Effects of model size and composition on quality of head-and-neck knowledge-based plans.

PURPOSE: Knowledge-based planning (KBP) aims to automate and standardize treatment planning. New KBP users are faced with many questions: How much does model size matter, and are multiple models needed to accommodate specific physician preferences? In this study, six head-and-neck KBP models were trained to address these questions. METHODS: The six models differed in training size and plan composition: The KBPFull (n = 203 plans), KBP101 (n = 101), KBP50 (n = 50), and KBP25 (n = 25) were trained with plans from two head-and-neck physicians. KBPA and KBPB each contained n = 101 plans from only one physician, respectively. An independent set of 39 patients treated to 6000-7000 cGy by a third physician was re-planned with all KBP models for validation. Standard head-and-neck dosimetric parameters were used to compare resulting plans. KBPFull plans were compared to the clinical plans to evaluate overall model quality. Additionally, clinical and KBPFull plans were presented to another physician for blind review. Dosimetric comparison of KBPFull against KBP101 , KBP50 , and KBP25 investigated the effect of model size. Finally, KBPA versus KBPB tested whether training KBP models on plans from one physician only influences the resulting output. Dosimetric differences were tested for significance using a paired t-test (p < 0.05). RESULTS: Compared to manual plans, KBPFull significantly increased PTV Low D95% and left parotid mean dose but decreased dose cochlea, constrictors, and larynx. The physician preferred the KBPFull plan over the manual plan in 20/39 cases. Dosimetric differences between KBPFull , KBP101 , KBP50 , and KBP25 plans did not exceed 187 cGy on aggregate, except for the cochlea. Further, average differences between KBPA and KBPB were below 110 cGy. CONCLUSIONS: Overall, all models were shown to produce high-quality plans. Differences between model outputs were small compared to the prescription. This indicates only small improvements when increasing model size and minimal influence of the physician when choosing treatment plans for training head-and-neck KBP models.

Evaluation of knowledge-based planning models for male pelvic CBCT-based online adaptive radiotherapy on conventional linear accelerators.

PURPOSE: To assess the practicality of employing a commercial knowledge-based planning tool (RapidPlan) to generate adapted intact prostate and prostate bed volumetric modulated arc therapy (VMAT) plans on iterative cone-beam computed tomography (iCBCT) datasets. METHODS AND MATERIALS: Intact prostate and prostate bed RapidPlan models were trained utilizing planning data from 50 and 44 clinical cases, respectively. To ensure that refined models were capable of producing adequate clinical plans with a single optimization, models were tested with 50 clinical planning CT datasets by comparing dose-volume histogram (DVH) and plan quality metric (PQM) values between clinical and RapidPlan-generated plans. The RapidPlan tool was then used to retrospectively generate adapted VMAT plans on daily iCBCT images for 20 intact prostate and 15 prostate bed cases. As before, DVH and PQM metrics were utilized to dosimetrically compare scheduled (iCBCT Verify) and adapted (iCBCT RapidPlan) plans. Timing data was collected to further evaluate the feasibility of integrating this approach within an online adaptive radiotherapy workflow. RESULTS: Model testing results confirmed the models were capable of producing VMAT plans within a single optimization that were overall improved upon or dosimetrically comparable to original clinical plans. Direct application of RapidPlan on iCBCT datasets produced satisfactory intact prostate and prostate bed plans with generally improved target volume coverage/conformality and rectal sparing relative to iCBCT Verify plans as indicated by DVH values, though bladder metrics were marginally increased on average. Average PQM values for iCBCT RapidPlans were significantly improved compared to iCBCT Verify plans. The average time required [in mm:ss] to generate adapted plans was 06:09 ± 02:06 (intact) and 07:12 ± 01:04 (bed). CONCLUSION: This study demonstrated the feasibility of leveraging RapidPlan to expeditiously generate adapted VMAT intact prostate and prostate bed plans on iCBCT datasets. In general, adapted plans were dosimetrically improved relative to scheduled plans, emphasizing the practicality of the proposed approach.

Multi-planner validation of RapidPlan knowledge-based model for volumetric modulated arc therapy in prostate cancer.

PURPOSE: To investigate the performance of a model-based optimization process for volumetric modulated arc therapy (VMAT) applied to prostate cancer patients with the multi-planner. METHODS AND MATERIALS: The 120 prostate plans for VMAT treatment were entered into the database system of the RapidPlan (RP) knowledge-based treatment planning. The treatment planning data for each plan was used to create and train the RP model. Twelve prostate cancer cases were selected and were used for planning by a manual of 12 planners based on the clinical protocol for dose constraints. Then, the treatment plans for each patient were compared with the RP model plans and analyzed with Wilcoxon tests. RESULTS: On average, the RP models can estimate comparable doses among all planner plans and clinical plans for the PTV, which Dmax , D95% , D98% , HI, and CI were used to evaluate. For the normal organ doses of the bladder, rectum, penile bulb, and femoral head, all RP model plans showed comparable or better dose sparing than all planner plans and clinical plans. Moreover, the average planning time of the RP model was faster than manual plans by about two times. The RP model can significantly reduce the variation dose of the normal organs compared with the manual plans among the planners. CONCLUSION: The automated plans of the RP model might benefit from further fine-tuning of the dose constraints of the normal organs, although both procedure plans are acceptable and fulfill the clinical protocol goals so that the RP model can enhance the efficacy and quality of plans.

Novel in-house knowledge-based automated planning system for lung cancer treated with intensity-modulated radiotherapy.

PURPOSE: The goal of this study was to propose a knowledge-based planning system which could automatically design plans for lung cancer patients treated with intensity-modulated radiotherapy (IMRT). METHODS AND MATERIALS: From May 2018 to June 2020, 612 IMRT treatment plans of lung cancer patients were retrospectively selected to construct a planning database. Knowledge-based planning (KBP) architecture named αDiar was proposed in this study. It consisted of two parts separated by a firewall. One was the in-hospital workstation, and the other was the search engine in the cloud. Based on our previous study, A­Net in the in-hospital workstation was used to generate predicted virtual dose images. A search engine including a three-dimensional convolutional neural network (3D CNN) was constructed to derive the feature vectors of dose images. By comparing the similarity of the features between virtual dose images and the clinical dose images in the database, the most similar feature was found. The optimization parameters (OPs) of the treatment plan corresponding to the most similar feature were assigned to the new plan, and the design of a new treatment plan was automatically completed. After αDiar was developed, we performed two studies. The first retrospective study was conducted to validate whether this architecture was qualified for clinical practice and involved 96 patients. The second comparative study was performed to investigate whether αDiar could assist dosimetrists in improving the quality of planning for the patients. Two dosimetrists were involved and designed plans for only one trial with and without αDiar; 26 patients were involved in this study. RESULTS: The first study showed that about 54% (52/96) of the automatically generated plans would achieve the dosimetric constraints of the Radiation Therapy Oncology Group (RTOG) and about 93% (89/96) of the automatically generated plans would achieve the dosimetric constraints of the National Comprehensive Cancer Network (NCCN). The second study showed that the quality of treatment planning designed by junior dosimetrists was improved with the help of αDiar. CONCLUSIONS: Our results showed that αDiar was an effective tool to improve planning quality. Over half of the patients' plans could be designed automatically. For the remaining patients, although the automatically designed plans did not fully meet the clinical requirements, their quality was also better than that of manual plans.

Improving knowledge-based treatment planning for lung cancer radiotherapy with automatic multi-criteria optimized training plans.

BACKGROUND: Knowledge-based planning (KBP) is a method for automated radiotherapy treatment planning where appropriate optimization objectives for new patients are predicted based on a library of training plans. KBP can save time and improve organ at-risk sparing and inter-patient consistency compared to manual planning, but its performance depends on the quality of the training plans. We used another system for automated planning, which generates multi-criteria optimized (MCO) plans based on a wish list, to create training plans for the KBP model, to allow seamless integration of knowledge from a new system into clinical routine. Model performance was compared for KBP models trained with manually created and automatic MCO treatment plans. MATERIAL AND METHODS: Two RapidPlan models with the same 30 locally advanced non-small cell lung cancer patients included were created, one containing manually created clinical plans (RP_CLIN) and one containing fully automatic multi-criteria optimized plans (RP_MCO). For 15 validation patients, model performance was compared in terms of dose-volume parameters and normal tissue complication probabilities, and an oncologist performed a blind comparison of the clinical (CLIN), RP_CLIN, and RP_MCO plans. RESULTS: The heart and esophagus doses were lower for RP_MCO compared to RP_CLIN, resulting in an average reduction in the risk of 2-year mortality by 0.9 percentage points and the risk of acute esophageal toxicity by 1.6 percentage points with RP_MCO. The oncologist preferred the RP_MCO plan for 8 patients and the CLIN plan for 7 patients, while the RP_CLIN plan was not preferred for any patients. CONCLUSION: RP_MCO improved OAR sparing compared to RP_CLIN and was selected for implementation in the clinic. Training a KBP model with clinical plans may lead to suboptimal output plans, and making an extra effort to optimize the library plans in the KBP model creation phase can improve the plan quality for many future patients.

A two-step treatment planning strategy incorporating knowledge-based planning for head-and-neck radiotherapy.

PURPOSE: There has been much research interest in automated head-and-neck (HN) planning with the goal of reducing planning time and inter-planner variability while improving plan quality. However, clinical uses are still limited and institution-dependent due to the plan complexity. This work aims to investigate whether the use of a novel semi-automated two-step optimization method (TSP) can improve the quality and efficiency of planning while providing a simple framework that other institutions can follow. METHODS AND MATERIALS: Forty patients (two and three prescription isodose levels) were retrospectively studied. Plans were generated by TSP which incorporates a knowledge-based planning solution. Comparisons were performed for plan conformity and selected dose-volume indices between clinical plan (CP) and TSP. Blind reviews were carried out by 15 clinicians to determine preference between the CP and TSP, as well as clinical suitability. RESULTS: For majority of patients studied, TSP had similar or slightly better conformity for the high-dose PTV, and better conformity for the low-dose PTV and 45 Gy isodose lines compared to CP. The only statistically significant difference observed for the serial organs was a reduction of the spinal cord maximum dose with TSP. Except for left parotid gland (Dmean and V30 for both 2R× and 3R× groups) and oral cavity (Dmean for 3R× group), TSP had significant dose reductions for all parallel organs compared to CP. Blind reviewers either showed preference/no preference for 57.2%/21.7% (2R×) and 57.5%/27.8% (3R×) of TSP compared with CP. Excluding no preference votes, 60% of TSP were preferred. TSP was selected majority of the time when looking at the vote distribution for each patient individually. CONCLUSION: Our TSP allows plans to be created within 90-min time frame while offering improvements in plan quality and less inter-planner variability as compared to traditional planning techniques.

Clinical application of a template-guided automated planning routine.

PURPOSE: Determine the dosimetric quality and the planning time reduction when utilizing a template-based automated planning application. METHODS: A software application integrated through the treatment planning system application programing interface, QuickPlan, was developed to facilitate automated planning using configurable templates for contouring, knowledge-based planning structure matching, field design, and algorithm settings. Validations are performed at various levels of the planning procedure and assist in the evaluation of readiness of the CT image, structure set, and plan layout for automated planning. QuickPlan is evaluated dosimetrically against 22 hippocampal-avoidance whole brain radiotherapy patients. The required times to treatment plan generation are compared for the validations set as well as 10 prospective patients whose plans have been automated by QuickPlan. RESULTS: The generations of 22 automated treatment plans are compared against a manual replanning using an identical process, resulting in dosimetric differences of minor clinical significance. The target dose to 2% volume and homogeneity index result in significantly decreased values for automated plans, whereas other dose metric evaluations are nonsignificant. The time to generate the treatment plans is reduced for all automated plans with a median difference of 9' 50″ ± 4' 33″. CONCLUSIONS: Template-based automated planning allows for reduced treatment planning time with consistent optimization structure creation, treatment field creation, plan optimization, and dose calculation with similar dosimetric quality. This process has potential expansion to numerous disease sites.

Automated evaluation for rapid implementation of knowledge-based radiotherapy planning models.

PURPOSE: Knowledge-based planning (KBP) offers the ability to predict dose-volume metrics based on information extracted from previous plans, reducing plan variability and improving plan quality. As clinical integration of KBP is increasing there is a growing need for quantitative evaluation of KBP models. A .NET-based application, RapidCompare, was created for automated plan creation and analysis of Varian RapidPlan models. METHODS: RapidCompare was designed to read calculation parameters and a list of reference plans. The tool copies the reference plan field geometry and structure set, applies the RapidPlan model, optimizes the KBP plan, and generates data for quantitative evaluation of dose-volume metrics. A cohort of 85 patients, divided into training (50), testing (10), and validation (25) groups, was used to demonstrate the utility of RapidCompare. After training and tuning, the KBP model was paired with three different optimization templates to compare various planning strategies in the validation cohort. All templates used the same set of constraints for the planning target volume (PTV). For organs-at-risk, the optimization template provided constraints using the whole dose-volume histogram (DVH), fixed-dose/volume points, or generalized equivalent uniform dose (gEUD). The resulting plans from each optimization approach were compared using DVH metrics. RESULTS: RapidCompare allowed for the automated generation of 75 total plans for comparison with limited manual intervention. In comparing optimization techniques, the Dose/Volume and Lines optimization templates generated plans with similar DVH metrics, with a slight preference for the Lines technique with reductions in heart V30Gy and spinal cord max dose. The gEUD model produced high target heterogeneity. CONCLUSION: Automated evaluation allowed for the exploration of multiple optimization templates in a larger validation cohort than would have been feasible using a manual approach. A final KBP model using line optimization objectives produced the highest quality plans without human intervention.

The effect of integrating knowledge-based planning with multicriteria optimization in treatment planning for prostate SBRT.

Knowledge-based planning (KBP) and multicriteria optimization (MCO) are two powerful tools to assist treatment planners in achieving optimal target coverage and organ-at-risk (OAR) sparing. The purpose of this work is to investigate if integrating MCO with conventional KBP can further improve treatment plan quality for prostate cancer stereotactic body radiation therapy (SBRT). A two-phase study was designed to investigate the impact of MCO and KBP in prostate SBRT treatment planning. The first phase involved the creation of a KBP model based on thirty clinical SBRT plans, generated by manual optimization (KBP_M). A ten-patient validation cohort was used to compare manual, MCO, and KBP_M optimization techniques. The next phase involved replanning the original model cohort with additional tradeoff optimization via MCO to create a second model, KBP_MCO. Plans were then generated using linear integration (KBP_M+MCO), non-linear integration (KBP_MCO), and a combination of integration methods (KBP_MCO+MCO). All plans were analyzed for planning target volume (PTV) coverage, OAR constraints, and plan quality metrics. Comparisons were generated to evaluate plan and model quality. Phase 1 highlighted the necessity of KBP and MCO in treatment planning, as both optimization methods improved plan quality metrics (Conformity and Heterogeneity Indices) and reduced mean rectal dose by 2 Gy, as compared to manual planning. Integrating MCO with KBP did not further improve plan quality, as little significance was seen over KBP or MCO alone. Principal component score (PCS) fitting showed KBP_MCO improved bladder and rectum estimated and modeled dose correlation by 5% and 22%, respectively; however, model improvements did not significantly impact plan quality. KBP and MCO have shown to reduce OAR dose while maintaining desired PTV coverage in this study. Further integration of KBP and MCO did not show marked improvements in treatment plan quality while requiring increased time in model generation and optimization time.

Dosimetric potential of knowledge-based planning model trained with HyperArc plans for brain metastases.

OBJECTIVE: Dosimetric potential of knowledge-based RapidPlan planning model trained with HyperArc plans (Model-HA) for brain metastases has not been reported. We developed a Model-HA and compared its performance with that of clinical volumetric modulated arc therapy (VMAT) plans. METHODS: From 67 clinical stereotactic radiosurgery (SRS) HyperArc plans for brain metastases, 47 plans were used to build and train a Model-HA. The other 20 clinical HyperArc plans were recalculated in RapidPlan system with Model-HA. The model performance was validated with the 20 plans by comparing dosimetric parameters for normal brain tissue between clinical plans and model-generated plans. The 20 clinical conventional VMAT-based SRS or stereotactic radiotherapy plans (CL-VMAT) were reoptimized with Model-HA (RP) and HyperArc system (HA), respectively. The dosimetric parameters were compared among three plans (CL-VMAT vs. RP vs. HA) in terms of planning target volume (PTV), normal brain excluding PTVs (Brain - PTV), brainstem, chiasm, and both optic nerves. RESULTS: In model validation, the optimization performance of Model-HA was comparable to that of HyperArc system. In comparison to CL-VMAT, there were no significant differences among three plans with respect to PTV coverage (p > 0.17) and maximum dose for brainstem, chiasm, and optic nerves (p > 0.40). RP provided significantly lower V20 Gy , V12 Gy , and V4 Gy for Brain - PTV than CL-VMAT (p < 0.01). CONCLUSION: The Model-HA has the potential to significantly reduce the normal brain dose of the original VMAT plans for brain metastases.

A comparative study of deep learning-based knowledge-based planning methods for 3D dose distribution prediction of head and neck.

PURPOSE: In this paper, we compare four novel knowledge-based planning (KBP) algorithms using deep learning to predict three-dimensional (3D) dose distributions of head and neck plans using the same patients' dataset and quantitative assessment metrics. METHODS: A dataset of 340 oropharyngeal cancer patients treated with intensity-modulated radiation therapy was used in this study, which represents the AAPM OpenKBP - 2020 Grand Challenge dataset. Four 3D convolutional neural network architectures were built. The models were trained on 64% of the data set and validated on 16% for voxel-wise dose predictions: U-Net, attention U-Net, residual U-Net (Res U-Net), and attention Res U-Net. The trained models were then evaluated for their performance on a test data set (20% of the data) by comparing the predicted dose distributions against the ground-truth using dose statistics and dose-volume indices. RESULTS: The four KBP dose prediction models exhibited promising performance with an averaged mean absolute dose error within the body contour <3 Gy on 68 plans in the test set. The average difference in predicting the D99 index for all targets was 0.92 Gy (p = 0.51) for attention Res U-Net, 0.94 Gy (p = 0.40) for Res U-Net, 2.94 Gy (p = 0.09) for attention U-Net, and 3.51 Gy (p = 0.08) for U-Net. For the OARs, the values for the D m a x ${D_{max}}$ and D m e a n ${D_{mean}}$ indices were 2.72 Gy (p < 0.01) for attention Res U-Net, 2.94 Gy (p < 0.01) for Res U-Net, 1.10 Gy (p < 0.01) for attention U-Net, 0.84 Gy (p < 0.29) for U-Net. CONCLUSION: All models demonstrated almost comparable performance for voxel-wise dose prediction. KBP models that employ 3D U-Net architecture as a base could be deployed for clinical use to improve cancer patient treatment by creating plans with consistent quality and making the radiotherapy workflow more efficient.

A practical method to improve the performance of knowledge-based VMAT planning for endometrial and cervical cancer.

PURPOSE: The aim of this work was to demonstrate a practical and effective method to improve the performance of RapidPlan (RP) model. METHODS: 203 consecutive clinical VMAT plans (P0) for cervical and endometrial cancer were used to train an RP model (M0). The plans were then reoptimized by M0 to generate 203 new plans (P1). Compared with P0, 150 plans with a lower mean dose (MD) of bladder, rectum and PBM were selected from P1 to configure a new RP model (M1). A final RP model (M2) was trained using plans in M1 and the remaining 53 plans from P1 (excluding OARs with worse MD) and the corresponding plans from P0 (only including OARs with better MD). The models were validated on the mentioned 53 plans (closed-loop set) and 46 patient cohorts outside the training library (open-loop set). p < 0.05 was considered statistically significant. RESULTS: For closed-loop validation, the difference of D2%, D98% and CI95% between groups was of no statistical significance, the homogeneity index (HI) was lower in the groups of RP models (p < 0.05). The MD of all OARs decreased monotonically in the sequence of the clinical group, group M0, M1 and M2, except the MD of bowel in M1 and MD of LFH in M2. Similarly, for open-loop validation, there was no significant difference in D2%, D98% and HI between groups, but CI95% was larger in the clinical group (p < 0.05). The MD of all OARs decreased monotonically in the sequence of the clinical group, group M0, M1 and M2, with the exception of bowel in M1. CONCLUSION: The practical method of incorporating plan data of better-sparing OARs from both the clinical VMAT plans and the re-optimized plans could further improve the performance of the RP model.

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.

Knowledge-based planning for the radiation therapy treatment plan quality assurance for patients with head and neck cancer.

This study aimed to investigate the feasibility of using a knowledge-based planning technique to detect poor quality VMAT plans for patients with head and neck cancer. We created two dose-volume histogram (DVH) prediction models using a commercial knowledge-based planning system (RapidPlan, Varian Medical Systems, Palo Alto, CA) from plans generated by manual planning (MP) and automated planning (AP) approaches. DVHs were predicted for evaluation cohort 1 (EC1) of 25 patients and compared with achieved DVHs of MP and AP plans to evaluate prediction accuracy. Additionally, we predicted DVHs for evaluation cohort 2 (EC2) of 25 patients for which we intentionally generated plans with suboptimal normal tissue sparing while satisfying dose-volume limits of standard practice. Three radiation oncologists reviewed these plans without seeing the DVH predictions. We found that predicted DVH ranges (upper-lower predictions) were consistently wider for the MP model than for the AP model for all normal structures. The average ranges of mean dose predictions among all structures was 9.7 Gy (MP model) and 3.4 Gy (AP model) for EC1 patients. RapidPlan models identified 7 MP plans as outliers according to mean dose or D1% for at least one structure, while none of AP plans were flagged. For EC2 patients, 22 suboptimal plans were identified by prediction. While re-generated AP plans validated that these suboptimal plans could be improved, 40 out of 45 structures with predicted poor sparing were also identified by oncologist reviews as requiring additional planning to improve sparing in the clinical setting. Our study shows that knowledge-based DVH prediction models can be sufficiently accurate for plan quality assurance purposes. A prediction model built by a small cohort automatically-generated plans was effective in detecting suboptimal plans. Such tools have potential to assist the plan quality assurance workflow for individual patients in the clinic.

Combining dense elements with attention mechanisms for 3D radiotherapy dose prediction on head and neck cancers.

PURPOSE: External radiation therapy planning is a highly complex and tedious process as it involves treating large target volumes, prescribing several levels of doses, as well as avoiding irradiating critical structures such as organs at risk close to the tumor target. This requires highly trained dosimetrists and physicists to generate a personalized plan and adapt it as treatment evolves, thus affecting the overall tumor control and patient outcomes. Our aim is to achieve accurate dose predictions for head and neck (H&N) cancer patients on a challenging in-house dataset that reflects realistic variability and to further compare and validate the method on a public dataset. METHODS: We propose a three-dimensional (3D) deep neural network that combines a hierarchically dense architecture with an attention U-net (HDA U-net). We investigate a domain knowledge objective, incorporating a weighted mean squared error (MSE) with a dose-volume histogram (DVH) loss function. The proposed HDA U-net using the MSE-DVH loss function is compared with two state-of-the-art U-net variants on two radiotherapy datasets of H&N cases. These include reference dose plans, computed tomography (CT) information, organs at risk (OARs), and planning target volume (PTV) delineations. All models were evaluated using coverage, homogeneity, and conformity metrics as well as mean dose error and DVH curves. RESULTS: Overall, the proposed architecture outperformed the comparative state-of-the-art methods, reaching 0.95 (0.98) on D95 coverage, 1.06 (1.07) on the maximum dose value, 0.10 (0.08) on homogeneity, 0.53 (0.79) on conformity index, and attaining the lowest mean dose error on PTVs of 1.7% (1.4%) for the in-house (public) dataset. The improvements are statistically significant ( p < 0.05 $p<0.05$ ) for the homogeneity and maximum dose value compared with the closest baseline. All models offer a near real-time prediction, measured between 0.43 and 0.88 s per volume. CONCLUSION: The proposed method achieved similar performance on both realistic in-house data and public data compared to the attention U-net with a DVH loss, and outperformed other methods such as HD U-net and HDA U-net with standard MSE losses. The use of the DVH objective for training showed consistent improvements to the baselines on most metrics, supporting its added benefit in H&N cancer cases. The quick prediction time of the proposed method allows for real-time applications, providing physicians a method to generate an objective end goal for the dosimetrist to use as reference for planning. This could considerably reduce the number of iterations between the two expert physicians thus reducing the overall treatment planning time.

Attention-aware 3D U-Net convolutional neural network for knowledge-based planning 3D dose distribution prediction of head-and-neck cancer.

PURPOSE: Deep learning-based knowledge-based planning (KBP) methods have been introduced for radiotherapy dose distribution prediction to reduce the planning time and maintain consistent high-quality plans. This paper presents a novel KBP model using an attention-gating mechanism and a three-dimensional (3D) U-Net for intensity-modulated radiation therapy (IMRT) 3D dose distribution prediction in head-and-neck cancer. METHODS: A total of 340 head-and-neck cancer plans, representing the OpenKBP-2020 AAPM Grand Challenge data set, were used in this study. All patients were treated with the IMRT technique and a dose prescription of 70 Gy. The data set was randomly divided into 64%/16%/20% as training/validation/testing cohorts. An attention-gated 3D U-Net architecture model was developed to predict full 3D dose distribution. The developed model was trained using the mean-squared error loss function, Adam optimization algorithm, a learning rate of 0.001, 120 epochs, and batch size of 4. In addition, a baseline U-Net model was also similarly trained for comparison. The model performance was evaluated on the testing data set by comparing the generated dose distributions against the ground-truth dose distributions using dose statistics and clinical dosimetric indices. Its performance was also compared to the baseline model and the reported results of other deep learning-based dose prediction models. RESULTS: The proposed attention-gated 3D U-Net model showed high capability in accurately predicting 3D dose distributions that closely replicated the ground-truth dose distributions of 68 plans in the test set. The average value of the mean absolute dose error was 2.972 ± 1.220 Gy (vs. 2.920 ± 1.476 Gy for a baseline U-Net) in the brainstem, 4.243 ± 1.791 Gy (vs. 4.530 ± 2.295 Gy for a baseline U-Net) in the left parotid, 4.622 ± 1.975 Gy (vs. 4.223 ± 1.816 Gy for a baseline U-Net) in the right parotid, 3.346 ± 1.198 Gy (vs. 2.958 ± 0.888 Gy for a baseline U-Net) in the spinal cord, 6.582 ± 3.748 Gy (vs. 5.114 ± 2.098 Gy for a baseline U-Net) in the esophagus, 4.756 ± 1.560 Gy (vs. 4.992 ± 2.030 Gy for a baseline U-Net) in the mandible, 4.501 ± 1.784 Gy (vs. 4.925 ± 2.347 Gy for a baseline U-Net) in the larynx, 2.494 ± 0.953 Gy (vs. 2.648 ± 1.247 Gy for a baseline U-Net) in the PTV_70, and 2.432 ± 2.272 Gy (vs. 2.811 ± 2.896 Gy for a baseline U-Net) in the body contour. The average difference in predicting the D99 value for the targets (PTV_70, PTV_63, and PTV_56) was 2.50 ± 1.77 Gy. For the organs at risk, the average difference in predicting the D m a x ${D_{max}}$ (brainstem, spinal cord, and mandible) and D m e a n ${D_{mean}}$ (left parotid, right parotid, esophagus, and larynx) values was 1.43 ± 1.01 and 2.44 ± 1.73 Gy, respectively. The average value of the homogeneity index was 7.99 ± 1.45 for the predicted plans versus 5.74 ± 2.95 for the ground-truth plans, whereas the average value of the conformity index was 0.63 ± 0.17 for the predicted plans versus 0.89 ± 0.19 for the ground-truth plans. The proposed model needs less than 5 s to predict a full 3D dose distribution of 64 × 64 × 64 voxels for a new patient that is sufficient for real-time applications. CONCLUSIONS: The attention-gated 3D U-Net model demonstrated a capability in predicting accurate 3D dose distributions for head-and-neck IMRT plans with consistent quality. The prediction performance of the proposed model was overall superior to a baseline standard U-Net model, and it was also competitive to the performance of the best state-of-the-art dose prediction method reported in the literature. The proposed model could be used to obtain dose distributions for decision-making before planning, quality assurance of planning, and guiding-automated planning for improved plan consistency, quality, and planning efficiency.

Knowledge-based intensity-modulated proton planning for gastroesophageal carcinoma.

PURPOSE: To investigate the performance of a narrow-scope knowledge-based RapidPlan (RP) model, for optimisation of intensity-modulated proton therapy (IMPT) plans applied to patients with locally advanced carcinoma in the gastroesophageal junction. METHODS: A cohort of 60 patients was retrospectively selected; 45 were used to 'train' a dose-volume histogram predictive model; the remaining 15 provided independent validation. The performance of the RP model was benchmarked against manual optimisation. Quantitative assessment was based on several dose-volume metrics. RESULTS: Manual and RP-optimised IMPT plans resulted dosimetrically similar, and the planning dose-volume objectives were met for all structures. Concerning the validation set, the comparison of the manual vs RP-based plans, respectively, showed for the target (PTV): the homogeneity index was 6.3 ± 2.2 vs 5.9 ± 1.2, and V98% was 89.3 ± 2.9 vs 91.4 ± 2.2% (this was 97.2 ± 1.9 vs 98.8 ± 1.1 for the CTV). Regarding the organs at risk, no significant differences were reported for the combined lungs, the whole heart, the left anterior descending artery, the kidneys, the spleen and the spinal canal. The D0.1 cm3 for the left ventricle resulted in 40.3 ± 3.4 vs 39.7 ± 4.3 Gy(RBE). The mean dose to the liver was 3.4 ± 1.3 vs 3.6 ± 1.5 Gy(RBE). CONCLUSION: A narrow-scope knowledge-based RP model was trained and validated for IMPT delivery in locally advanced cancer of the gastroesophageal junction. The results demonstrate that RP can create models for effective IMPT. Furthermore, the equivalence between manual interactive and unattended RP-based optimisation could be displayed. The data also showed a high correlation between predicted and achieved doses in support of the valuable predictive power of the RP method.

Optimal beam angle selection and knowledge-based planning significantly reduces radiotherapy dose to organs at risk for lung cancer patients.

BACKGROUND: Lung cancer patients struggle with high toxicity rates. This study investigates if IMRT plans with individually set beam angles or uni-lateral VMAT plans results in dose reduction to OARs. We investigate if introduction of a RapidPlan model leads to reduced dose to OARs. Finally, the model is validated prospectively. MATERIAL AND METHODS: Seventy-four consecutive lung cancer patients treated with IMRT were included. For all patients, new IMRT plans were made by an experienced dose planner re-tuning beam angles aiming for minimized dose to the lungs and heart. Additionally, VMAT plans were made. The IMRT plans were selected as input for a RapidPlan model, which was used to generate 74 new IMRT plans. The new IMRT plans were used as input for a second RapidPlan model. This model was clinically implemented and used for generation of clinical treatment plans. Dosimetric parameters were compared using a Wilcoxon signed rank test or a 1-sided student's t-test. p < .05 was considered significant. RESULTS: IMRT plans significantly reduced mean doses to lungs (MLD) and heart (MHD) by 1.6 Gy and 1.7 Gy in mean compared to VMAT plans. MLD was significantly (p < .001) reduced from 10.8 Gy to 9.4 Gy by using the second RapidPlan model. MHD was significantly (p < .001) reduced from 4.9 Gy to 3.9 Gy. The model was validated in prospectively collected treatment plans showing significantly lower MLD after the implementation of the second RapidPlan model. CONCLUSION: Introduction of RapidPlan and beam angles selected based on the target and OARs position reduces dose to OARs.

An updating approach for knowledge-based planning models to improve plan quality and variability in volumetric-modulated arc therapy for prostate cancer.

PURPOSE: The purpose of this study was to compare the dose-volume parameters and regression scatter plots of the iteratively improved RapidPlan (RP) models, specific knowledge-based planning (KBP) models, in volumetric-modulated arc therapy (VMAT) for prostate cancer over three periods. METHODS: A RP1 model was created from 47 clinical intensity-modulated radiation therapy (IMRT)/VMAT plans. A RP2 model was created to exceed dosimetric goals which set as the mean values +1SD of the dose-volume parameters of RP1 (50 consecutive new clinical VMAT plans). A RP3 model was created with more strict dose constraints for organs at risks (OARs) than RP1 and RP2 models (50 consecutive anew clinical VMAT plans). Each RP model was validated against 30 validation plans (RP1, RP2, and RP3) that were not used for model configuration, and the dose-volume parameters were compared. The Cook's distances of regression scatterplots of each model were also evaluated. RESULTS: Significant differences (p < 0.05) between RP1 and RP2 were found in Dmean (101.5% vs. 101.9%), homogeneity index (3.90 vs. 4.44), 95% isodose conformity index (1.22 vs. 1.20) for the target, V40Gy (47.3% vs. 45.7%), V60Gy (27.9% vs. 27.1%), V70Gy (16.4% vs. 15.2%), and V78Gy (0.4% vs. 0.2%) for the rectal wall, and V40Gy (43.8% vs. 41.8%) and V70Gy (21.3% vs. 20.5%) for the bladder wall, whereas only V70Gy (15.2% vs. 15.8%) of the rectal wall differed significantly between RP2 and RP3. The proportions of cases with a Cook's distance of <1.0 (RP1, RP2, and RP3 models) were 55%, 78%, and 84% for the rectal wall, and 77%, 68%, and 76% for the bladder wall, respectively. CONCLUSIONS: The iteratively improved RP models, reflecting the clear dosimetric goals based on the RP feedback (dose-volume parameters) and more strict dose constraints for the OARs, generated superior dose-volume parameters and the regression scatterplots in the model converged. This approach could be used to standardize the inverse planning strategies.

Knowledge-based planning in robotic intracranial stereotactic radiosurgery treatments.

PURPOSE: To develop a knowledge-based planning (KBP) model that predicts dosimetric indices and facilitates planning in CyberKnife intracranial stereotactic radiosurgery/radiotherapy (SRS/SRT). METHODS: Forty CyberKnife SRS/SRT plans were retrospectively used to build a linear KBP model which correlated the equivalent radius of the PTV (req_PTV ) and the equivalent radius of volume that receives a set of prescription dose (req_Vi , where Vi  = V10% , V20% … V120% ). To evaluate the model's predictability, a fourfold cross-validation was performed for dosimetric indices such as gradient measure (GM) and brain V50% . The accuracy of the prediction was quantified by the mean and the standard deviation of the difference between planned and predicted values, (i.e., ΔGM = GMpred - GMclin and fractional ΔV50%  = (V50%pred - V50%clin )/V50%clin ) and a coefficient of determination, R2 . Then, the KBP model was incorporated into the planning for another 22 clinical cases. The training plans and the KBP test plans were compared in terms of the new conformity index (nCI) as well as the planning efficiency. RESULTS: Our KBP model showed desirable predictability. For the 40 training plans, the average prediction error from cross-validation was only 0.36 ± 0.06 mm for ΔGM, and 0.12 ± 0.08 for ΔV50% . The R2 for the linear fit between req_PTV and req_vi was 0.985 ± 0.019 for isodose volumes ranging from V10% to V120% ; particularly, R2  = 0.995 for V50% and R2  = 0.997 for V100% . Compared to the training plans, our KBP test plan nCI was improved from 1.31 ± 0.15 to 1.15 ± 0.08 (P < 0.0001). The efficient automatic generation of the optimization constraints by using our model requested no or little planner's intervention. CONCLUSION: We demonstrated a linear KBP based on PTV volumes that accurately predicts CyberKnife SRS/SRT planning dosimetric indices and greatly helps achieve superior plan quality and planning efficiency.