Online adaptive radiotherapy for bladder cancer using a simultaneous integrated boost and fiducial markers

PURPOSE

The aim was to assess the feasibility of online adaptive radiotherapy (oART) for bladder cancer using a focal boost by focusing on the quality of the online treatment plan and automatic target delineation, duration of the workflow and performance in the presence of fiducial markers for tumor bed localization.

METHODS

Fifteen patients with muscle invasive bladder cancer received daily oART with Cone Beam CT (CBCT), artificial intelligence (AI)-assisted automatic delineation of the daily anatomy and online plan reoptimization. The bladder and pelvic lymph nodes received a total dose of 40 Gy in 20 fractions, the tumor received an additional simultaneously integrated boost (SIB) of 15 Gy. The dose distribution of the reference plan was calculated for the daily anatomy, i.e. the scheduled plan. Simultaneously, a reoptimization of the plan was performed i.e. the adaptive plan. The target coverage and V95% outside the target were evaluated for both plans. The need for manual adjustments of the GTV delineation, the duration of the workflow and the influence of fiducial markers were assessed.

RESULTS

All 300 adaptive plans met the requirement of the CTV-coverage V95%≥98% for both the boost (55 Gy) and elective volume (40 Gy). For the scheduled plans the CTV-coverage was 53.5% and 98.5%, respectively. Significantly less tissue outside the targets received 55 Gy in case of the adaptive plans as compared to the scheduled plans. Manual corrections of the GTV were performed in 67% of the sessions. In 96% of these corrections the GTV was enlarged and resulted in a median improvement of 1% for the target coverage. The median on-couch time was 22 min. A third of the session time consisted of reoptimization of the treatment plan. Fiducial markers were visible on the CBCTs and aided the tumor localization.

CONCLUSIONS

AI-driven CBCT-guided oART aided by fiducial markers is feasible for bladder cancer radiotherapy treatment including a SIB. The quality of the adaptive plans met the clinical requirements and fiducial markers were visible enabling consistent daily tumor localization. Improved automatic delineation to lower the need for manual corrections and faster reoptimization would result in shorter session time.

Practical and technical key challenges in head and neck adaptive radiotherapy: The GORTEC point of view

Anatomical variations occur during head and neck (H&N) radiotherapy (RT) treatment. These variations may result in underdosage to the target volume or overdosage to the organ at risk. Replanning during the treatment course can be triggered to overcome this issue. Due to technological, methodological and clinical evolutions, tools for adaptive RT (ART) are becoming increasingly sophisticated. The aim of this paper is to give an overview of the key steps of an H&N ART workflow and tools from the point of view of a group of French-speaking medical physicists and physicians (from GORTEC). Focuses are made on image registration, segmentation, estimation of the delivered dose of the day, workflow and quality assurance for an implementation of H&N offline and online ART. Practical recommendations are given to assist physicians and medical physicists in a clinical workflow.

Image-guided Adaptive Radiotherapy for Bladder Cancer

Technological advancement has facilitated patient-specific radiotherapy in bladder cancer. This has been made possible by developments in image-guided radiotherapy (IGRT). Particularly transformative has been the integration of volumetric imaging into the workflow. The ability to visualise the bladder target using cone beam computed tomography and magnetic resonance imaging initially assisted with determining the magnitude of inter- and intra-fraction target change. It has led to greater confidence in ascertaining true anatomy at each fraction. The increased certainty of dose delivered to the bladder has permitted the safe reduction of planning target volume margins. IGRT has therefore improved target coverage with a reduction in integral dose to the surrounding tissue. Use of IGRT to feed back into plan and dose delivery optimisation according to the anatomy of the day has enabled adaptive radiotherapy bladder solutions. Here we undertake a review of the stepwise developments underpinning IGRT and adaptive radiotherapy strategies for external beam bladder cancer radiotherapy. We present the evidence in accordance with the framework for systematic clinical evaluation of technical innovations in radiation oncology (R-IDEAL).

Domain adversarial networks and intensity-based data augmentation for male pelvic organ segmentation in cone beam CT

In radiation therapy, a CT image is used to manually delineate the organs and plan the treatment. During the treatment, a cone beam CT (CBCT) is often acquired to monitor the anatomical modifications. For this purpose, automatic organ segmentation on CBCT is a crucial step. However, manual segmentations on CBCT are scarce, and models trained with CT data do not generalize well to CBCT images. We investigate adversarial networks and intensity-based data augmentation, two strategies leveraging large databases of annotated CTs to train neural networks for segmentation on CBCT. Adversarial networks consist of a 3D U-Net segmenter and a domain classifier. The proposed framework is aimed at encouraging the learning of filters producing more accurate segmentations on CBCT. Intensity-based data augmentation consists in modifying the training CT images to reduce the gap between CT and CBCT distributions. The proposed adversarial networks reach DSCs of 0.787, 0.447, and 0.660 for the bladder, rectum, and prostate respectively, which is an improvement over the DSCs of 0.749, 0.179, and 0.629 for "source only" training. Our brightness-based data augmentation reaches DSCs of 0.837, 0.701, and 0.734, which outperforms the morphons registration algorithms for the bladder (0.813) and rectum (0.653), while performing similarly on the prostate (0.731). The proposed adversarial training framework can be used for any segmentation application where training and test distributions differ. Our intensity-based data augmentation can be used for CBCT segmentation to help achieve the prescribed dose on target and lower the dose delivered to healthy organs.

Head-and-neck organs-at-risk auto-delineation using dual pyramid networks for CBCT-guided adaptive radiotherapy

Organ-at-risk (OAR) delineation is a key step for cone-beam CT (CBCT) based adaptive radiotherapy planning that can be a time-consuming, labor-intensive, and subject-to-variability process. We aim to develop a fully automated approach aided by synthetic MRI for rapid and accurate CBCT multi-organ contouring in head-and-neck (HN) cancer patients. MRI has superb soft-tissue contrasts, while CBCT offers bony-structure contrasts. Using the complementary information provided by MRI and CBCT is expected to enable accurate multi-organ segmentation in HN cancer patients. In our proposed method, MR images are firstly synthesized using a pre-trained cycle-consistent generative adversarial network given CBCT. The features of CBCT and synthetic MRI (sMRI) are then extracted using dual pyramid networks for final delineation of organs. CBCT images and their corresponding manual contours were used as pairs to train and test the proposed model. Quantitative metrics including Dice similarity coefficient (DSC), Hausdorff distance 95% (HD95), mean surface distance, and residual mean square distance (RMS) were used to evaluate the proposed method. The proposed method was evaluated on a cohort of 65 HN cancer patients. CBCT images were collected from those patients who received proton therapy. Overall, DSC values of 0.87 ± 0.03, 0.79 ± 0.10/0.79 ± 0.11, 0.89 ± 0.08/0.89 ± 0.07, 0.90 ± 0.08, 0.75 ± 0.06/0.77 ± 0.06, 0.86 ± 0.13, 0.66 ± 0.14, 0.78 ± 0.05/0.77 ± 0.04, 0.96 ± 0.04, 0.89 ± 0.04/0.89 ± 0.04, 0.83 ± 0.02, and 0.84 ± 0.07 for commonly used OARs for treatment planning including brain stem, left/right cochlea, left/right eye, larynx, left/right lens, mandible, optic chiasm, left/right optic nerve, oral cavity, left/right parotid, pharynx, and spinal cord, respectively, were achieved. This study provides a rapid and accurate OAR auto-delineation approach, which can be used for adaptive radiation therapy.

Auto-segmentations by convolutional neural network in cervical and anorectal cancer with clinical structure sets as the ground truth

•

The network provided auto-segmentations with high overlap with ground truth volumes.

•

Evaluation of femoral heads/bladder auto-segmentations showed highest overlap.

•

Annotated structure sets from daily clinical practice is feasible as ground truth.

Background

It is time-consuming for oncologists to delineate volumes for radiotherapy treatment in computer tomography (CT) images. Automatic delineation based on image processing exists, but with varied accuracy and moderate time savings. Using convolutional neural network (CNN), delineations of volumes are faster and more accurate. We have used CTs with the annotated structure sets to train and evaluate a CNN.

Material and methods

The CNN is a standard segmentation network modified to minimize memory usage. We used CTs and structure sets from 75 cervical cancers and 191 anorectal cancers receiving radiation therapy at Skåne University Hospital 2014-2018. Five structures were investigated: left/right femoral heads, bladder, bowel bag, and clinical target volume of lymph nodes (CTVNs). Dice score and mean surface distance (MSD) (mm) evaluated accuracy, and one oncologist qualitatively evaluated auto-segmentations.

Results

Median Dice/MSD scores for anorectal cancer: 0.91–0.92/1.93–1.86 femoral heads, 0.94/2.07 bladder, and 0.83/6.80 bowel bag. Median Dice scores for cervical cancer were 0.93–0.94/1.42–1.49 femoral heads, 0.84/3.51 bladder, 0.88/5.80 bowel bag, and 0.82/3.89 CTVNs. With qualitative evaluation, performance on femoral heads and bladder auto-segmentations was mostly excellent, but CTVN auto-segmentations were not acceptable to a larger extent.

Discussion

It is possible to train a CNN with high overlap using structure sets as ground truth. Manually delineated pelvic volumes from structure sets do not always strictly follow volume boundaries and are sometimes inaccurately defined, which leads to similar inaccuracies in the CNN output. More data that is consistently annotated is needed to achieve higher CNN accuracy and to enable future clinical implementation.

Pelvic multi-organ segmentation on cone-beam CT for prostate adaptive radiotherapy

BACKGROUND AND PURPOSE

The purpose of this study is to develop a deep learning-based approach to simultaneously segment five pelvic organs including prostate, bladder, rectum, left and right femoral heads on cone-beam CT (CBCT), as required elements for prostate adaptive radiotherapy planning.

MATERIALS AND METHODS

We propose to utilize both CBCT and CBCT-based synthetic MRI (sMRI) for the segmentation of soft tissue and bony structures, as they provide complementary information for pelvic organ segmentation. CBCT images have superior bony structure contrast and sMRIs have superior soft tissue contrast. Prior to segmentation, sMRI was generated using a cycle-consistent adversarial networks (CycleGAN), which was trained using paired CBCT-MR images. To combine the advantages of both CBCT and sMRI, we developed a cross-modality attention pyramid network with late feature fusion. Our method processes CBCT and sMRI inputs separately to extract CBCT-specific and sMRI-specific features prior to combining them in a late-fusion network for final segmentation. The network was trained and tested using 100 patients' datasets, with each dataset including the CBCT and manual physician contours. For comparison, we trained another two networks with different network inputs and architectures. The segmentation results were compared to manual contours for evaluations.

RESULTS

For the proposed method, dice similarity coefficients and mean surface distances between the segmentation results and the ground truth were 0.96 ± 0.03, 0.65 ± 0.67 mm; 0.91 ± 0.08, 0.93 ± 0.96 mm; 0.93 ± 0.04, 0.72 ± 0.61 mm; 0.95 ± 0.05, 1.05 ± 1.40 mm; and 0.95 ± 0.05, 1.08 ± 1.48 mm for bladder, prostate, rectum, left and right femoral heads, respectively. As compared to the other two competing methods, our method has shown superior performance in terms of the segmentation accuracy.

CONCLUSION

We developed a deep learning-based segmentation method to rapidly and accurately segment five pelvic organs simultaneously from daily CBCTs. The proposed method could be used in the clinic to support rapid target and organs-at-risk contouring for prostate adaptive radiation therapy.

Cross-Domain Data Augmentation for Deep-Learning-Based Male Pelvic Organ Segmentation in Cone Beam CT

For prostate cancer patients, large organ deformations occurring between radiotherapy treatment sessions create uncertainty about the doses delivered to the tumor and surrounding healthy organs. Segmenting those regions on cone beam CT (CBCT) scans acquired on treatment day would reduce such uncertainties. In this work, a 3D U-net deep-learning architecture was trained to segment bladder, rectum, and prostate on CBCT scans. Due to the scarcity of contoured CBCT scans, the training set was augmented with CT scans already contoured in the current clinical workflow. Our network was then tested on 63 CBCT scans. The Dice similarity coefficient (DSC) increased significantly with the number of CBCT and CT scans in the training set, reaching 0.874 ± 0.096 , 0.814 ± 0.055 , and 0.758 ± 0.101 for bladder, rectum, and prostate, respectively. This was about 10% better than conventional approaches based on deformable image registration between planning CT and treatment CBCT scans, except for prostate. Interestingly, adding 74 CT scans to the CBCT training set allowed maintaining high DSCs, while halving the number of CBCT scans. Hence, our work showed that although CBCT scans included artifacts, cross-domain augmentation of the training set was effective and could rely on large datasets available for planning CT scans.

Automatic segmentation of the clinical target volume and organs at risk in the planning CT for rectal cancer using deep dilated convolutional neural networks

PURPOSE

Delineation of the clinical target volume (CTV) and organs at risk (OARs) is very important for radiotherapy but is time-consuming and prone to inter-observer variation. Here, we proposed a novel deep dilated convolutional neural network (DDCNN)-based method for fast and consistent auto-segmentation of these structures.

METHODS

Our DDCNN method was an end-to-end architecture enabling fast training and testing. Specifically, it employed a novel multiple-scale convolutional architecture to extract multiple-scale context features in the early layers, which contain the original information on fine texture and boundaries and which are very useful for accurate auto-segmentation. In addition, it enlarged the receptive fields of dilated convolutions at the end of networks to capture complementary context features. Then, it replaced the fully connected layers with fully convolutional layers to achieve pixel-wise segmentation. We used data from 278 patients with rectal cancer for evaluation. The CTV and OARs were delineated and validated by senior radiation oncologists in the planning computed tomography (CT) images. A total of 218 patients chosen randomly were used for training, and the remaining 60 for validation. The Dice similarity coefficient (DSC) was used to measure segmentation accuracy.

RESULTS

Performance was evaluated on segmentation of the CTV and OARs. In addition, the performance of DDCNN was compared with that of U-Net. The proposed DDCNN method outperformed the U-Net for all segmentations, and the average DSC value of DDCNN was 3.8% higher than that of U-Net. Mean DSC values of DDCNN were 87.7% for the CTV, 93.4% for the bladder, 92.1% for the left femoral head, 92.3% for the right femoral head, 65.3% for the intestine, and 61.8% for the colon. The test time was 45 s per patient for segmentation of all the CTV, bladder, left and right femoral heads, colon, and intestine. We also assessed our approaches and results with those in the literature: our system showed superior performance and faster speed.

CONCLUSIONS

These data suggest that DDCNN can be used to segment the CTV and OARs accurately and efficiently. It was invariant to the body size, body shape, and age of the patients. DDCNN could improve the consistency of contouring and streamline radiotherapy workflows.

Dosimetric advantages of a clinical daily adaptive plan selection strategy compared with a non-adaptive strategy in cervical cancer radiation therapy

BACKGROUND

Radiation therapy (RT) using a daily plan selection adaptive strategy can be applied to account for interfraction organ motion while limiting organ at risk dose. The aim of this study was to quantify the dosimetric consequences of daily plan selection compared with non-adaptive RT in cervical cancer.

MATERIAL AND METHODS

Ten consecutive patients who received pelvic irradiation, planning CTs (full and empty bladder), weekly post-fraction CTs and pre-fraction CBCTs were included. Non-adaptive plans were generated based on the PTV defined using the full bladder planning CT. For the adaptive strategy, multiple PTVs were created based on both planning CTs by ITVs of the primary CTVs (i.e., GTV, cervix, corpus-uterus and upper part of the vagina) and corresponding library plans were generated. Daily CBCTs were rigidly aligned to the full bladder planning CT for plan selection. For daily plan recalculation, selected CTs based on initial similarity were deformably registered to CBCTs. Differences in daily target coverage (D_98% > 95%) and in V_0.5Gy, V_1.5Gy, V_2Gy, D_50% and D_2% for rectum, bladder and bowel were assessed.

RESULTS

Non-adaptive RT showed inadequate primary CTV coverage in 17% of the daily fractions. Plan selection compensated for anatomical changes and improved primary CTV coverage significantly (p < 0.01) to 98%. Compared with non-adaptive RT, plan selection decreased the fraction dose to rectum and bowel indicated by significant (p < 0.01) improvements for daily V_0.5Gy, V_1.5Gy, V_2Gy, D_50% and D_2%. However, daily plan selection significantly increased the bladder V_1.5Gy, V_2Gy, D_50% and D_2%.

CONCLUSIONS

In cervical cancer RT, a non-adaptive strategy led to inadequate target coverage for individual patients. Daily plan selection corrected for day-to-day anatomical variations and resulted in adequate target coverage in all fractions. The dose to bowel and rectum was decreased significantly when applying adaptive RT.

Dosimetric advantages of proton therapy compared with photon therapy using an adaptive strategy in cervical cancer

Background Image-guided adaptive proton therapy (IGAPT) can potentially be applied to take into account interfraction motion while limiting organ at risk (OAR) dose in cervical cancer radiation therapy (RT). In this study, the potential dosimetric advantages of IGAPT compared with photon-based image-guided adaptive RT (IGART) were investigated. Material and methods For 13 cervical cancer patients, full and empty bladder planning computed tomography (CT) images and weekly CTs were acquired. Based on both primary clinical target volumes (pCTVs) [i.e. gross tumor volume (GTV), cervix, corpus-uterus and upper part of the vagina] on planning CTs, the pretreatment observed full range primary internal target volume (pITV) was interpolated to derive pITV subranges. Given corresponding ITVs (i.e. pITVs including lymph nodes), patient-specific photon and proton plan libraries were generated. Using all weekly CTs, IGART and IGAPT treatments were simulated by selecting library plans and recalculating the dose. For each recalculated IGART and IGAPT fraction, CTV (i.e. pCTV including lymph nodes) coverage was assessed and differences in fractionated substitutes of dose-volume histogram (DVH) parameters (V15Gy, V30Gy, V45Gy, Dmean, D2cc) for bladder, bowel and rectum were tested for significance (Wilcoxon signed-rank test). Also, differences in toxicity-related DVH parameters (rectum V30Gy, bowel V45Gy) were approximated based on accumulated dose distributions. Results In 92% (96%) of all recalculated IGAPT (IGART) fractions adequate CTV coverage (V95% >98%) was obtained. All dose parameters for bladder, bowel and rectum, except the fractionated substitute for rectum V45Gy, were improved using IGAPT. Also, IGAPT reduced the mean dose to bowel, bladder and rectum significantly (p < 0.01). In addition, an average decrease of rectum V30Gy and bowel V45Gy indicated reductions in toxicity probabilities when using IGAPT. Conclusion This study demonstrates the feasibility of IGAPT in cervical cancer using a plan-library based plan-of-the-day approach. Compared to photon-based IGART, IGAPT maintains target coverage while significant dose reductions for the bladder, bowel and rectum can be achieved.

Beam configuration selection for robust intensity-modulated proton therapy in cervical cancer using Pareto front comparison

The Pareto front reflects the optimal trade-offs between conflicting objectives and can be used to quantify the effect of different beam configurations on plan robustness and dose-volume histogram parameters. Therefore, our aim was to develop and implement a method to automatically approach the Pareto front in robust intensity-modulated proton therapy (IMPT) planning. Additionally, clinically relevant Pareto fronts based on different beam configurations will be derived and compared to enable beam configuration selection in cervical cancer proton therapy. A method to iteratively approach the Pareto front by automatically generating robustly optimized IMPT plans was developed. To verify plan quality, IMPT plans were evaluated on robustness by simulating range and position errors and recalculating the dose. For five retrospectively selected cervical cancer patients, this method was applied for IMPT plans with three different beam configurations using two, three and four beams. 3D Pareto fronts were optimized on target coverage (CTV D(99%)) and OAR doses (rectum V30Gy; bladder V40Gy). Per patient, proportions of non-approved IMPT plans were determined and differences between patient-specific Pareto fronts were quantified in terms of CTV D(99%), rectum V(30Gy) and bladder V(40Gy) to perform beam configuration selection. Per patient and beam configuration, Pareto fronts were successfully sampled based on 200 IMPT plans of which on average 29% were non-approved plans. In all patients, IMPT plans based on the 2-beam set-up were completely dominated by plans with the 3-beam and 4-beam configuration. Compared to the 3-beam set-up, the 4-beam set-up increased the median CTV D(99%) on average by 0.2 Gy and decreased the median rectum V(30Gy) and median bladder V(40Gy) on average by 3.6% and 1.3%, respectively. This study demonstrates a method to automatically derive Pareto fronts in robust IMPT planning. For all patients, the defined four-beam configuration was found optimal in terms of plan robustness, target coverage and OAR sparing.