Sparsity constrained split feasibility for dose-volume constraints in inverse planning of intensity-modulated photon or proton therapy

Superiorization of projection algorithms for linearly constrained inverse radiotherapy treatment planning

Objective

We apply the superiorization methodology to the constrained intensity-modulated radiation therapy (IMRT) treatment planning problem. Superiorization combines a feasibility-seeking projection algorithm with objective function reduction: The underlying projection algorithm is perturbed with gradient descent steps to steer the algorithm towards a solution with a lower objective function value compared to one obtained solely through feasibility-seeking.

Approach

Within the open-source inverse planning toolkit matRad, we implement a prototypical algorithmic framework for superiorization using the well-established Agmon, Motzkin, and Schoenberg (AMS) feasibility-seeking projection algorithm and common nonlinear dose optimization objective functions. Based on this prototype, we apply superiorization to intensity-modulated radiation therapy treatment planning and compare it with (i) bare feasibility-seeking (i.e., without any objective function) and (ii) nonlinear constrained optimization using first-order derivatives. For these comparisons, we use the TG119 water phantom, the head-and-neck and the prostate patient of the CORT dataset.

Main results

Bare feasibility-seeking with AMS confirms previous studies, showing it can find solutions that are nearly equivalent to those found by the established piece-wise least-squares optimization approach. The superiorization prototype solved the linearly constrained planning problem with similar dosimetric performance to that of a general-purpose nonlinear constrained optimizer while showing smooth convergence in both constraint proximity and objective function reduction.

Significance

Superiorization is a useful alternative to constrained optimization in radiotherapy inverse treatment planning. Future extensions with other approaches to feasibility-seeking, e.g., with dose-volume constraints and more sophisticated perturbations, may unlock its full potential for high performant inverse treatment planning.

Dynamic string-averaging CQ-methods for the split feasibility problem with percentage violation constraints arising in radiation therapy treatment planning

We study a feasibility-seeking problem with percentage violation constraints (PVCs). These are additional constraints that are appended to an existing family of constraints, which single out certain subsets of the existing constraints and declare that up to a specified fraction of the number of constraints in each subset is allowed to be violated by up to a specified percentage of the existing bounds. Our motivation to investigate problems with PVCs comes from the field of radiation therapy treatment planning (RTTP) wherein the fully discretized inverse planning problem is formulated as a split feasibility problem and the PVCs give rise to nonconvex constraints. Following the CQ algorithm of Byrne (2002, Inverse Problems, Vol. 18, pp. 441-53), we develop a string-averaging CQ-method that uses only projections onto the individual sets that are half-spaces represented by linear inequalities. The question of extending our theoretical results to the nonconvex sets case is still open. We describe how our results apply to RTTP and provide a numerical example.

A dose volume constraint (DVC) projection-based algorithm for IMPT inverse planning optimization

PURPOSE

Provide a projection-based algorithm to solve the class of optimization problems encountered in intensity modulated proton therapy (IMPT). The algorithm can handle percentage dose-volume constraints that are usually found in such problems.

METHODS

To seek a feasible solution, the automatic relaxation method was used to project the spot weight vector onto the interval defined by lower and upper bound target dose constraints. The obtained solution was optimized separately based on the objective of each OAR in addition to maximizing the minimum target dose using the bisection search method using a stopping criterion of 10 cGy. The combined weight was used in the CQ algorithm to solve the split feasibility problem but with a special projection technique due to the non-convexity of dose volume constraints. The algorithm was applied to four clinical IMPT cases (meningioma, prostate, tongue, and oropharynx) and compared to the corresponding treatment plans optimized in Eclipse.

RESULTS

The treatment plans obtained, for the four cases, using the BCQ-ARM algorithm have dosimetric endpoints that are similar to their counterparts generated from Eclipse. The algorithm worked equally well with all cases, including the complex head and neck ones. The stopping criterion of 10 cGy results in making the generated plans slightly less optimal (ε-optimal) rather than optimal, but with the advantage of the possibility of generating a database of plans.

CONCLUSIONS

The application of the BCQ-ARM algorithm to different cases of IMPT plans with dose volume constraints was demonstrated. The algorithm is successful in generating plans that are dosimetrically equivalent to their corresponding Eclipse plans. Thus, it is suitable to generate optimized treatment plans in a clinically reasonable time frame. This article is protected by copyright. All rights reserved.

Intelligent IoT-based large-scale inverse planning system considering postmodulation factors

The model and algorithm of intensity-modulated radiotherapy (IMRT) are updated increasingly quickly, but the hardware upgrade of primary hospitals often lags behind. The new generation of intelligent precise radiotherapy platforms provides users with intelligent medical consortium services using big data, artificial intelligence and industrial Internet of Things technology. This technology can ensure that under the real-time guidance of a professional medical consortium, primary hospitals can realize rapid large-scale reverse planning design and can more accurately consider many factors of postprocessing. Although large-scale healthcare systems, such as volumetric-modulated arc therapy and other accurate radiotherapy technologies, have developed rapidly, the development of step-and-shoot-mode IMRT technology is still very important for developing countries. For software, in addition to the conformity of the dose distribution, the modulation speed, convenience and stability of the later dose delivery should also be considered in inverse planning. Therefore, this paper analyzes the main problems in conventional IMRT inverse planning, including the smoothing of the fluence map, the selection of the gantry angle and the dose leakage of tongue–groove effects. To address these issues, a novel Intelligent IoT-based large-scale inverse planning strategy with the key factors of the postmodulation is developed, and a detailed flow chart is also provided. The scheme consists of two steps. The first step is to obtain a relatively optimal combination of gantry angles by considering the dose distribution requirements and constraints and the modulation requirements and constraints. The second step is to optimize the intensity map, to smooth the map based on prior knowledge according to the determined angles, and to obtain the final modulation scheme according to the relevant objectives and constraints of the map decomposition (leaf sequencing). In an experiment, we calculate and validate the clinical head and neck case. Because of the special gantry angle selection, the angle combination is optimized from the initial equivalent distribution to adapt to the target area and protect the nontarget area. The value of the objective function varies greatly after the optimization, especially in the target area, and the target value decreases by approximately 10%. On this basis, we smooth the fluence map by a partial differential equation with prior knowledge and a minimization of the total number of monitor units. It is also shown from the objective function value that the target value is essentially unchanged for the target area, while for the nontarget area, the value decreases by 16%, which is very impressive.

Incorporating human and learned domain knowledge into training deep neural networks: A differentiable dose-volume histogram and adversarial inspired framework for generating Pareto optimal dose distributions in radiation therapy

PURPOSE

We propose a novel domain-specific loss, which is a differentiable loss function based on the dose-volume histogram (DVH), and combine it with an adversarial loss for the training of deep neural networks. In this study, we trained a neural network for generating Pareto optimal dose distributions, and evaluate the effects of the domain-specific loss on the model performance.

METHODS

In this study, three loss functions - mean squared error (MSE) loss, DVH loss, and adversarial (ADV) loss - were used to train and compare four instances of the neural network model: (a) MSE, (b) MSE + ADV, (c) MSE + DVH, and (d) MSE + DVH+ADV. The data for 70 prostate patients, including the planning target volume (PTV), and the organs at risk (OAR) were acquired as 96 × 96 × 24 dimension arrays at 5 mm3 voxel size. The dose influence arrays were calculated for 70 prostate patients, using a 7 equidistant coplanar beam setup. Using a scalarized multicriteria optimization for intensity-modulated radiation therapy, 1200 Pareto surface plans per patient were generated by pseudo-randomizing the PTV and OAR tradeoff weights. With 70 patients, the total number of plans generated was 84 000 plans. We divided the data into 54 training, 6 validation, and 10 testing patients. Each model was trained for a total of 100,000 iterations, with a batch size of 2. All models used the Adam optimizer, with a learning rate of 1 × 10-3 .

RESULTS

Training for 100 000 iterations took 1.5 days (MSE), 3.5 days (MSE+ADV), 2.3 days (MSE+DVH), and 3.8 days (MSE+DVH+ADV). After training, the prediction time of each model is 0.052 s. Quantitatively, the MSE+DVH+ADV model had the lowest prediction error of 0.038 (conformation), 0.026 (homogeneity), 0.298 (R50), 1.65% (D95), 2.14% (D98), and 2.43% (D99). The MSE model had the worst prediction error of 0.134 (conformation), 0.041 (homogeneity), 0.520 (R50), 3.91% (D95), 4.33% (D98), and 4.60% (D99). For both the mean dose PTV error and the max dose PTV, Body, Bladder and rectum error, the MSE+DVH+ADV outperformed all other models. Regardless of model, all predictions have an average mean and max dose error <2.8% and 4.2%, respectively.

CONCLUSION

The MSE+DVH+ADV model performed the best in these categories, illustrating the importance of both human and learned domain knowledge. Expert human domain-specific knowledge can be the largest driver in the performance improvement, and adversarial learning can be used to further capture nuanced attributes in the data. The real-time prediction capabilities allow for a physician to quickly navigate the tradeoff space for a patient, and produce a dose distribution as a tangible endpoint for the dosimetrist to use for planning. This is expected to considerably reduce the treatment planning time, allowing for clinicians to focus their efforts on the difficult and demanding cases.

A feasibility study for predicting optimal radiation therapy dose distributions of prostate cancer patients from patient anatomy using deep learning

With the advancement of treatment modalities in radiation therapy for cancer patients, outcomes have improved, but at the cost of increased treatment plan complexity and planning time. The accurate prediction of dose distributions would alleviate this issue by guiding clinical plan optimization to save time and maintain high quality plans. We have modified a convolutional deep network model, U-net (originally designed for segmentation purposes), for predicting dose from patient image contours of the planning target volume (PTV) and organs at risk (OAR). We show that, as an example, we are able to accurately predict the dose of intensity-modulated radiation therapy (IMRT) for prostate cancer patients, where the average Dice similarity coefficient is 0.91 when comparing the predicted vs. true isodose volumes between 0% and 100% of the prescription dose. The average value of the absolute differences in [max, mean] dose is found to be under 5% of the prescription dose, specifically for each structure is [1.80%, 1.03%](PTV), [1.94%, 4.22%](Bladder), [1.80%, 0.48%](Body), [3.87%, 1.79%](L Femoral Head), [5.07%, 2.55%](R Femoral Head), and [1.26%, 1.62%](Rectum) of the prescription dose. We thus managed to map a desired radiation dose distribution from a patient's PTV and OAR contours. As an additional advantage, relatively little data was used in the techniques and models described in this paper.