Search

Purpose: Treatment planning for volumetric modulated arc therapy (VMAT) typically involves the use of multiple arcs to achieve sufficient intensity modulation. Alternatively, we can perform segment boosting to achieve similar intensity modulation while also reducing the number of control points used. Here, we propose the MetaPlanner Boosted VMAT (MPBV) approach, which generates boosted VMAT plans through a fully automated framework. Methods: The proposed MPBV approach is an open‐source framework that consists of three main stages: meta‐optimization of treatment plan hyperparameters, fast beam angle optimization on a coarse dose grid to select desirable segments for boosting, and final plan generation (i.e., constructing the boosted VMAT arc and performing optimization). Results: Performance for the MPBV approach is evaluated on 21 prostate cases and 6 head and neck cases using clinically relevant plan quality metrics (i.e., target coverage, dose conformity, dose homogeneity, and OAR sparing). As compared to two baseline methods with multiple arcs, MPBV maintains or improves dosimetric performance for the evaluated metrics while substantially reducing average estimated delivery times (from 2.6 to 2.1 min). Conclusion: Our proposed MPBV approach provides an automated framework for producing high‐quality VMAT plans that uses fewer control points and reduces delivery time as compared to traditional approaches with multiple arcs. MPBV applies automated treatment planning to segmentally boosted VMAT to address the beam utilization inefficiencies of traditional VMAT approaches that use multiple full arcs. [ABSTRACT FROM AUTHOR]

Background: Knowledge‐based planning (KBP) is increasingly implemented clinically because of its demonstrated ability to improve treatment planning efficiency and reduce plan quality variations. However, cases with large dose‐volume histogram (DVH) prediction uncertainties may still need manual adjustments by the planner to achieve high plan quality. Purpose: The purpose of this study is to develop a data‐driven method to detect patients with high prediction uncertainties so that intentional effort is directed to these patients. Methods: We apply an anomaly detection method known as the local outlier factor (LOF) to a dataset consisting of the training set and each of the prospective patients considered, to evaluate their likelihood of being an anomaly when compared with the training cases. Features used in the LOF analysis include anatomical features and the model‐generated DVH principal component scores. To test the efficacy of the proposed model, 142 prostate patients were retrieved from the clinical database and split into a training dataset of 100 patients and a test dataset of 42 patients. The outlier identification performance was quantified by the difference between the DVH prediction root‐mean‐squared errors (RMSE) of the identified outlier cases and that of the remaining inlier cases. Results: With a predefined LOF threshold of 1.4, the inlier cases achieved average RMSEs of 5.0 and 6.7 for bladder and rectum, while the outlier cases have substantially higher RMSEs of 6.7 and 13.0 in comparison. Conclusions: We propose a method that can determine the prospective patient's outlier status. This method can be integrated into existing automated treatment planning workflows to reduce the risk of generating suboptimal treatment plans while providing an upfront alert to the treatment planner. [ABSTRACT FROM AUTHOR]

Purpose: To advance fair and consistent comparisons of dose prediction methods for knowledge‐based planning (KBP) in radiation therapy research. Methods: We hosted OpenKBP, a 2020 AAPM Grand Challenge, and challenged participants to develop the best method for predicting the dose of contoured computed tomography (CT) images. The models were evaluated according to two separate scores: (a) dose score, which evaluates the full three‐dimensional (3D) dose distributions, and (b) dose‐volume histogram (DVH) score, which evaluates a set DVH metrics. We used these scores to quantify the quality of the models based on their out‐of‐sample predictions. To develop and test their models, participants were given the data of 340 patients who were treated for head‐and‐neck cancer with radiation therapy. The data were partitioned into training (n=200), validation (n=40), and testing (n=100) datasets. All participants performed training and validation with the corresponding datasets during the first (validation) phase of the Challenge. In the second (testing) phase, the participants used their model on the testing data to quantify the out‐of‐sample performance, which was hidden from participants and used to determine the final competition ranking. Participants also responded to a survey to summarize their models. Results: The Challenge attracted 195 participants from 28 countries, and 73 of those participants formed 44 teams in the validation phase, which received a total of 1750 submissions. The testing phase garnered submissions from 28 of those teams, which represents 28 unique prediction methods. On average, over the course of the validation phase, participants improved the dose and DVH scores of their models by a factor of 2.7 and 5.7, respectively. In the testing phase one model achieved the best dose score (2.429) and DVH score (1.478), which were both significantly better than the dose score (2.564) and the DVH score (1.529) that was achieved by the runner‐up models. Lastly, many of the top performing teams reported that they used generalizable techniques (e.g., ensembles) to achieve higher performance than their competition. Conclusion: OpenKBP is the first competition for knowledge‐based planning research. The Challenge helped launch the first platform that enables researchers to compare KBP prediction methods fairly and consistently using a large open‐source dataset and standardized metrics. OpenKBP has also democratized KBP research by making it accessible to everyone, which should help accelerate the progress of KBP research. The OpenKBP datasets are available publicly to help benchmark future KBP research. [ABSTRACT FROM AUTHOR]

Purpose: To develop a knowledge‐based automated planning pipeline that generates treatment plans without feature engineering, using deep neural network architectures for predicting three‐dimensional (3D) dose. Methods: Our knowledge‐based automated planning (KBAP) pipeline consisted of a knowledge‐based planning (KBP) method that predicts dose for a contoured computed tomography (CT) image followed by two optimization models that learn objective function weights and generate fluence‐based plans, respectively. We developed a novel generative adversarial network (GAN)‐based KBP approach, a 3D GAN model, which predicts dose for the full 3D CT image at once and accounts for correlations between adjacent CT slices. Baseline comparisons were made against two state‐of‐the‐art deep learning–based KBP methods from the literature. We also developed an additional benchmark, a two‐dimensional (2D) GAN model which predicts dose to each axial slice independently. For all models, we investigated the impact of multiplicatively scaling the predictions before optimization, such that the predicted dose distributions achieved all target clinical criteria. Each KBP model was trained on 130 previously delivered oropharyngeal treatment plans. Performance was tested on 87 out‐of‐sample previously delivered treatment plans. All KBAP plans were evaluated using clinical planning criteria and compared to their corresponding clinical plans. KBP prediction quality was assessed using dose‐volume histogram (DVH) differences from the corresponding clinical plans. Results: The best performing KBAP plans were generated using predictions from the 3D GAN model that were multiplicatively scaled. These plans satisfied 77% of all clinical criteria, compared to the clinical plans, which satisfied 67% of all criteria. In general, multiplicatively scaling predictions prior to optimization increased the fraction of clinical criteria satisfaction by 11% relative to the plans generated with nonscaled predictions. Additionally, these KBAP plans satisfied the same criteria as the clinical plans 84% and 8% more frequently as compared to the two benchmark methods, respectively. Conclusions: We developed the first knowledge‐based automated planning framework using a 3D generative adversarial network for prediction. Our results, based on 217 oropharyngeal cancer treatment plans, demonstrated superior performance in satisfying clinical criteria and generated more realistic plans as compared to the previous state‐of‐the‐art approaches. [ABSTRACT FROM AUTHOR]

Purpose: To develop and implement a fully automated approach to intensity modulated radiation therapy (IMRT) treatment planning. Method: The optimization algorithm is developed based on a hierarchical constrained optimization technique and is referred internally at our institution as expedited constrained hierarchical optimization (ECHO). Beamlet contributions to regions‐of‐interest are precomputed and captured in the influence matrix. Planning goals are of two classes: hard constraints that are strictly enforced from the first step (e.g., maximum dose to spinal cord), and desirable goals that are sequentially introduced in three constrained optimization problems (better planning target volume (PTV) coverage, lower organ at risk (OAR) doses, and smoother fluence map). After solving the optimization problems using external commercial optimization engines, the optimal fluence map is imported into an FDA‐approved treatment planning system (TPS) for leaf sequencing and accurate full dose calculation. The dose‐discrepancy between the optimization and TPS dose calculation is then calculated and incorporated into optimization by a novel dose correction loop technique using Lagrange multipliers. The correction loop incorporates the leaf sequencing and scattering effects into optimization to improve the plan quality and reduce the calculation time. The resultant optimal fluence map is again imported into TPS for leaf sequencing and final dose calculation for plan evaluation and delivery. The workflow is automated using application program interface (API) scripting, requiring user interaction solely to prepare the contours and beam arrangement prior to launching the ECHO plug‐in from the TPS. For each site, parameters and objective functions are chosen to represent clinical priorities. The first site chosen for clinical implementation was metastatic paraspinal lesions treated with stereotactic body radiotherapy (SBRT). As a first step, 75 ECHO paraspinal plans were generated retrospectively and compared with clinically treated plans generated by planners using VMAT (volumetric modulated arc therapy) with 4 to 6 partial arcs. Subsequently, clinical deployment began in April, 2017. Results: In retrospective study, ECHO plans were found to be dosimetrically superior with respect to tumor coverage, plan conformity, and OAR sparing. For example, the average PTV D95%, cord and esophagus max doses, and Paddick Conformity Index were improved, respectively, by 1%, 6%, 14%, and 15%, at a negligible 3% cost of the average skin D10cc dose. Conclusion: Hierarchical constrained optimization is a powerful and flexible tool for automated IMRT treatment planning. The dosimetric correction step accurately accounts for detailed dosimetric multileaf collimator and scattering effects. The system produces high‐quality, Pareto optimal plans and avoids the time‐consuming trial‐and‐error planning process. [ABSTRACT FROM AUTHOR]

Purpose: The planning of an intensity modulated radiation therapy treatment requires the optimization of the fluence intensities. The fluence map optimization (FMO) is many times based on a nonlinear continuous programming problem, being necessary for the planner to define a priori weights and/or lower bounds that are iteratively changed within a trial-and-error procedure until an acceptable plan is reached. In this work, the authors describe an alternative approach for FMO that releases the human planner from trial-and-error procedures, contributing for the automation of the planning process. Methods: The FMO is represented by a voxel-based convex penalty continuous nonlinear model. This model makes use of both weights and lower/upper bounds to guide the optimization process toward interesting solutions that are able to satisfy all the constraints defined for the treatment. All the model's parameters are iteratively changed by resorting to a fuzzy inference system. This system analyzes how far the current solution is from a desirable solution, changing in a completely automated way both weights and lower/upper bounds. The fuzzy inference system is based on fuzzy reasoning that enables the use of common-sense rules within an iterative optimization process. The method is built in two stages: in a first stage, an admissible solution is calculated, trying to guarantee that all the treatment planning constraints are being satisfied. In this first stage, the algorithm tries to improve as much as possible the irradiation of the planning target volumes. In a second stage, the algorithm tries to improve organ sparing, without jeopardizing tumor coverage. Results: The proposed methodology was applied to ten head-and-neck cancer cases already treated in the Portuguese Oncology Institute of Coimbra (IPOCFG) and signalized as complex cases. IMRT treatment was considered, with 7, 9, and 11 equidistant beam angles. It was possible to obtain admissible solutions for all the patients considered and with no human planner intervention. The results obtained were compared with the optimized solution using a similar optimization model but with human planner intervention. For the vast majority of cases, it was possible to improve organ sparing and at the same time to assure better tumor coverage. Conclusions: Embedding a fuzzy inference system into FMO allows human planner reasoning to be used in the guidance of the optimization process toward interesting regions in a truly automated way. The proposed methodology is capable of calculating high quality plans within reasonable computational times and can be an important contribution toward fully automated radiation therapy treatment planning. [ABSTRACT FROM AUTHOR]

Purpose: Prior work by the authors and other groups has studied the creation of automated intensity modulated radiotherapy (IMRT) plans of equivalent quality to those in a patient database of manually created clinical plans; those database plans provided guidance on the achievable sparing to organs-at-risk (OARs). However, in certain sites, such as head-and-neck, the clinical plans may not be sufficiently optimized because of anatomical complexity and clinical time constraints. This could lead to automated plans that suboptimally exploit OAR sparing. This work investigates a novel dose warping and scaling scheme that attempts to reduce effects of suboptimal sparing in clinical database plans, thus improving the quality of semiautomated head-and-neck cancer (HNC) plans. Methods: Knowledge-based radiotherapy (KBRT) plans for each of ten "query" patients were semiautomatically generated by identifying the most similar "match" patient in a database of 103 clinical manually created patient plans. The match patient's plans were adapted to the query case by: (1) deforming the match beam fluences to suit the query target volume and (2) warping the match primary/boost dose distribution to suit the query geometry and using the warped distribution to generate query primary/boost optimization dose-volume constraints. Item (2) included a distance scaling factor to improve query OAR dose sparing with respect to the possibly suboptimal clinical match plan. To further compensate for a component plan of the match case (primary/boost) not optimally sparing OARs, the query dose volume constraints were reduced using a dose scaling factor to be the minimum from either (a) the warped component plan (primary or boost) dose distribution or (b) the warped total plan dose distribution (primary + boost) scaled in proportion to the ratio of component prescription dose to total prescription dose. The dose-volume constraints were used to plan the query case with no human intervention to adjust constraints during plan optimization. Results: KBRT and original clinical plans were dosimetrically equivalent for parotid glands (mean/median doses), spinal cord, and brainstem (maximum doses). KBRT plans significantly reduced larynx median doses (21.5±6.6 Gy to 17.9±3.9 Gy), and oral cavity mean (32.3±6.2 Gy to 28.9±5.4 Gy) and median (28.7±5.7 Gy to 23.2±5.3 Gy) doses. Doses to ipsilateral parotid gland, larynx, oral cavity, and brainstem were lower or equivalent in the KBRT plans for the majority of cases. By contrast, KBRT plans generated without the dose warping and dose scaling steps were not significantly different from the clinical plans. Conclusions: Fast, semiautomatically generated HNC IMRT plans adapted from existing plans in a clinical database can be of equivalent or better quality than manually created plans. The reductions in OAR doses in the semiautomated plans, compared to the clinical plans, indicate that the proposed dose warping and scaling method shows promise in mitigating the impact of suboptimal clinical plans. [ABSTRACT FROM AUTHOR]