Metric-guided Image Reconstruction Bounds via Conformal Prediction

Recent advancements in machine learning have led to novel imaging systems and algorithms that address ill-posed problems. Assessing their trustworthiness and understanding how to deploy them safely at test time remains an important and open problem. We propose a method that leverages conformal prediction to retrieve upper/lower bounds and statistical inliers/outliers of reconstructions based on the prediction intervals of downstream metrics. We apply our method to sparse-view CT for downstream radiotherapy planning and show 1) that metric-guided bounds have valid coverage for downstream metrics while conventional pixel-wise bounds do not and 2) anatomical differences of upper/lower bounds between metric-guided and pixel-wise methods. Our work paves the way for more meaningful reconstruction bounds. Code available at https://github.com/matthewyccheung/conformal-metric.

Automatic end-to-end VMAT treatment planning for rectal cancers

BACKGROUND

The treatment planning process from segmentation to producing a deliverable plan is time-consuming and labor-intensive. Existing solutions automate the segmentation and planning processes individually. The feasibility of combining auto-segmentation and auto-planning for volumetric modulated arc therapy (VMAT) for rectal cancers in an end-to-end process is not clear.

PURPOSE

To create and clinically evaluate a complete end-to-end process for auto-segmentation and auto-planning of VMAT for rectal cancer requiring only the gross tumor volume contour and a CT scan as inputs.

METHODS

Patient scans and data were retrospectively selected from our institutional records for patients treated for malignant neoplasm of the rectum. We trained, validated, and tested deep learning auto-segmentation models using nnU-Net architecture for clinical target volume (CTV), bowel bag, large bowel, small bowel, total bowel, femurs, bladder, bone marrow, and female and male genitalia. For the CTV, we identified 174 patients with clinically drawn CTVs. We used data for 18 patients for all structures other than the CTV. The structures were contoured under the guidance of and reviewed by a gastrointestinal (GI) radiation oncologist. The predicted results for CTV in 35 patients and organs at risk (OAR) in six patients were scored by the GI radiation oncologist using a five-point Likert scale. For auto-planning, a RapidPlan knowledge-based planning solution was modeled for VMAT delivery with a prescription of 25 Gy in five fractions. The model was trained and tested on 20 and 34 patients, respectively. The resulting plans were scored by two GI radiation oncologists using a five-point Likert scale. Finally, the end-to-end pipeline was evaluated on 16 patients, and the resulting plans were scored by two GI radiation oncologists.

RESULTS

In 31 of 35 patients, CTV contours were clinically acceptable without necessary modifications. The CTV achieved a Dice similarity coefficient of 0.85 (±0.05) and 95% Hausdorff distance of 15.25 (±5.59) mm. All OAR contours were clinically acceptable without edits, except for large and small bowel which were challenging to differentiate. However, contours for total, large, and small bowel were clinically acceptable. The two physicians accepted 100% and 91% of the auto-plans. For the end-to-end pipeline, the two physicians accepted 88% and 62% of the auto-plans.

CONCLUSIONS

This study demonstrated that the VMAT treatment planning technique for rectal cancer can be automated to generate clinically acceptable and safe plans with minimal human interventions.

SC-GAN: Structure-completion generative adversarial network for synthetic CT generation from MR images with truncated anatomy

Creating synthetic CT (sCT) from magnetic resonance (MR) images enables MR-based treatment planning in radiation therapy. However, the MR images used for MR-guided adaptive planning are often truncated in the boundary regions due to the limited field of view and the need for sequence optimization. Consequently, the sCT generated from these truncated MR images lacks complete anatomic information, leading to dose calculation error for MR-based adaptive planning. We propose a novel structure-completion generative adversarial network (SC-GAN) to generate sCT with full anatomic details from the truncated MR images. To enable anatomy compensation, we expand input channels of the CT generator by including a body mask and introduce a truncation loss between sCT and real CT. The body mask for each patient was automatically created from the simulation CT scans and transformed to daily MR images by rigid registration as another input for our SC-GAN in addition to the MR images. The truncation loss was constructed by implementing either an auto-segmentor or an edge detector to penalize the difference in body outlines between sCT and real CT. The experimental results show that our SC-GAN achieved much improved accuracy of sCT generation in both truncated and untruncated regions compared to the original cycleGAN and conditional GAN methods.

Artificial Intelligence-Based Radiotherapy Contouring and Planning to Improve Global Access to Cancer Care

PURPOSE

Increased automation has been identified as one approach to improving global cancer care. The Radiation Planning Assistant (RPA) is a web-based tool offering automated radiotherapy (RT) contouring and planning to low-resource clinics. In this study, the RPA workflow and clinical acceptability were assessed by physicians around the world.

METHODS

The RPA output for 75 cases was reviewed by at least three physicians; 31 radiation oncologists at 16 institutions in six countries on five continents reviewed RPA contours and plans for clinical acceptability using a 5-point Likert scale.

RESULTS

For cervical cancer, RPA plans using bony landmarks were scored as usable as-is in 81% (with minor edits 93%); using soft tissue contours, plans were scored as usable as-is in 79% (with minor edits 96%). For postmastectomy breast cancer, RPA plans were scored as usable as-is in 44% (with minor edits 91%). For whole-brain treatment, RPA plans were scored as usable as-is in 67% (with minor edits 99%). For head/neck cancer, the normal tissue autocontours were acceptable as-is in 89% (with minor edits 97%). The clinical target volumes (CTVs) were acceptable as-is in 40% (with minor edits 93%). The volumetric-modulated arc therapy (VMAT) plans were acceptable as-is in 87% (with minor edits 96%). For cervical cancer, the normal tissue autocontours were acceptable as-is in 92% (with minor edits 99%). The CTVs for cervical cancer were scored as acceptable as-is in 83% (with minor edits 92%). The VMAT plans for cervical cancer were acceptable as-is in 99% (with minor edits 100%).

CONCLUSION

The RPA, a web-based tool designed to improve access to high-quality RT in low-resource settings, has high rates of clinical acceptability by practicing clinicians around the world. It has significant potential for successful implementation in low-resource clinics.

Deep learning-based automatic segmentation of cardiac substructures for lung cancers

PURPOSE

Accurate and comprehensive segmentation of cardiac substructures is crucial for minimizing the risk of radiation-induced heart disease in lung cancer radiotherapy. We sought to develop and validate deep learning-based auto-segmentation models for cardiac substructures.

MATERIALS AND METHODS

Nineteen cardiac substructures (whole heart, 4 heart chambers, 6 great vessels, 4 valves, and 4 coronary arteries) in 100 patients treated for non-small cell lung cancer were manually delineated by two radiation oncologists. The valves and coronary arteries were delineated as planning risk volumes. An nnU-Net auto-segmentation model was trained, validated, and tested on this dataset with a split ratio of 75:5:20. The auto-segmented contours were evaluated by comparing them with manually drawn contours in terms of Dice similarity coefficient (DSC) and dose metrics extracted from clinical plans. An independent dataset of 42 patients was used for subjective evaluation of the auto-segmentation model by 4 physicians.

RESULTS

The average DSCs were 0.95 (+/- 0.01) for the whole heart, 0.91 (+/- 0.02) for 4 chambers, 0.86 (+/- 0.09) for 6 great vessels, 0.81 (+/- 0.09) for 4 valves, and 0.60 (+/- 0.14) for 4 coronary arteries. The average absolute errors in mean/max doses to all substructures were 1.04 (+/- 1.99) Gy and 2.20 (+/- 4.37) Gy. The subjective evaluation revealed that 94% of the auto-segmented contours were clinically acceptable.

CONCLUSION

We demonstrated the effectiveness of our nnU-Net model for delineating cardiac substructures, including coronary arteries. Our results indicate that this model has promise for studies regarding radiation dose to cardiac substructures.

Random Forest Modeling of Acute Toxicity in Anal Cancer: Effects of Peritoneal Cavity Contouring Approaches on Model Performance

PURPOSE

Our purpose was to analyze the effect on gastrointestinal (GI) toxicity models when their dose-volume metrics predictors are derived from segmentations of the peritoneal cavity after different contouring approaches.

METHODS AND MATERIALS

A random forest machine learning approach was used to predict acute grade ≥3 GI toxicity from dose-volume metrics and clinicopathologic factors for 246 patients (toxicity incidence = 9.5%) treated with definitive chemoradiation for squamous cell carcinoma of the anus. Three types of random forest models were constructed based on different bowel bag segmentation approaches: (1) physician-delineated after Radiation Therapy Oncology Group (RTOG) guidelines, (2) autosegmented by a deep learning model (nnU-Net) following RTOG guidelines, and (3) autosegmented but spanning the entire bowel space. Each model type was evaluated using repeated cross-validation (100 iterations; 50%/50% training/test split). The performance of the models was assessed using area under the precision-recall curve (AUPRC) and the receiver operating characteristic curve (AUROCC), as well as optimal F1 score.

RESULTS

When following RTOG guidelines, the models based on the nnU-Net auto segmentations (mean values: AUROCC, 0.71 ± 0.07; AUPRC, 0.42 ± 0.09; F1 score, 0.46 ± 0.08) significantly outperformed (P < .001) those based on the physician-delineated contours (mean values: AUROCC, 0.67 ± 0.07; AUPRC, 0.34 ± 0.08; F1 score, 0.36 ± 0.07). When spanning the entire bowel space, the performance of the autosegmentation models improved considerably (mean values: AUROCC, 0.87 ± 0.05; AUPRC, 0.70 ± 0.09; F1 score, 0.68 ± 0.09).

CONCLUSIONS

Random forest models were superior at predicting acute grade ≥3 GI toxicity when based on RTOG-defined bowel bag autosegmentations rather than physician-delineated contours. Models based on autosegmentations spanning the entire bowel space show further considerable improvement in model performance. The results of this study should be further validated using an external data set.

Clinical acceptability of automatically generated lymph node levels and structures of deglutition and mastication for head and neck radiation therapy

Background and Purpose

Auto-contouring of complex anatomy in computed tomography (CT) scans is a highly anticipated solution to many problems in radiotherapy. In this study, artificial intelligence (AI)-based auto-contouring models were clinically validated for lymph node levels and structures of swallowing and chewing in the head and neck.

Materials and Methods

CT scans of 145 head and neck radiotherapy patients were retrospectively curated. One cohort (n = 47) was used to analyze seven lymph node levels and the other (n = 98) used to analyze 17 swallowing and chewing structures. Separate nnUnet models were trained and validated using the separate cohorts. For the lymph node levels, preference and clinical acceptability of AI vs human contours were scored. For the swallowing and chewing structures, clinical acceptability was scored. Quantitative analyses of the test sets were performed for AI vs human contours for all structures using overlap and distance metrics.

Results

Median Dice Similarity Coefficient ranged from 0.77 to 0.89 for lymph node levels and 0.86 to 0.96 for chewing and swallowing structures. The AI contours were superior to or equally preferred to the manual contours at rates ranging from 75% to 91%; there was not a significant difference in clinical acceptability for nodal levels I-V for manual versus AI contours. Across all AI-generated lymph node level contours, 92% were rated as usable with stylistic to no edits. Of the 340 contours in the chewing and swallowing cohort, 4% required minor edits.

Conclusions

An accurate approach was developed to auto-contour lymph node levels and chewing and swallowing structures on CT images for patients with intact nodal anatomy. Only a small portion of test set auto-contours required minor edits.

Monitoring Variations in the Use of Automated Contouring Software

PURPOSE

Our purpose was to identify variations in the clinical use of automatically generated contours that could be attributed to software error, off-label use, or automation bias.

METHODS AND MATERIALS

For 500 head and neck patients who were contoured by an in-house automated contouring system, Dice similarity coefficient and added path length were calculated between the contours generated by the automated system and the final contours after editing for clinical use. Statistical process control was used and control charts were generated with control limits at 3 standard deviations. Contours that exceeded the thresholds were investigated to determine the cause. Moving mean control plots were then generated to identify dosimetrists who were editing less over time, which could be indicative of automation bias.

RESULTS

Major contouring edits were flagged for: 1.0% brain, 3.1% brain stem, 3.5% left cochlea, 2.9% right cochlea, 4.8% esophagus, 4.1% left eye, 4.0% right eye, 2.2% left lens, 4.9% right lens, 2.5% mandible, 11% left optic nerve, 6.1% right optic nerve, 3.8% left parotid, 5.9% right parotid, and 3.0% of spinal cord contours. Identified causes of editing included unexpected patient positioning, deviation from standard clinical practice, and disagreement between dosimetrist preference and automated contouring style. A statistically significant (P < .05) difference was identified between the contour editing practice of dosimetrists, with 1 dosimetrist editing more across all organs at risk. Eighteen percent (27/150) of moving mean control plots created for 5 dosimetrists indicated the amount of contour editing was decreasing over time, possibly corresponding to automation bias.

CONCLUSIONS

The developed system was used to detect statistically significant edits caused by software error, unexpected clinical use, and automation bias. The increased ability to detect systematic errors that occur when editing automatically generated contours will improve the safety of the automatic treatment planning workflow.

Fully-automated, CT-only GTV contouring for palliative head and neck radiotherapy

Planning for palliative radiotherapy is performed without the advantage of MR or PET imaging in many clinics. Here, we investigated CT-only GTV delineation for palliative treatment of head and neck cancer. Two multi-institutional datasets of palliative-intent treatment plans were retrospectively acquired: a set of 102 non-contrast-enhanced CTs and a set of 96 contrast-enhanced CTs. The nnU-Net auto-segmentation network was chosen for its strength in medical image segmentation, and five approaches separately trained: (1) heuristic-cropped, non-contrast images with a single GTV channel, (2) cropping around a manually-placed point in the tumor center for non-contrast images with a single GTV channel, (3) contrast-enhanced images with a single GTV channel, (4) contrast-enhanced images with separate primary and nodal GTV channels, and (5) contrast-enhanced images along with synthetic MR images with separate primary and nodal GTV channels. Median Dice similarity coefficient ranged from 0.6 to 0.7, surface Dice from 0.30 to 0.56, and 95th Hausdorff distance from 14.7 to 19.7 mm across the five approaches. Only surface Dice exhibited statistically-significant difference across these five approaches using a two-tailed Wilcoxon Rank-Sum test (p ≤ 0.05). Our CT-only results met or exceeded published values for head and neck GTV autocontouring using multi-modality images. However, significant edits would be necessary before clinical use in palliative radiotherapy.

Identifying the optimal deep learning architecture and parameters for automatic beam aperture definition in 3D radiotherapy

PURPOSE

Two-dimensional radiotherapy is often used to treat cervical cancer in low- and middle-income countries, but treatment planning can be challenging and time-consuming. Neural networks offer the potential to greatly decrease planning time through automation, but the impact of the wide range of hyperparameters to be set during training on model accuracy has not been exhaustively investigated. In the current study, we evaluated the effect of several convolutional neural network architectures and hyperparameters on 2D radiotherapy treatment field delineation.

METHODS

Six commonly used deep learning architectures were trained to delineate four-field box apertures on digitally reconstructed radiographs for cervical cancer radiotherapy. A comprehensive search of optimal hyperparameters for all models was conducted by varying the initial learning rate, image normalization methods, and (when appropriate) convolutional kernel size, the number of learnable parameters via network depth and the number of feature maps per convolution, and nonlinear activation functions. This yielded over 1700 unique models, which were all trained until performance converged and then tested on a separate dataset.

RESULTS

Of all hyperparameters, the choice of initial learning rate was most consistently significant for improved performance on the test set, with all top-performing models using learning rates of 0.0001. The optimal image normalization was not consistent across architectures. High overlap (mean Dice similarity coefficient = 0.98) and surface distance agreement (mean surface distance < 2 mm) were achieved between the treatment field apertures for all architectures using the identified best hyperparameters. Overlap Dice similarity coefficient (DSC) and distance metrics (mean surface distance and Hausdorff distance) indicated that DeepLabv3+ and D-LinkNet architectures were least sensitive to initial hyperparameter selection.

CONCLUSION

DeepLabv3+ and D-LinkNet are most robust to initial hyperparameter selection. Learning rate, nonlinear activation function, and kernel size are also important hyperparameters for improving performance.

Synthetic Megavoltage Cone Beam Computed Tomography Image Generation for Improved Contouring Accuracy of Cardiac Pacemakers

In this study, we aimed to enhance the contouring accuracy of cardiac pacemakers by improving their visualization using deep learning models to predict MV CBCT images based on kV CT or CBCT images. Ten pacemakers and four thorax phantoms were included, creating a total of 35 combinations. Each combination was imaged on a Varian Halcyon (kV/MV CBCT images) and Siemens SOMATOM CT scanner (kV CT images). Two generative adversarial network (GAN)-based models, cycleGAN and conditional GAN (cGAN), were trained to generate synthetic MV (sMV) CBCT images from kV CT/CBCT images using twenty-eight datasets (80%). The pacemakers in the sMV CBCT images and original MV CBCT images were manually delineated and reviewed by three users. The Dice similarity coefficient (DSC), 95% Hausdorff distance (HD95), and mean surface distance (MSD) were used to compare contour accuracy. Visual inspection showed the improved visualization of pacemakers on sMV CBCT images compared to original kV CT/CBCT images. Moreover, cGAN demonstrated superior performance in enhancing pacemaker visualization compared to cycleGAN. The mean DSC, HD95, and MSD for contours on sMV CBCT images generated from kV CT/CBCT images were 0.91 ± 0.02/0.92 ± 0.01, 1.38 ± 0.31 mm/1.18 ± 0.20 mm, and 0.42 ± 0.07 mm/0.36 ± 0.06 mm using the cGAN model. Deep learning-based methods, specifically cycleGAN and cGAN, can effectively enhance the visualization of pacemakers in thorax kV CT/CBCT images, therefore improving the contouring precision of these devices.

Deep learning-based dose prediction to improve the plan quality of volumetric modulated arc therapy for gynecologic cancers

BACKGROUND

In recent years, deep-learning models have been used to predict entire three-dimensional dose distributions. However, the usability of dose predictions to improve plan quality should be further investigated.

PURPOSE

To develop a deep-learning model to predict high-quality dose distributions for volumetric modulated arc therapy (VMAT) plans for patients with gynecologic cancer and to evaluate their usability in driving plan quality improvements.

METHODS

A total of 79 VMAT plans for the female pelvis were used to train (47 plans), validate (16 plans), and test (16 plans) 3D dense dilated U-Net models to predict 3D dose distributions. The models received the normalized CT scan, dose prescription, and target and normal tissue contours as inputs. Three models were used to predict the dose distributions for plans in the test set. A radiation oncologist specializing in the treatment of gynecologic cancers scored the test set predictions using a 5-point scale (5, acceptable as-is; 4, prefer minor edits; 3, minor edits needed; 2, major edits needed; and 1, unacceptable). The clinical plans for which the dose predictions indicated that improvements could be made were reoptimized with constraints extracted from the predictions.

RESULTS

The predicted dose distributions in the test set were of comparable quality to the clinical plans. The mean voxel-wise dose difference was -0.14 ± 0.46 Gy. The percentage dose differences in the predicted target metrics ofD 1 % ${D}_{1{\mathrm{\% }}}$ andD 98 % ${D}_{98{\mathrm{\% }}}$ were -1.05% ± 0.59% and 0.21% ± 0.28%, respectively. The dose differences in the predicted organ at risk mean and maximum doses were -0.30 ± 1.66 Gy and -0.42 ± 2.07 Gy, respectively. A radiation oncologist deemed all of the predicted dose distributions clinically acceptable; 12 received a score of 5, and four received a score of 4. Replanning of flagged plans (five plans) showed that the original plans could be further optimized to give dose distributions close to the predicted dose distributions.

CONCLUSIONS

Deep-learning dose prediction can be used to predict high-quality and clinically acceptable dose distributions for VMAT female pelvis plans, which can then be used to identify plans that can be improved with additional optimization.

Clinical Acceptability of Automatically Generated Elective Lymph Node Volumes for Head and Neck Cancer Patients

PURPOSE/OBJECTIVE(S)

Manual contouring of head and neck lymph node levels is a time-intensive process prone to provider-specific variation. The purpose of this work is to generate a clinical segmentation tool while minimizing the amount of manual effort required by physicians to develop training datasets and review contours. Here we investigate an approach to curate, develop, and clinically validate an auto-contouring model for standard cervical lymph node volumes in the head and neck using a publicly available deep learning architecture. This model updates our previously validated tool to reflect modern practices in lymph node segmentation.

MATERIALS/METHODS

With the assistance of a resident physician, five radiation oncologists manually contoured individual lymph node levels on CT scans for three separate patients treated definitively with radiation or chemoradiation for oropharynx cancer, resulting in 15 unique ground truth cases. These cases were then used to train an nnUnet deep-learning model to generate automated contours for 32 additional cases. These 32 cases were reviewed, manually edited, and used to create the final model. Finally, the model was used to generate contours on the original 15 CT scans (testing cohort), and providers compared these automated contours with the ground-truth (manual) contours. Two blinded studies were performed. In a double-blinded fashion, providers were first asked to select which set of contours they would prefer to use in clinical practice as a starting point for actual cases. Second, they scored each contour on a Likert scale (1-5) to indicate clinical acceptability, ranging from completely unusable to usable without modification.

RESULTS

Across all lymph node levels (IA, IB, II, III, IV, V, RP), average Dice Similarity Coefficient ranged from 0.77 to 0.89 for AI vs manual contours in the testing cohort. These AI and manual lymph node contours were reviewed by 5 physicians each, resulting in 525 preference scores. Across all lymph nodes, the AI contour was superior to or equally preferred to the manual contours at rates ranging from 75% to 91% in the first blinded study. In the second blinded study, physician preference for the manual vs AI contour was statistically different for only the RP contours (p < 0.01). Thus, there was not a significant difference in clinical acceptability for nodal levels I-V for manual versus AI contours. Across all physician-generated contours, 82% were rated as usable with stylistic to no edits, and across all AI-generated contours, 92% were rated as usable with stylistic to no edits.

CONCLUSION

An approach to generate clinically acceptable automated contours for cervical lymph node levels in the head and neck was demonstrated. Furthermore, for nodal levels I-V, there was no significant difference in clinical acceptability in manual vs AI contours. Because we were able to generate and validate a model for each lymph node level individually, the output is applicable to a complete range of disease in which cervical lymph nodes are treated.

Cobalt Compensator-Based IMRT for Gynecologic Cancer Treatment in Low- and Middle-Income Countries: Equivalence to LINAC-Based IMRT

PURPOSE/OBJECTIVE(S)

Cobalt intensity modulated radiation therapy (IMRT) has the potential to impact global oncology by improving access to precision radiation therapy, particularly in low- and middle-income countries, and a novel compensator-based machine has been developed. Previously, cobalt compensator-based IMRT and LINAC-based IMRT were shown to achieve similar target goals and dose constraints for head and neck clinical cases. Additionally, the fidelity of those calculations was supported by Monte Carlo simulations and gamma analysis. The current study investigates the use of cobalt compensator IMRT for another application - gynecologic cancer patients.

MATERIALS/METHODS

A commercial treatment planning system was previously commissioned for the cobalt compensator-based device using Monte Carlo simulations from a simulation toolkit. Patient-specific compensators were modeled. Clinical gynecologic cancer cases were planned and compared to clinical plans with a 6MV LINAC using IMRT. The prescribed dose was 45 Gy in 25 fractions. Dosimetric plan quality endpoints for the planning target volume (PTV) and organs-at-risk (OARs) were compared. Parametric t-testing was used for statistical analysis, and criteria for significance was p<0.05.

RESULTS

Dose statistics of 10 cases showed similar target coverage and normal tissue sparing between IMRT plans using a 6 MV LINAC and using a cobalt compensator-based device. Greater than 95% of the dose to greater than 95% of the PTV was achieved for all 10 cases using both the LINAC-based and cobalt compensator-based plans. There was no significant difference in the mean percentage of the PTV receiving 95% of the dose between LINAC-based and cobalt compensator-based plans (98.0 ± 1.88% vs 97.1 ± 1.90%, respectively, p = 0.313). Additionally, both the LINAC-based and cobalt compensator-based plans, for all 10 cases, met dose constraints for the OARs, including Spinal Cord Dmax < 4500 cGy, Kidney Dmean < 1800 cGy, Bladder Dmax < 5750 cGy, Rectum Dmax < 5500 cGy, Small Bowel Dmax < 5400 cGy, Bone Marrow V4000 < 37%, and Femoral Heads Dmax < 5000 cGy.

CONCLUSION

Cobalt compensator-based IMRT may provide comparable quality treatment plans to LINAC-based IMRT for gynecologic cancer patients. Physical development of the cobalt compensator device has been completed and will be commissioned for further research to assess clinical outcomes.

A real-time contouring feedback tool for consensus-based contour training

Purpose

Variability in contouring structures of interest for radiotherapy continues to be challenging. Although training can reduce such variability, having radiation oncologists provide feedback can be impractical. We developed a contour training tool to provide real-time feedback to trainees, thereby reducing variability in contouring.

Methods

We developed a novel metric termed localized signed square distance (LSSD) to provide feedback to the trainee on how their contour compares with a reference contour, which is generated real-time by combining trainee contour and multiple expert radiation oncologist contours. Nine trainees performed contour training by using six randomly assigned training cases that included one test case of the heart and left ventricle (LV). The test case was repeated 30 days later to assess retention. The distribution of LSSD maps of the initial contour for the training cases was combined and compared with the distribution of LSSD maps of the final contours for all training cases. The difference in standard deviations from the initial to final LSSD maps, ΔLSSD, was computed both on a per-case basis and for the entire group.

Results

For every training case, statistically significant ΔLSSD were observed for both the heart and LV. When all initial and final LSSD maps were aggregated for the training cases, before training, the mean LSSD ([range], standard deviation) was -0.8 mm ([-37.9, 34.9], 4.2) and 0.3 mm ([-25.1, 32.7], 4.8) for heart and LV, respectively. These were reduced to -0.1 mm ([-16.2, 7.3], 0.8) and 0.1 mm ([-6.6, 8.3], 0.7) for the final LSSD maps during the contour training sessions. For the retention case, the initial and final LSSD maps of the retention case were aggregated and were -1.5 mm ([-22.9, 19.9], 3.4) and -0.2 mm ([-4.5, 1.5], 0.7) for the heart and 1.8 mm ([-16.7, 34.5], 5.1) and 0.2 mm ([-3.9, 1.6],0.7) for the LV.

Conclusions

A tool that uses real-time contouring feedback was developed and successfully used for contour training of nine trainees. In all cases, the utility was able to guide the trainee and ultimately reduce the variability of the trainee's contouring.

A transformer-based hierarchical registration framework for multimodality deformable image registration

Deformable image registration (DIR) between daily and reference images is fundamentally important for adaptive radiotherapy. In the last decade, deep learning-based image registration methods have been developed with faster computation time and improved robustness compared to traditional methods. However, the registration performance is often degraded in extra-cranial sites with large volume containing multiple anatomic regions, such as Computed Tomography (CT)/Magnetic Resonance (MR) images used in head and neck (HN) radiotherapy. In this study, we developed a hierarchical deformable image registration (DIR) framework, Patch-based Registration Network (Patch-RegNet), to improve the accuracy and speed of CT-MR and MR-MR registration for head-and-neck MR-Linac treatments. Patch-RegNet includes three steps: a whole volume global registration, a patch-based local registration, and a patch-based deformable registration. Following a whole-volume rigid registration, the input images were divided into overlapping patches. Then a patch-based rigid registration was applied to achieve accurate local alignment for subsequent DIR. We developed a ViT-Morph model, a combination of a convolutional neural network (CNN) and the Vision Transformer (ViT), for the patch-based DIR. A modality independent neighborhood descriptor was adopted in our model as the similarity metric to account for both inter-modality and intra-modality registration. The CT-MR and MR-MR DIR models were trained with 242 CT-MR and 213 MR-MR image pairs from 36 patients, respectively, and both tested with 24 image pairs (CT-MR and MR-MR) from 6 other patients. The registration performance was evaluated with 7 manually contoured organs (brainstem, spinal cord, mandible, left/right parotids, left/right submandibular glands) by comparing with the traditional registration methods in Monaco treatment planning system and the popular deep learning-based DIR framework, Voxelmorph. Evaluation results show that our method outperformed VoxelMorph by 6 % for CT-MR registration, and 4 % for MR-MR registration based on DSC measurements. Our hierarchical registration framework has been demonstrated achieving significantly improved DIR accuracy of both CT-MR and MR-MR registration for head-and-neck MR-guided adaptive radiotherapy.

Commissioning and dosimetric validation of a novel compensator-based Co-60 IMRT system for evaluating suitability to automated treatment planning

PURPOSE

A novel compensator-based system has been proposed which delivers intensity-modulated radiation therapy (IMRT) with cobalt-60 beams. This could improve access to advanced radiotherapy in low- and middle-income countries. For this system to be clinically viable and to be adapted into the Radiation Planning Assistant (RPA), being developed to offer automated planning services in low- and middle-income countries, it is necessary to commission and validate it in a commercial treatment planning system (TPS).

METHODS

The novel treatment device considered here employs a cobalt-60 source and nine compensators. Each compensator is produced by 3-D printing a thin plastic mold which is then filled on-demand within the machine with reusable 2-mm-diameter spherical tungsten balls. This system was commissioned in the Eclipse TPS and validation tests were conducted with Monte Carlo using Geant4 Application for Tomographic Emission for percentage depth dose, in-plane profiles, penumbra, and IMRT dose validation. And the American Association of Physicists in Medicine Task Group 119 benchmarking testing was performed. Additionally, compensator-based cobalt-60 IMRT plans were created for 46 head-and-neck cancer cases and compared to the linac-based volumetric modulated arc therapy (VMAT) plans used clinically, then dosimetric parameters were evaluated. Beam-on time for each field was calculated. In addition, the measurement was also performed in a limited environment and compared with the Monte Carlo simulations.

RESULTS

The differences in percent depth doses and in-plane profiles between the Eclipse and Monte Carlo simulations were 0.65% ± 0.41% and 1.02% ± 0.99%, respectively, and the 80%-20% penumbra agreed within 0.46 ± 0.27 mm. For the Task Group 119 validation plans, all treatment planning goals were met and gamma passing rates were >95% (3%/3 mm criteria). In 46 clinical head-and-neck cases, the cobalt-60 compensator-based IMRT plans had planning target volume (PTV) coverages similar to linac-based VMAT plans: all dosimetric values for PTV were within 1.5%. The organs at risk dose parameters were somewhat higher in cobalt-60 compensator-based IMRT plans versus linac-based VMAT plans. The mean dose differences for the spinal cord, brain, and brainstem were 4.43 ± 1.92, 3.39 ± 4.67, and 2.40 ± 3.71 Gy, while those for the rest of the organs were <1 Gy. The average beam-on time per field was 0.42 ± 0.10 min for the 6 MV multi-leaf-collimator plans while those for the cobalt-60 compensator plans were 0.17 ± 0.01 and 0.31 ± 0.01 min at the dose rates of 350 and 175 cGy/min. There was a good agreement between in-plane profiles from measurements and Monte Carlo simulations, which differences are 1.34 ± 1.90% and 0.13 ± 2.16% for two different fields.

CONCLUSIONS

A novel compensator-based IMRT system using cobalt-60 beams was commissioned and validated in a commercial TPS. Plan quality with this system was comparable to that of linac-based plans in all test cases with shorter estimated beam-on times. This system enables reliable, high-quality plans with reduced cost and complexity and may have benefits for underserved regions of the world. This system is being integrated into the RPA, a web-based platform for auto-contouring and auto-planning.

Hazard testing to reduce risk in the development of automated planning tools

PURPOSE

Hazard scenarios were created to assess and reduce the risk of planning errors in automated planning processes. This was accomplished through iterative testing and improvement of examined user interfaces.

METHODS

Automated planning requires three user inputs: a computed tomography (CT), a prescription document, known as the service request, and contours. We investigated the ability of users to catch errors that were intentionally introduced into each of these three stages, according to an FMEA analysis. Five radiation therapists each reviewed 15 patient CTs, containing three errors: inappropriate field of view, incorrect superior border, and incorrect identification of isocenter. Four radiation oncology residents reviewed 10 service requests, containing two errors: incorrect prescription and treatment site. Four physicists reviewed 10 contour sets, containing two errors: missing contour slices and inaccurate target contour. Reviewers underwent video training prior to reviewing and providing feedback for various mock plans.

RESULTS

Initially, 75% of hazard scenarios were detected in the service request approval. The visual display of prescription information was then updated to improve the detectability of errors based on user feedback. The change was then validated with five new radiation oncology residents who detected 100% of errors present. 83% of the hazard scenarios were detected in the CT approval portion of the workflow. For the contour approval portion of the workflow none of the errors were detected by physicists, indicating this step will not be used for quality assurance of contours. To mitigate the risk from errors that could occur at this step, radiation oncologists must perform a thorough review of contour quality prior to final plan approval.

CONCLUSIONS

Hazard testing was used to pinpoint the weaknesses of an automated planning tool and as a result, subsequent improvements were made. This study identified that not all workflow steps should be used for quality assurance and demonstrated the importance of performing hazard testing to identify points of risk in automated planning tools.

A deep learning-based approach for statistical robustness evaluation in proton therapy treatment planning: a feasibility study

OBJECTIVE

Robustness evaluation is critical in particle radiotherapy due to its susceptibility to uncertainties. However, the customary method for robustness evaluation only considers a few uncertainty scenarios, which are insufficient to provide a consistent statistical interpretation. We propose an artificial intelligence-based approach that overcomes this limitation by predicting a set of percentile dose values at every voxel and allows for the evaluation of planning objectives at specific confidence levels. &#xD;Approach: We built and trained a deep learning (DL) model to predict the 5th and 95th percentile dose distributions, which corresponds to the lower and upper bounds of a two-tailed 90% confidence interval (CI), respectively. Predictions were made directly from the nominal dose distribution and planning computed tomography scan. The data used to train and test the model consisted of proton plans from 543 prostate cancer patients. The ground truth percentile values were estimated for each patient using 600 dose recalculations representing randomly sampled uncertainty scenarios. For comparison, we also tested whether a common worst-case scenario (WCS) robustness evaluation (voxel-wise minimum and maximum) corresponding to a 90% CI could reproduce the ground truth 5th and 95th percentile doses. &#xD;Main Results: The percentile dose distributions predicted by DL yielded excellent agreements with the ground truth dose distributions, with mean dose errors below 0.15Gy and average gamma passing rates (GPR) at 1 mm/1% above 93.9, which were substantially better than the WCS dose distributions (mean dose error above 2.2Gy and GPR at 1 mm/1% below 54). We observed similar outcomes in a dose-volume histogram error analysis, where the DL predictions generally yielded smaller mean errors and standard deviations than the WCS evaluation doses. &#xD;Significance: The proposed method produces accurate and fast predictions (~2.5s for one percentile dose distribution) for a given confidence level. Thus, the method has the potential to improve robustness evaluation.

A deep-learning-based dose verification tool utilizing fluence maps for a cobalt-60 compensator-based intensity-modulated radiation therapy system

Background and purpose

A novel cobalt-60 compensator-based intensity-modulated radiation therapy (IMRT) system was developed for a resource-limited environment but lacked an efficient dose verification algorithm. The aim of this study was to develop a deep-learning-based dose verification algorithm for accurate and rapid dose predictions.

Materials and methods

A deep-learning network was employed to predict the doses from static fields related to beam commissioning. Inputs were a cube-shaped phantom, a beam binary mask, and an intersecting volume of the phantom and beam binary mask, while output was a 3-dimensional (3D) dose. The same network was extended to predict patient-specific doses for head and neck cancers using two different approaches. A field-based method predicted doses for each field and combined all calculated doses into a plan, while the plan-based method combined all nine fluences into a plan to predict doses. Inputs included patient computed tomography (CT) scans, binary beam masks, and fluence maps truncated to the patient's CT in 3D.

Results

For static fields, predictions agreed well with ground truths with average deviations of less than 0.5% for percent depth doses and profiles. Even though the field-based method showed excellent prediction performance for each field, the plan-based method showed better agreement between clinical and predicted dose distributions. The distributed dose deviations for all planned target volumes and organs at risk were within 1.3 Gy. The calculation speed for each case was within two seconds.

Conclusions

A deep-learning-based dose verification tool can accurately and rapidly predict doses for a novel cobalt-60 compensator-based IMRT system.

Evaluation of repeatability and reproducibility of radiomic features produced by the fan-beam KV-CT on a novel ring gantry-based PET/CT linear accelerator

BACKGROUND

The RefleXion^® X1 is a novel radiotherapy delivery system on a ring gantry equipped with fan-beam kV-CT and PET imaging subsystems. The day-to-day scanning variability of radiomics features must be evaluated before any attempt to utilize radiomics features PURPOSE: This study aims to characterize the repeatability and reproducibility of radiomic features produced by the RefleXion^® X1 kV-CT.

MATERIALS AND METHODS

The Credence Cartridge Radiomics (CCR) phantom includes six cartridges of varied materials. It was scanned ten times on the RefleXion X1 kVCT imaging subsystem over a three-month period using the two most frequently used scanning protocols (BMS and BMF). Fifty-five radiomic features were extracted for each ROI on each CT scan and analyzed using LifeX software. The coefficient of variation (COV) was computed to evaluate the repeatability. Intraclass correlation coefficient (ICC) and concordance correlation coefficient (CCC) were used to evaluate the repeatability and reproducibility of the scanned images using 0.9 as the threshold. This process is repeated on a GE PET-CT scanner using several built-in protocols as a comparison.

RESULTS

On average, 87% of the features on both scan protocols on the RefleXion X1 kVCT imaging subsystem can be considered repeatable as they met COV < 10% criteria. On GE PET-CT, this number is similar at 86%. When we tighten the criteria to COV<5%, the RefleXion X1 kVCT imaging subsystem showed much better repeatability with 81% of features on average whereas GE PET-CT showed only 73.5% on average. About 91% and 89% of the features with ICC>0.9 respectively for BMS and BMF protocols on RefleXion X1. On the other hand, the percentage of features with ICC>0.9 on GE PET-CT ranges from 67% to 82%. The RefleXion X1 kVCT imaging subsystem showed excellent intra-scanner reproducibility between the scanning protocols much better than the GE PET CT scanner. For the inter-scanner reproducibility, the percentage of features with CCC>0.9 ranged from 49% to 80%. between X1 and GE PET-CT scanning protocols.

CONCLUSIONS

Clinically useful CT radiomic features produced by the RefleXion X1 kVCT imaging subsystem are reproducible and stable over time, demonstrating its utility as a quantitative imaging platform. This article is protected by copyright. All rights reserved.

Bayesian feature selection for radiomics using reliability metrics

Introduction: Imaging of tumors is a standard step in diagnosing cancer and making subsequent treatment decisions. The field of radiomics aims to develop imaging based biomarkers using methods rooted in artificial intelligence applied to medical imaging. However, a challenging aspect of developing predictive models for clinical use is that many quantitative features derived from image data exhibit instability or lack of reproducibility across different imaging systems or image-processing pipelines.

Methods: To address this challenge, we propose a Bayesian sparse modeling approach for image classification based on radiomic features, where the inclusion of more reliable features is favored via a probit prior formulation.

Results: We verify through simulation studies that this approach can improve feature selection and prediction given correct prior information. Finally, we illustrate the method with an application to the classification of head and neck cancer patients by human papillomavirus status, using as our prior information a reliability metric quantifying feature stability across different imaging systems.

Automated Contouring and Planning in Radiation Therapy: What Is 'Clinically Acceptable'?

Developers and users of artificial-intelligence-based tools for automatic contouring and treatment planning in radiotherapy are expected to assess clinical acceptability of these tools. However, what is 'clinical acceptability'? Quantitative and qualitative approaches have been used to assess this ill-defined concept, all of which have advantages and disadvantages or limitations. The approach chosen may depend on the goal of the study as well as on available resources. In this paper, we discuss various aspects of 'clinical acceptability' and how they can move us toward a standard for defining clinical acceptability of new autocontouring and planning tools.

Dose Escalation for Pancreas SBRT: Potential and Limitations of using Daily Online Adaptive Radiation Therapy and an Iterative Isotoxicity Automated Planning Approach

Purpose

To determine the dosimetric limitations of daily online adaptive pancreas stereotactic body radiation treatment by using an automated dose escalation approach.

Methods and Materials

We collected 108 planning and daily computed tomography (CT) scans from 18 patients (18 patients × 6 CT scans) who received 5-fraction pancreas stereotactic body radiation treatment at MD Anderson Cancer Center. Dose metrics from the original non-dose-escalated clinical plan (non-DE), the dose-escalated plan created on the original planning CT (DE-ORI), and the dose-escalated plan created on daily adaptive radiation therapy CT (DE-ART) were analyzed. We developed a dose-escalation planning algorithm within the radiation treatment planning system to automate the dose-escalation planning process for efficiency and consistency. In this algorithm, the prescription dose of the dose-escalation plan was escalated before violating any organ-at-risk (OAR) dose constraint. Dose metrics for 3 targets (gross target volume [GTV], tumor vessel interface [TVI], and dose-escalated planning target volume [DE-PTV]) and 9 OARs (duodenum, large bowel, small bowel, stomach, spinal cord, kidneys, liver, and skin) for the 3 plans were compared. Furthermore, we evaluated the effectiveness of the online adaptive dose-escalation planning process by quantifying the effect of the interfractional dose distribution variations among the DE-ART plans.

Results

The median D_95% dose to the GTV/TVI/DE-PTV was 33.1/36.2/32.4 Gy, 48.5/50.9/40.4 Gy, and 53.7/58.2/44.8 Gy for non-DE, DE-ORI, and DE-ART, respectively. Most OAR dose constraints were not violated for the non-DE and DE-ART plans, while OAR constraints were violated for the majority of the DE-ORI patients due to interfractional motion and lack of adaptation. The maximum difference per fraction in D_95%, due to interfractional motion, was 2.5 ± 2.7 Gy, 3.0 ± 2.9 Gy, and 2.0 ± 1.8 Gy for the TVI, GTV, and DE-PTV, respectively.

Conclusions

Most patients require daily adaptation of the radiation planning process to maximally escalate delivered dose to the pancreatic tumor without exceeding OAR constraints. Using our automated approach, patients can receive higher target dose than standard of care without violating OAR constraints.

Cobalt compensator-based IMRT device: A treatment planning study of head and neck cases

PURPOSE

Our goal is to develop a novel cobalt-compensator-based IMRT device for low- and middle-income countries that is reliable and cost-effective while delivering treatment plans of equal quality to those from linac-MLC devices. The present study examines the quality of treatment plans using this device.

METHODS

A commercial treatment planning system (TPS; RayStation v.8B) was commissioned for this device using Monte Carlo simulations from the Geant4 toolkit. Patient-specific compensators were created as regions-of-interest. Thirty clinical head & neck cases were planned and compared to clinical plans with a 6MV linac using IMRT. The mock head and neck plan from TG-119 was used for further validation.

RESULTS

PTV objectives were achieved in all 30 plans with PTV V95% >95 %. OAR sparing was similar to clinical plans. There were 14 cases where OAR dose limits exceeded the recommended QUANTEC limits in the clinical plan in order to achieve target coverage. OAR sparing was better in the cobalt compensator plan in 8 cases and worse in 3 cases, in the latter cases exceeding the clinical plan doses by an average of 8.22 % (0.0 %-13.5 %). Average field-by-field gamma pass-rate were 93.7 % (2 %/2mm). Estimated treatment times using the Co-60 compensator device were 1 min 27 s vs 1 min 2 s for the clinical system.

CONCLUSION

This system is the first of its kind to allow for IMRT with a Co-60 device. Data here suggests that the delivery meets plan quality criteria while maintaining short treatment times which may offer a sustainable and cost-low option for IMRT on the global scale.

Compensation cycle consistent generative adversarial networks (Comp-GAN) for synthetic CT generation from MR scans with truncated anatomy

BACKGROUND

MR scans used in radiotherapy can be partially truncated due to the limited field of view (FOV), affecting dose calculation accuracy in MR-based radiation treatment planning.

PURPOSE

We proposed a novel Compensation-cycleGAN (Comp-cycleGAN) by modifying the cycle-consistent generative adversarial network (cycleGAN), to simultaneously create synthetic CT (sCT) images and compensate the missing anatomy from the truncated MR images.

METHODS

Computed tomography (CT) and T1 MR images with complete anatomy of 79 head-and-neck patients were used for this study. The original MR images were manually cropped 10-25 mm off at the posterior head to simulate clinically truncated MR images. Fifteen patients were randomly chosen for testing and the rest of the patients were used for model training and validation. Both the truncated and original MR images were used in the Comp-cycleGAN training stage, which enables the model to compensate for the missing anatomy by learning the relationship between the truncation and known structures. After the model was trained, sCT images with complete anatomy can be generated by feeding only the truncated MR images into the model. In addition, the external body contours acquired from the CT images with full anatomy could be an optional input for the proposed method to leverage the additional information of the actual body shape for each test patient. The mean absolute error (MAE) of Hounsfield units (HU), peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM) were calculated between sCT and real CT images to quantify the overall sCT performance. To further evaluate the shape accuracy, we generated the external body contours for sCT and original MR images with full anatomy. The Dice similarity coefficient (DSC) and mean surface distance (MSD) were calculated between the body contours of sCT and original MR images for the truncation region to assess the anatomy compensation accuracy.

RESULTS

The average MAE, PSNR, and SSIM calculated over test patients were 93.1 HU/91.3 HU, 26.5 dB/27.4 dB, and 0.94/0.94 for the proposed Comp-cycleGAN models trained without/with body-contour information, respectively. These results were comparable with those obtained from the cycleGAN model which is trained and tested on full-anatomy MR images, indicating the high quality of the sCT generated from truncated MR images by the proposed method. Within the truncated region, the mean DSC and MSD were 0.85/0.89 and 1.3/0.7 mm for the proposed Comp-cycleGAN models trained without/with body contour information, demonstrating good performance in compensating the truncated anatomy.

CONCLUSIONS

We developed a novel Comp-cycleGAN model that can effectively create sCT with complete anatomy compensation from truncated MR images, which could potentially benefit the MRI-based treatment planning.

Deep Learning-Based Dose Prediction for Automated, Individualized Quality Assurance of Head and Neck Radiation Therapy Plans

PURPOSE

This study aimed to use deep learning-based dose prediction to assess head and neck (HN) plan quality and identify suboptimal plans.

METHODS AND MATERIALS

A total of 245 volumetric modulated arc therapy HN plans were created using RapidPlan knowledge-based planning (KBP). A subset of 112 high-quality plans was selected under the supervision of an HN radiation oncologist. We trained a 3D Dense Dilated U-Net architecture to predict 3-dimensional dose distributions using 3-fold cross-validation on 90 plans. Model inputs included computed tomography images, target prescriptions, and contours for targets and organs at risk (OARs). The model's performance was assessed on the remaining 22 test plans. We then tested the application of the dose prediction model for automated review of plan quality. Dose distributions were predicted on 14 clinical plans. The predicted versus clinical OAR dose metrics were compared to flag OARs with suboptimal normal tissue sparing using a 2 Gy dose difference or 3% dose-volume threshold. OAR flags were compared with manual flags by 3 HN radiation oncologists.

RESULTS

The predicted dose distributions were of comparable quality to the KBP plans. The differences between the predicted and KBP-planned D_1%,D_95%, and D_99% across the targets were within -2.53% ± 1.34%, -0.42% ± 1.27%, and -0.12% ± 1.97%, respectively, and the OAR mean and maximum doses were within -0.33 ± 1.40 Gy and -0.96 ± 2.08 Gy, respectively. For the plan quality assessment study, radiation oncologists flagged 47 OARs for possible plan improvement. There was high interphysician variability; 83% of physician-flagged OARs were flagged by only one of 3 physicians. The comparative dose prediction model flagged 63 OARs, including 30 of 47 physician-flagged OARs.

CONCLUSIONS

Deep learning can predict high-quality dose distributions, which can be used as comparative dose distributions for automated, individualized assessment of HN plan quality.

Clinical acceptability of fully automated external beam radiotherapy for cervical cancer with three different beam delivery techniques

Purpose

To fully automate CT‐based cervical cancer radiotherapy by automating contouring and planning for three different treatment techniques.

Methods

We automated three different radiotherapy planning techniques for locally advanced cervical cancer: 2D 4‐field‐box (4‐field‐box), 3D conformal radiotherapy (3D‐CRT), and volumetric modulated arc therapy (VMAT). These auto‐planning algorithms were combined with a previously developed auto‐contouring system. To improve the quality of the 4‐field‐box and 3D‐CRT plans, we used an in‐house, field‐in‐field (FIF) automation program. Thirty‐five plans were generated for each technique on CT scans from multiple institutions and evaluated by five experienced radiation oncologists from three different countries. Every plan was reviewed by two of the five radiation oncologists and scored using a 5‐point Likert scale.

Results

Overall, 87%, 99%, and 94% of the automatically generated plans were found to be clinically acceptable without modification for the 4‐field‐box, 3D‐CRT, and VMAT plans, respectively. Some customizations of the FIF configuration were necessary on the basis of radiation oncologist preference. Additionally, in some cases, it was necessary to renormalize the plan after it was generated to satisfy radiation oncologist preference.

Conclusion

Approximately, 90% of the automatically generated plans were clinically acceptable for all three planning techniques. This fully automated planning system has been implemented into the radiation planning assistant for further testing in resource‐constrained radiotherapy departments in low‐ and middle‐income countries.

Assessing the practicality of using a single knowledge‐based planning model for multiple linac vendors

Purpose

Knowledge‐based planning (KBP) has been shown to be an effective tool in quality control for intensity‐modulated radiation therapy treatment planning and generating high‐quality plans. Previous studies have evaluated its ability to create consistent plans across institutions and between planners within the same institution as well as its use as teaching tool for inexperienced planners. This study evaluates whether planning quality is consistent when using a KBP model to plan across different treatment machines.

Materials and methods

This study used a RapidPlan model (Varian Medical Systems) provided by the vendor, to which we added additional planning objectives, maximum dose limits, and planning structures, such that a clinically acceptable plan is achieved in a single optimization. This model was used to generate and optimize volumetric‐modulated arc therapy plans for a cohort of 50 patients treated for head‐neck cancer. Plans were generated using the following treatment machines: Varian 2100, Elekta Versa HD, and Varian Halcyon. A noninferiority testing methodology was used to evaluate the hypothesis that normal and target metrics in our autoplans were no worse than a set of clinically‐acceptable baseline plans by a margin of 1.8 Gy or 3% dose‐volume. The quality of these plans were also compared through the use of common clinical dose‐volume histogram criteria.

Results

The Versa HD met our noninferiority criteria for 23 of 34 normal and target metrics; while the Halcyon and Varian 2100 machines met our criteria for 24 of 34 and 26 of 34 metrics, respectively. The experimental plans tended to have less volume coverage for prescription dose planning target volume and larger hotspot volumes. However, comparable plans were generated across different treatment machines.

Conclusions

These results support the use of a head‐neck RapidPlan models in centralized planning workflows that support clinics with different linac models/vendors, although some fine‐tuning for targets may be necessary.

Development and validation of a checklist for use with automatically generated radiotherapy plans

Purpose

To develop a checklist that improves the rate of error detection during the plan review of automatically generated radiotherapy plans.

Methods

A custom checklist was developed using guidance from American Association of Physicists in Medicine task groups 275 and 315 and the results of a failure modes and effects analysis of the Radiation Planning Assistant (RPA), an automated contouring and treatment planning tool. The preliminary checklist contained 90 review items for each automatically generated plan. In the first study, eight physicists were recruited from our institution who were familiar with the RPA. Each physicist reviewed 10 artificial intelligence‐generated resident treatment plans from the RPA for safety and plan quality, five of which contained errors. Physicists performed plan checks, recorded errors, and rated each plan's clinical acceptability. Following a 2‐week break, physicists reviewed 10 additional plans with a similar distribution of errors using our customized checklist. Participants then provided feedback on the usability of the checklist and it was modified accordingly. In a second study, this process was repeated with 14 senior medical physics residents who were randomly assigned to checklist or no checklist for their reviews. Each reviewed 10 plans, five of which contained errors, and completed the corresponding survey.

Results

In the first study, the checklist significantly improved the rate of error detection from 3.4 ± 1.1 to 4.4 ± 0.74 errors per participant without and with the checklist, respectively (p = 0.02). Error detection increased by 20% when the custom checklist was utilized. In the second study, 2.9 ± 0.84 and 3.5 ± 0.84 errors per participant were detected without and with the revised checklist, respectively (p = 0.08). Despite the lack of statistical significance for this cohort, error detection increased by 18% when the checklist was utilized.

Conclusion

Our results indicate that the use of a customized checklist when reviewing automated treatment plans will result in improved patient safety.

Automatic contouring QA method using a deep learning-based autocontouring system

Purpose

To determine the most accurate similarity metric when using an independent system to verify automatically generated contours.

Methods

A reference autocontouring system (primary system to create clinical contours) and a verification autocontouring system (secondary system to test the primary contours) were used to generate a pair of 6 female pelvic structures (UteroCervix [uterus + cervix], CTVn [nodal clinical target volume (CTV)], PAN [para‐aortic lymph nodes], bladder, rectum, and kidneys) on 49 CT scans from our institution and 38 from other institutions. Additionally, clinically acceptable and unacceptable contours were manually generated using the 49 internal CT scans. Eleven similarity metrics (volumetric Dice similarity coefficient (DSC), Hausdorff distance, 95% Hausdorff distance, mean surface distance, and surface DSC with tolerances from 1 to 10 mm) were calculated between the reference and the verification autocontours, and between the manually generated and the verification autocontours. A support vector machine (SVM) was used to determine the threshold that separates clinically acceptable and unacceptable contours for each structure. The 11 metrics were investigated individually and in certain combinations. Linear, radial basis function, sigmoid, and polynomial kernels were tested using the combinations of metrics as inputs for the SVM.

Results

The highest contouring error detection accuracies were 0.91 for the UteroCervix, 0.90 for the CTVn, 0.89 for the PAN, 0.92 for the bladder, 0.95 for the rectum, and 0.97 for the kidneys and were achieved using surface DSCs with a thickness of 1, 2, or 3 mm. The linear kernel was the most accurate and consistent when a combination of metrics was used as an input for the SVM. However, the best model accuracy from the combinations of metrics was not better than the best model accuracy from a surface DSC as an input.

Conclusions

We distinguished clinically acceptable contours from clinically unacceptable contours with an accuracy higher than 0.9 for the targets and critical structures in patients with cervical cancer; the most accurate similarity metric was surface DSC with a thickness of 1, 2, or 3 mm.

Barriers and Facilitators of Implementing Automated Radiotherapy Planning: A Multisite Survey of Low- and Middle-Income Country Radiation Oncology Providers

PURPOSE

Global access to radiotherapy (RT) is inequitable, with obstacles to implementing modern technologies in low- and middle- income countries (LMICs). The Radiation Planning Assistant (RPA) is a web-based automated RT planning software package intended to increase accessibility of high-quality RT planning. We surveyed LMIC RT providers to identify barriers and facilitators of future RPA deployment and uptake.

METHODS

RT providers underwent a pilot RPA teaching session in sub-Saharan Africa (Botswana, South Africa, and Tanzania) and Central America (Guatemala). Thirty providers (30 of 33, 90.9% response rate) participated in a postsession survey.

RESULTS

Respondents included physicians (n = 10, 33%), physicists (n = 9, 30%), dosimetrists (n = 8, 27%), residents/registrars (n = 1, 3.3%), radiation therapists (n = 1, 3.3%), and administrators (n = 1, 3.3%). Overall, 86.7% expressed interest in RPA; more respondents expected that RPA would be usable in 2 years (80%) compared with now (60%). Anticipated barriers were lack of reliable internet (80%), potential subscription fees (60%), and need for functionality in additional disease sites (48%). Expected facilitators included decreased workload (80%), decreased planning time (72%), and ability to treat more patients (64%). Forty-four percent anticipated that RPA would help transition from 2-dimensional to 3-dimensional techniques and 48% from 3-dimensional to intensity-modulated radiation treatment. Of a maximum acceptability/feasibility score of 60, physicians (45.6, standard deviation [SD] = 7.5) and dosimetrists (44.3, SD = 9.1) had lower scores than the mean for all respondents (48.3, SD = 7.7) although variation in scores by roles was not significantly different (P = .21).

CONCLUSION

These data provide an early assessment and create an initial framework to identify stakeholder needs and establish priorities to address barriers and promote facilitators of RPA deployment and uptake across global sites, as well as to tailor to needs in LMICs.

Radiation therapy related cardiac disease risk in childhood cancer survivors: Updated dosimetry analysis from the Childhood Cancer Survivor Study

BACKGROUND AND PURPOSE

We previously evaluated late cardiac disease in long-term survivors in the Childhood Cancer Survivor Study (CCSS) based on heart radiation therapy (RT) doses estimated from an age-scaled phantom with a simple atlas-based heart model (H_Atlas). We enhanced our phantom with a high-resolution CT-based anatomically realistic and validated age-scalable cardiac model (H_Hybrid). We aimed to evaluate how this update would impact our prior estimates of RT-related late cardiac disease risk in the CCSS cohort.

METHODS

We evaluated 24,214 survivors from the CCSS diagnosed from 1970 to 1999. RT fields were reconstructed on an age-scaled phantom with H_Hybrid and mean heart dose (D_m), percent volume receiving ≥ 20 Gy (V₂₀) and ≥ 5 Gy with V₂₀ = 0 ( [Formula: see text] ) were calculated. We reevaluated cumulative incidences and adjusted relative rates of grade 3-5 Common Terminology Criteria for Adverse Events outcomes for any cardiac disease, coronary artery disease (CAD), and heart failure (HF) in association with D_m, V₂₀, and [Formula: see text] (as categorical variables). Dose-response relationships were evaluated using piecewise-exponential models, adjusting for attained age, sex, cancer diagnosis age, race/ethnicity, time-dependent smoking history, diagnosis year, and chemotherapy exposure and doses. For relative rates, D_m was also considered as a continuous variable.

RESULTS

Consistent with previous findings with H_Atlas, reevaluation using H_Hybrid dosimetry found that, D_m ≥ 10 Gy, V₂₀ ≥ 0.1%, and [Formula: see text]  ≥ 50% were all associated with increased cumulative incidences and relative rates for any cardiac disease, CAD, and HF. While updated risk estimates were consistent with previous estimates overall without statistically significant changes, there were some important and significant (P < 0.05) increases in risk with updated dosimetry for D_m in the category of 20 to 29.9 Gy and V₂₀ in the category of 30% to 79.9%. When changes in the linear dose-response relationship for D_m were assessed, the slopes of the dose response were steeper (P < 0.001) with updated dosimetry. Changes were primarily observed among individuals with chest-directed RT with prescribed doses ≥ 20 Gy.

CONCLUSION

These findings present a methodological advancement in heart RT dosimetry with improved estimates of RT-related late cardiac disease risk. While results are broadly consistent with our prior study, we report that, with updated cardiac dosimetry, risks of cardiac disease are significantly higher in two dose and volume categories and slopes of the Dm-specific RT-response relationships are steeper. These data support the use of contemporary RT to achieve lower heart doses for pediatric patients, particularly those requiring chest-directed RT.

VerSe: A Vertebrae labelling and segmentation benchmark for multi-detector CT images

Vertebral labelling and segmentation are two fundamental tasks in an automated spine processing pipeline. Reliable and accurate processing of spine images is expected to benefit clinical decision support systems for diagnosis, surgery planning, and population-based analysis of spine and bone health. However, designing automated algorithms for spine processing is challenging predominantly due to considerable variations in anatomy and acquisition protocols and due to a severe shortage of publicly available data. Addressing these limitations, the Large Scale Vertebrae Segmentation Challenge (VerSe) was organised in conjunction with the International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) in 2019 and 2020, with a call for algorithms tackling the labelling and segmentation of vertebrae. Two datasets containing a total of 374 multi-detector CT scans from 355 patients were prepared and 4505 vertebrae have individually been annotated at voxel level by a human-machine hybrid algorithm (https://osf.io/nqjyw/, https://osf.io/t98fz/). A total of 25 algorithms were benchmarked on these datasets. In this work, we present the results of this evaluation and further investigate the performance variation at the vertebra level, scan level, and different fields of view. We also evaluate the generalisability of the approaches to an implicit domain shift in data by evaluating the top-performing algorithms of one challenge iteration on data from the other iteration. The principal takeaway from VerSe: the performance of an algorithm in labelling and segmenting a spine scan hinges on its ability to correctly identify vertebrae in cases of rare anatomical variations. The VerSe content and code can be accessed at: https://github.com/anjany/verse.

Clinical implementation of automated treatment planning for whole-brain radiotherapy

The purpose of the study was to develop and clinically deploy an automated, deep learning‐based approach to treatment planning for whole‐brain radiotherapy (WBRT). We collected CT images and radiotherapy treatment plans to automate a beam aperture definition from 520 patients who received WBRT. These patients were split into training (n = 312), cross‐validation (n = 104), and test (n = 104) sets which were used to train and evaluate a deep learning model. The DeepLabV3+ architecture was trained to automatically define the beam apertures on lateral‐opposed fields using digitally reconstructed radiographs (DRRs).

For the beam aperture evaluation, 1st quantitative analysis was completed using a test set before clinical deployment and 2nd quantitative analysis was conducted 90 days after clinical deployment. The mean surface distance and the Hausdorff distances were compared in the anterior‐inferior edge between the clinically used and the predicted fields. Clinically used plans and deep‐learning generated plans were evaluated by various dose–volume histogram metrics of brain, cribriform plate, and lens.

The 1st quantitative analysis showed that the average mean surface distance and Hausdorff distance were 7.1 mm (±3.8 mm) and 11.2 mm (±5.2 mm), respectively, in the anterior–inferior edge of the field. The retrospective dosimetric comparison showed that brain dose coverage (D99%, D95%, D1%) of the automatically generated plans was 29.7, 30.3, and 32.5 Gy, respectively, and the average dose of both lenses was up to 19.0% lower when compared to the clinically used plans. Following the clinical deployment, the 2nd quantitative analysis showed that the average mean surface distance and Hausdorff distance between the predicted and clinically used fields were 2.6 mm (±3.2 mm) and 4.5 mm (±5.6 mm), respectively.

In conclusion, the automated patient‐specific treatment planning solution for WBRT was implemented in our clinic. The predicted fields appeared consistent with clinically used fields and the predicted plans were dosimetrically comparable.

Technical Note: Dose prediction for head and neck radiotherapy using a three-dimensional dense dilated U-net architecture

PURPOSE

Radiation therapy treatment planning is a time-consuming and iterative manual process. Consequently, plan quality varies greatly between and within institutions. Artificial intelligence shows great promise in improving plan quality and reducing planning times. This technical note describes our participation in the American Association of Physicists in Medicine Open Knowledge-Based Planning Challenge (OpenKBP), a competition to accurately predict radiation therapy dose distributions.

METHODS

A three-dimensional (3D) densely connected U-Net with dilated convolutions was developed to predict 3D dose distributions given contoured CT images of head and neck patients as input. While traditional augmentation techniques such as rotations and translations were explored, it was found that training on random patches alone resulted in the greatest model performance. A custom-weighted mean squared error loss function was employed. Finally, an ensemble of best-performing networks was used to generate the final challenge predictions.

RESULTS

Our team (SuperPod) placed second in the dose stream of the OpenKBP challenge. The average mean absolute difference between the predicted and clinical dose distributions of the testing dataset was 2.56 Gy. On average, the predicted normalized target DVH metrics were within 3% of the clinical plans, and the predicted organ at risk DVH metrics were within 2 Gy of the clinical plans.

CONCLUSIONS

The developed 3D dense dilated U-Net architecture can accurately predict 3D radiotherapy dose distributions and can be used as part of a fully automated radiation therapy planning pipeline.

18FDG positron emission tomography mining for metabolic imaging biomarkers of radiation-induced xerostomia in patients with oropharyngeal cancer

Purpose

Head and neck cancers radiotherapy (RT) is associated with inevitable injury to parotid glands and subsequent xerostomia. We investigated the utility of SUV derived from ¹⁸FDG-PET to develop metabolic imaging biomarkers (MIBs) of RT-related parotid injury.

Methods

Data for oropharyngeal cancer (OPC) patients treated with RT at our institution between 2005 and 2015 with available planning computed tomography (CT), dose grid, pre- & first post-RT ¹⁸FDG-PET-CT scans, and physician-reported xerostomia assessment at 3-6 months post-RT (Xero 3-6 ms) per CTCAE, was retrieved, following an IRB approval. A CT-CT deformable image co-registration followed by voxel-by-voxel resampling of pre & post-RT ¹⁸FDG activity and dose grid were performed. Ipsilateral (Ipsi) and contralateral (contra) parotid glands were sub-segmented based on the received dose in 5 Gy increments, i.e. 0-5 Gy, 5-10 Gy sub-volumes, etc. Median and dose-weighted SUV were extracted from whole parotid volumes and sub-volumes on pre- & post-RT PET scans, using in-house code that runs on MATLAB. Wilcoxon signed-rank and Kruskal-Wallis tests were used to test differences pre- and post-RT.

Results

432 parotid glands, belonging to 108 OPC patients treated with RT, were sub-segmented & analyzed. Xero 3-6 ms was reported as: non-severe (78.7%) and severe (21.3%). SUV- median values were significantly reduced post-RT, irrespective of laterality (p = 0.02). A similar pattern was observed in parotid sub-volumes, especially ipsi parotid gland sub-volumes receiving doses 10-50 Gy (p < 0.05). Kruskal-Wallis test showed a significantly higher mean RT dose in the contra parotid in the patients with more severe Xero 3-6mo (p = 0.03). Multiple logistic regression showed a combined clinical-dosimetric-metabolic imaging model could predict the severity of Xero 3-6mo; AUC = 0.78 (95%CI: 0.66-0.85; p < 0.0001).

Conclusion

We sought to quantify pre- and post-RT ¹⁸FDG-PET metrics of parotid glands in patients with OPC. Temporal dynamics of PET-derived metrics can potentially serve as MIBs of RT-related xerostomia in concert with clinical and dosimetric variables.

Beam energy metrics for the acceptance and quality assurance of Halcyon linear accelerator

Purpose

Establish and compare two metrics for monitoring beam energy changes in the Halcyon platform and evaluate the accuracy of these metrics across multiple Halcyon linacs.

Method

The first energy metric is derived from the diagonal normalized flatness (F_DN), which is defined as the ratio of the average measurements at a fixed off‐axis equal distance along the open profiles in two diagonals to the measurement at the central axis with an ionization chamber array (ICA). The second energy metric comes from the area ratio (AR) of the quad wedge (QW) profiles measured with the QW on the top of the ICA. Beam energy is changed by adjusting the magnetron current in a non‐clinical Halcyon. With D_10cm measured in water at each beam energy, the relationships between F_DN or AR energy metrics to D_10cm in water is established with linear regression across six energy settings. The coefficients from these regressions allow D_10cm(F_DN) calculation from F_DN using open profiles and D_10cm(QW) calculation from AR using QW profiles.

Results

Five Halcyon linacs from five institutions were used to evaluate the accuracy of the D_10cm(F_DN) and the D_10cm(QW) energy metrics by comparing to the D_10cm values computed from the treatment planning system (TPS) and D_10cm measured in water. For the five linacs, the D_10cm(F_DN) reported by the ICA based on F_DN from open profiles agreed with that calculated by TPS within –0.29 ± 0.23% and 0.61% maximum discrepancy; the D_10cm(QW) reported by the QW profiles agreed with that calculated by TPS within –0.82 ± 1.27% and –2.43% maximum discrepancy.

Conclusion

The F_DN‐based energy metric D_10cm(F_DN) can be used for acceptance testing of beam energy, and also for the verification of energy in periodic quality assurance (QA) processes.

Development and dosimetric assessment of an automatic dental artifact classification tool to guide artifact management techniques in a fully automated treatment planning workflow

Purpose:

We conducted our study to develop a tool capable of automatically detecting dental artifacts in a CT scan on a slice-by-slice basis and to assess the dosimetric impact of implementing the tool into the Radiation Planning Assistant (RPA), a web-based platform designed to fully automate the radiation therapy treatment planning process.

Methods:

We developed an automatic dental artifact identification tool and assessed the dosimetric impact of its use in the RPA. Three users manually annotated 83,676 head-and-neck (HN) CT slices (549 patients). Majority-voting was applied to the individual annotations to determine the presence or absence of dental artifacts. The patients were divided into train, cross-validation, and test data sets (ratio: 3:1:1, respectively). A random subset of images without dental artifacts was used to balance classes (1:1) in the training data set. The Inception-V3 deep learning model was trained with the binary cross-entropy loss function. With use of this model, we automatically identified artifacts on 15 RPA HN plans on a slice-by-slice basis and investigated three dental artifact management methods applied before and after volumetric modulated arc therapy (VMAT) plan optimization. The resulting dose distributions and target coverage were quantified.

Results:

Per-slice accuracy, sensitivity, and specificity were 99 %, 91 %, and 99 %, respectively. The model identified all patients with artifacts. Small dosimetric differences in total plan dose were observed between the various density-override methods (±1 Gy). For the pre- and post-optimized plans, 90 % and 99 %, respectively, of dose comparisons resulted in normal structure dose differences of ±1 Gy. Differences in the volume of structures receiving 95 % of the prescribed dose (V95[%]) were ≤0.25 % for 100 % of plans.

Conclusion:

The dosimetric impact of applying dental artifact management before and after artifact plan optimization was minor. Our results suggest that not accounting for dental artifacts in the current RPA workflow (where only post-optimization dental artifact management is possible) may result in minor dosimetric differences. If RPA users choose to override CT densities as a solution to managing dental artifacts, our results suggest segmenting the volume of the artifact and overriding its density to water is a safe option.

Computed Tomography Radiomics Kinetics as Early Imaging Correlates of Osteoradionecrosis in Oropharyngeal Cancer Patients

Osteoradionecrosis (ORN) is a major side-effect of radiation therapy in oropharyngeal cancer (OPC) patients. In this study, we demonstrate that early prediction of ORN is possible by analyzing the temporal evolution of mandibular subvolumes receiving radiation. For our analysis, we use computed tomography (CT) scans from 21 OPC patients treated with Intensity Modulated Radiation Therapy (IMRT) with subsequent radiographically-proven ≥ grade II ORN, at three different time points: pre-IMRT, 2-months, and 6-months post-IMRT. For each patient, radiomic features were extracted from a mandibular subvolume that developed ORN and a control subvolume that received the same dose but did not develop ORN. We used a Multivariate Functional Principal Component Analysis (MFPCA) approach to characterize the temporal trajectories of these features. The proposed MFPCA model performs the best at classifying ORN vs. Control subvolumes with an area under curve (AUC) = 0.74 [95% confidence interval (C.I.): 0.61–0.90], significantly outperforming existing approaches such as a pre-IMRT features model or a delta model based on changes at intermediate time points, i.e., at 2- and 6-month follow-up. This suggests that temporal trajectories of radiomics features derived from sequential pre- and post-RT CT scans can provide markers that are correlates of RT-induced mandibular injury, and consequently aid in earlier management of ORN.

Training deep-learning segmentation models from severely limited data

PURPOSE

To enable generation of high-quality deep learning segmentation models from severely limited contoured cases (e.g., ~10 cases).

METHODS

Thirty head and neck computed tomography (CT) scans with well-defined contours were deformably registered to 200 CT scans of the same anatomic site without contours. Acquired deformation vector fields were used to train a principal component analysis (PCA) model for each of the 30 contoured CT scans by capturing the mean deformation and most prominent variations. Each PCA model can produce an infinite number of synthetic CT scans and corresponding contours by applying random deformations. We used 300, 600, 1000, and 2000 synthetic CT scans and contours generated from one PCA model to train V-Net, a 3D convolutional neural network architecture, to segment parotid and submandibular glands. We repeated the training using same numbers of training cases generated from 7, 10, 20, and 30 PCA models, with the data distributed evenly between each PCA model. Performance of the segmentation models was evaluated with Dice similarity coefficients between auto-generated contours and physician-drawn contours on 162 test CT scans for parotid glands and another 21 test CT scans for submandibular glands.

RESULTS

Dice values varied with the number of synthetic CT scans and the number of PCA models used to train the network. By using 2000 synthetic CT scans generated from 10 PCA models, we achieved Dice values of 82.8% ± 6.8% for right parotid, 82.0% ± 6.9% for left parotid, and 74.2% ± 6.8% for submandibular glands. These results are comparable with those obtained from state-of-the-art auto-contouring approaches, including a deep learning network trained from more than 1000 contoured patients and a multi-atlas algorithm from 12 well-contoured atlases. Improvement was marginal when >10 PCA models or >2000 synthetic CT scans were used.

CONCLUSIONS

We demonstrated an effective data augmentation approach to train high-quality deep learning segmentation models from a limited number of well-contoured patient cases.

Impact of slice thickness, pixel size, and CT dose on the performance of automatic contouring algorithms

Purpose

To investigate the impact of computed tomography (CT) image acquisition and reconstruction parameters, including slice thickness, pixel size, and dose, on automatic contouring algorithms.

Methods

Eleven scans from patients with head‐and‐neck cancer were reconstructed with varying slice thicknesses and pixel sizes. CT dose was varied by adding noise using low‐dose simulation software. The impact of these imaging parameters on two in‐house auto‐contouring algorithms, one convolutional neural network (CNN)‐based and one multiatlas‐based system (MACS) was investigated for 183 reconstructed scans. For each algorithm, auto‐contours for organs‐at‐risk were compared with auto‐contours from scans with 3 mm slice thickness, 0.977 mm pixel size, and 100% CT dose using Dice similarity coefficient (DSC), Hausdorff distance (HD), and mean surface distance (MSD).

Results

Increasing the slice thickness from baseline value of 3 mm gave a progressive reduction in DSC and an increase in HD and MSD on average for all structures. Reducing the CT dose only had a relatively minimal effect on DSC and HD. The rate of change with respect to dose for both auto‐contouring methods is approximately 0. Changes in pixel size had a small effect on DSC and HD for CNN‐based auto‐contouring with differences in DSC being within 0.07. Small structures had larger deviations from the baseline values than large structures for DSC. The relative differences in HD and MSD between the large and small structures were small.

Conclusions

Auto‐contours can deviate substantially with changes in CT acquisition and reconstruction parameters, especially slice thickness and pixel size. The CNN was less sensitive to changes in pixel size, and dose levels than the MACS. The results contraindicated more restrictive values for the parameters should be used than a typical imaging protocol for head‐and‐neck.

Clinical Acceptability of Automated Radiation Treatment Planning for Head and Neck Cancer Using the Radiation Planning Assistant

PURPOSE

Radiation treatment planning for head and neck cancer is a complex process with much variability; automated treatment planning is a promising option to improve plan quality and efficiency. This study compared radiation plans generated from a fully automated radiation treatment planning system to plans generated manually that had been clinically approved and delivered.

METHODS AND MATERIALS

The study cohort consisted of 50 patients treated by a specialized head and neck cancer team at a tertiary care center. An automated radiation treatment planning system, the Radiation Planning Assistant, was used to create autoplans for all patients using their original, approved contours. Common dose-volume histogram (DVH) criteria were used to compare the quality of autoplans to the clinical plans. Fourteen radiation oncologists, each from a different institution, then reviewed and compared the autoplans and clinical plans in a blinded fashion.

RESULTS

Autoplans and clinical plans were very similar with regard to DVH metrics for coverage and critical structure constraints. Physician reviewers found both the clinical plans and autoplans acceptable for use; overall, 78% of the clinical plans and 88% of the autoplans were found to be usable as is (without any edits). When asked to choose which plan would be preferred for approval, 27% of physician reviewers selected the clinical plan, 47% selected the autoplan, 25% said both were equivalent, and 0% said neither. Hence, overall, 72% of physician reviewers believed the autoplan or either the clinical or autoplan was preferable.

CONCLUSIONS

Automated radiation treatment planning creates consistent, clinically acceptable treatment plans that meet DVH criteria and are found to be appropriate on physician review.

Tissue-specific deformable image registration using a spatial-contextual filter

Intensity-based deformable registration with spatial-invariant regularization generally fails when distinct motion exists across different types of tissues. The purpose of this work was to develop and validate a new regularization approach for deformable image registration that is tissue-specific and able to handle motion discontinuities. Our approach was built upon a Demons registration framework, and used the image context supplementing the original spatial constraint to regularize displacement vector fields in iterative image registration process. The new regularization was implemented as a spatial-contextual filter, which favors the motion vectors within the same tissue type but penalizes the motion vectors from different tissues. This approach was validated using five public lung cancer patients, each with 300 landmark pairs identified by a thoracic radiation oncologist. The mean and standard deviation of the landmark registration errors were 1.3 ± 0.8 mm, compared with those of 2.3 ± 2.9 mm using the original Demons algorithm. Particularly, for the case with the largest initial landmark displacement of 15 ± 9 mm, the modified Demons algorithm had a registration error of 1.3 ± 1.1 mm, while the original Demons algorithm had a registration error of 3.6 ± 5.9 mm. We also qualitatively evaluated the modified Demons algorithm using two difficult cases in our routine clinic: one lung case with large sliding motion and one head and neck case with large anatomical changes in air cavity. Visual evaluation on the deformed image created by the deformable image registration showed that the modified Demons algorithm achieved reasonable registration accuracy, but the original Demons algorithm produced distinct registration errors.

The Emergence of Artificial Intelligence within Radiation Oncology Treatment Planning

BACKGROUND

The future of artificial intelligence (AI) heralds unprecedented change for the field of radiation oncology. Commercial vendors and academic institutions have created AI tools for radiation oncology, but such tools have not yet been widely adopted into clinical practice. In addition, numerous discussions have prompted careful thoughts about AI's impact upon the future landscape of radiation oncology: How can we preserve innovation, creativity, and patient safety? When will AI-based tools be widely adopted into the clinic? Will the need for clinical staff be reduced? How will these devices and tools be developed and regulated?

SUMMARY

In this work, we examine how deep learning, a rapidly emerging subset of AI, fits into the broader historical context of advancements made in radiation oncology and medical physics. In addition, we examine a representative set of deep learning-based tools that are being made available for use in external beam radiotherapy treatment planning and how these deep learning-based tools and other AI-based tools will impact members of the radiation treatment planning team. Key Messages: Compared to past transformative innovations explored in this article, such as the Monte Carlo method or intensity-modulated radiotherapy, the development and adoption of deep learning-based tools is occurring at faster rates and promises to transform practices of the radiation treatment planning team. However, accessibility to these tools will be determined by each clinic's access to the internet, web-based solutions, or high-performance computing hardware. As seen by the trends exhibited by many technologies, high dependence on new technology can result in harm should the product fail in an unexpected manner, be misused by the operator, or if the mitigation to an expected failure is not adequate. Thus, the need for developers and researchers to rigorously validate deep learning-based tools, for users to understand how to operate tools appropriately, and for professional bodies to develop guidelines for their use and maintenance is essential. Given that members of the radiation treatment planning team perform many tasks that are automatable, the use of deep learning-based tools, in combination with other automated treatment planning tools, may refocus tasks performed by the treatment planning team and may potentially reduce resource-related burdens for clinics with limited resources.