segmentation accuracy of nasopharyngeal carcinoma target area delineation. At the same time, compared with the attention mechanism with similar reasons, 2D-PE-GAN network can reduce the occupation of computing resources when the segmentation accuracy is not much different.

Geometric and Dosimetric Evaluation of Deep Learning-Based Automatic Delineation on CBCT-Synthesized CT and Planning CT for Breast Cancer Adaptive Radiotherapy: A Multi-Institutional Study.

PURPOSE: We developed a deep learning model to achieve automatic multitarget delineation on planning CT (pCT) and synthetic CT (sCT) images generated from cone-beam CT (CBCT) images. The geometric and dosimetric impact of the model was evaluated for breast cancer adaptive radiation therapy. METHODS: We retrospectively analyzed 1,127 patients treated with radiotherapy after breast-conserving surgery from two medical institutions. The CBCT images for patient setup acquired utilizing breath-hold guided by optical surface monitoring system were used to generate sCT with a generative adversarial network. Organs at risk (OARs), clinical target volume (CTV), and tumor bed (TB) were delineated automatically with a 3D U-Net model on pCT and sCT images. The geometric accuracy of the model was evaluated with metrics, including Dice similarity coefficient (DSC) and 95% Hausdorff distance (HD95). Dosimetric evaluation was performed by quick dose recalculation on sCT images relying on gamma analysis and dose-volume histogram (DVH) parameters. The relationship between ΔD95, ΔV95 and DSC-CTV was assessed to quantify the clinical impact of the geometric changes of CTV. RESULTS: The ranges of DSC and HD95 were 0.73-0.97 and 2.22-9.36 mm for pCT, 0.63-0.95 and 2.30-19.57 mm for sCT from institution A, 0.70-0.97 and 2.10-11.43 mm for pCT from institution B, respectively. The quality of sCT was excellent with an average mean absolute error (MAE) of 71.58 ± 8.78 HU. The mean gamma pass rate (3%/3 mm criterion) was 91.46 ± 4.63%. DSC-CTV down to 0.65 accounted for a variation of more than 6% of V95 and 3 Gy of D95. DSC-CTV up to 0.80 accounted for a variation of less than 4% of V95 and 2 Gy of D95. The mean ΔD90/ΔD95 of CTV and TB were less than 2Gy/4Gy, 4Gy/5Gy for all the patients. The cardiac dose difference in left breast cancer cases was larger than that in right breast cancer cases. CONCLUSIONS: The accurate multitarget delineation is achievable on pCT and sCT via deep learning. The results show that dose distribution needs to be considered to evaluate the clinical impact of geometric variations during breast cancer radiotherapy.

Clinical implementation of deep-learning based auto-contouring tools-Experience of three French radiotherapy centers.

Deep-learning (DL)-based auto-contouring solutions have recently been proposed as a convincing alternative to decrease workload of target volumes and organs-at-risk (OAR) delineation in radiotherapy planning and improve inter-observer consistency. However, there is minimal literature of clinical implementations of such algorithms in a clinical routine. In this paper we first present an update of the state-of-the-art of DL-based solutions. We then summarize recent recommendations proposed by the European society for radiotherapy and oncology (ESTRO) to be followed before any clinical implementation of artificial intelligence-based solutions in clinic. The last section describes the methodology carried out by three French radiation oncology departments to deploy CE-marked commercial solutions. Based on the information collected, a majority of OAR are retained by the centers among those proposed by the manufacturers, validating the usefulness of DL-based models to decrease clinicians' workload. Target volumes, with the exception of lymph node areas in breast, head and neck and pelvic regions, whole breast, breast wall, prostate and seminal vesicles, are not available in the three commercial solutions at this time. No implemented workflows are currently available to continuously improve the models, but these can be adapted/retrained in some solutions during the commissioning phase to best fit local practices. In reported experiences, automatic workflows were implemented to limit human interactions and make the workflow more fluid. Recommendations published by the ESTRO group will be of importance for guiding physicists in the clinical implementation of patient specific and regular quality assurances.

CT based automatic clinical target volume delineation using a dense-fully connected convolution network for cervical Cancer radiation therapy.

BACKGROUND: It is very important to accurately delineate the CTV on the patient's three-dimensional CT image in the radiotherapy process. Limited to the scarcity of clinical samples and the difficulty of automatic delineation, the research of automatic delineation of cervical cancer CTV based on CT images for new patients is slow. This study aimed to assess the value of Dense-Fully Connected Convolution Network (Dense V-Net) in predicting Clinical Target Volume (CTV) pre-delineation in cervical cancer patients for radiotherapy. METHODS: In this study, we used Dense V-Net, a dense and fully connected convolutional network with suitable feature learning in small samples to automatically pre-delineate the CTV of cervical cancer patients based on computed tomography (CT) images and then we assessed the outcome. The CT data of 133 patients with stage IB and IIA postoperative cervical cancer with a comparable delineation scope was enrolled in this study. One hundred and thirteen patients were randomly designated as the training set to adjust the model parameters. Twenty cases were used as the test set to assess the network performance. The 8 most representative parameters were also used to assess the pre-sketching accuracy from 3 aspects: sketching similarity, sketching offset, and sketching volume difference. RESULTS: The results presented that the DSC, DC/mm, HD/cm, MAD/mm, ∆V, SI, IncI and JD of CTV were 0.82 ± 0.03, 4.28 ± 2.35, 1.86 ± 0.48, 2.52 ± 0.40, 0.09 ± 0.05, 0.84 ± 0.04, 0.80 ± 0.05, and 0.30 ± 0.04, respectively, and the results were greater than those with a single network. CONCLUSIONS: Dense V-Net can correctly predict CTV pre-delineation of cervical cancer patients and can be applied in clinical practice after completing simple modifications.

Deep learning-based auto-delineation of gross tumour volumes and involved nodes in PET/CT images of head and neck cancer patients.

PURPOSE: Identification and delineation of the gross tumour and malignant nodal volume (GTV) in medical images are vital in radiotherapy. We assessed the applicability of convolutional neural networks (CNNs) for fully automatic delineation of the GTV from FDG-PET/CT images of patients with head and neck cancer (HNC). CNN models were compared to manual GTV delineations made by experienced specialists. New structure-based performance metrics were introduced to enable in-depth assessment of auto-delineation of multiple malignant structures in individual patients. METHODS: U-Net CNN models were trained and evaluated on images and manual GTV delineations from 197 HNC patients. The dataset was split into training, validation and test cohorts (n= 142, n = 15 and n = 40, respectively). The Dice score, surface distance metrics and the new structure-based metrics were used for model evaluation. Additionally, auto-delineations were manually assessed by an oncologist for 15 randomly selected patients in the test cohort. RESULTS: The mean Dice scores of the auto-delineations were 55%, 69% and 71% for the CT-based, PET-based and PET/CT-based CNN models, respectively. The PET signal was essential for delineating all structures. Models based on PET/CT images identified 86% of the true GTV structures, whereas models built solely on CT images identified only 55% of the true structures. The oncologist reported very high-quality auto-delineations for 14 out of the 15 randomly selected patients. CONCLUSIONS: CNNs provided high-quality auto-delineations for HNC using multimodality PET/CT. The introduced structure-wise evaluation metrics provided valuable information on CNN model strengths and weaknesses for multi-structure auto-delineation.

RefineNet-based automatic delineation of the clinical target volume and organs at risk for three-dimensional brachytherapy for cervical cancer.

BACKGROUND: In recent years, high-precision image-guided intensity-modulated radiation therapy combined with three-dimensional (3D) high-dose-rate (HDR) brachytherapy (BT) has become a recommended technique for radical radiotherapy for cervical cancer. This study first employed contrast-limited adaptive histogram equalization (CLAHE) for preprocessing of input data to achieve image enhancement. In this way, rapid and accurate automatic delineation of the clinical target volume (CTV) and organs at risk (OARs) in 3D BT for cervical cancer was achieved. METHODS: Two hundred cervical cancer patients who underwent radical radiotherapy from January 2016 to December 2018 were selected. After collecting the computed tomography (CT) image data of a patient, we constructed the radiotherapy CTV and OAR image libraries. A RefineNet-based deep learning protocol was used to segment the CTV and OARs for 3D BT for cervical cancer. In this study, a total of 1,000 rounds of training were carried out, and the model with the best performance was selected for subsequent iterative tuning. Finally, the clinical test was carried out, in which the CT images of 10 cases were tested one by one. The manual delineation results and the model output results for the CTV and OARs were compared to measure the performance of the model. RESULTS: Compared with the manually delineated CTV, the RefineNet model-based segmented CTV had a higher Dice similarity coefficient (DSC), Hausdorff distance (HD), and overlap index (OI), which were 0.861, 6.005, and 0.839, respectively. For OARs, the RefineNet-based model obtained the best results for bladder segmentation (DSC: 85.96%), respectively. The mean duration of RefineNet-based automatic contour processing of the CTV was 70 s, and the mean durations of RefineNet-based automatic delineation of the bladder, rectum, sigmoid colon, and small intestine were 67, 67.4, 63.8, and 60.8 s, respectively. The total time saved by RefineNet was approximately 60%. CONCLUSIONS: The RefineNet-based automatic delineation model for 3D BT for cervical cancer is a stable and highly consistent automatic delineation algorithmic model that has the potential to improve the consistency of target region delineation, simplify the radiotherapy procedure, and achieve rapid and accurate automatic delineation of CTVs and OARs.

Comparison of the automatic delineation of two kinds of stomach by the AccuContour software for patients with thoracic and abdominal tumors / 中国辐射卫生

Objective To delineate the normal stomach and thoracic stomach structure of patients with thoracic and abdominal tumor automatically using the AccuContour software based on deep learning in order to evaluate and compare the results. Methods Thirty-six patients with choracic and abdominal tumors were chosen for this study, and were divided into two groups. Group A included 18 patients with normal stomach, and group B included the other 18 patients undergoing esophageal carcinoma operation with thoracic stomach. The stomach structures were automatically delineated by the AccuContour software in the simulation CT series. Statistical analysis was carried out to data of the differences in volume, position and shape between the automatic and manual delineations, and data of the two kinds of stomach were compared. Results For group A, the differences in volume (ΔV%) between the automatic and manual delineations was (−1.82 ± 9.65)%, the total position difference (ΔL) was (0.51 ± 0.37) cm, the values of dice similarity coefficient (DSC) was 0.89 ± 0.04. There were significant differences in values of ΔV%、ΔL and DSC (P < 0.05). Conclusion The used version of AccuContour software in this study had a satisfactory result of automatic delineation of the normal stomach structure larger than certain volume, but could not delineate the thoracic stomach structures effectively for patients undergoing esophageal carcinoma operation.