Monte Carlo Dose Distributions Research Articles

Practical experience with initial quality assurance of a high dose rate brachytherapy grid-based Boltzmann solver algorithm.

The purpose of this study was to validate the use of a model-based dose calculation algorithm (MBDCA), Acuros BV, for high dose rate brachytherapy treatment planning for a community-based hospital with a Bravos afterloader. Based on published AAPM recommendations, this work details a practical approach for community-based clinics to complete initial validation of Acuros BV, in order to add a MBDCA to a TG-43 based brachytherapy treatment planning program. Source dimensions and materials used in Acuros BV and TG-43 source models were compared to the physical source. TG-186 testing was completed with standardized test cases externally calculated with Monte Carlo compared to locally calculated with Acuros BV. Point doses calculated using TG-43 were compared to those calculated with Acuros BV in water at various dose grid settings. Secondary dose check software was used to evaluate dose distributions resembling clinical patient plans, both in water and on CT datasets representative of patient anatomy. The major source of discrepancy of source models was the length of modeled steel cable. TG-186 testing showed that the largest differences between Monte Carlo and Acuros BV dose distributions were located along the source axis for cases calculated in water, as well as located in regions of high dose gradients and within the applicator for the case calculated with a generic shielded applicator. An audit of point doses calculated with both TG-43 and Acuros BV in water found that dose grid settings significantly affected agreement. Secondary dose check software indicated that Acuros BV functioned satisfactorily, and a 5% threshold was adopted for secondary dose checks on gynecologic plans. This validation process indicated that Acuros BV met expected standards and affirmed its suitability for integration into this clinical practice's brachytherapy treatment planning.

Development and optimisation of grid inserts for a preclinical radiotherapy system and corresponding Monte Carlo beam simulations.

Objective. To develop a physical grid collimator compatible with the X-RAD preclinical radiotherapy system and create a corresponding Monte Carlo (MC) model.Approach. This work presents a methodology for the fabrication of a grid collimator designed for utilisation on the X-RAD preclinical radiotherapy system. Additionally, a MC simulation of the grid is developed, which is compatible with the X-RAD treatment planning system. The grid was manufactured by casting a low melting point alloy, cerrobend, into a silicone mould. The silicone was moulded around a 3D-printed replica of the grid, enabling the production of diverging holes with precise radii and spacing. A MC simulation was conducted on an equivalent 3D grid model and validated using 11 layers of GAFChromic EBT-3 film interspersed in a 3D-printed water-equivalent phantom. A 3D dose distribution was constructed from the film layers, enabling a direct comparison with the MC Simulation.Main results. The film and the MC dose distribution demonstrated a gamma passing rate of 99% for a 1%, 0.5 mm criteria with a 10% threshold applied. The peak-to-valley dose ratio and output factor at the surface were determined to be 20.4 and 0.79, respectively.Significance. The pairing of the grid collimator with a MC simulation can significantly enhance the practicality of grid therapy on the X-RAD. This combination enables further exploration of the biological implications of grid therapy, supported by a knowledge of the complex dose distributions. Moreover, this methodology can be adapted for use in other systems and scenarios.

Evaluation of RADIANCE Monte Carlo algorithm for treatment planning in electron based Intraoperative Radiotherapy (IOERT)

PurposeTo perform experimental as well as independent Monte Carlo (MC) evaluation of the MC algorithm implemented in RADIANCE version 4.0.8, a dedicated treatment planning system (TPS) for 3D electron dose calculations in intraoperative radiation therapy (IOERT). Methods and materialsThe MOBETRON 2000 (IntraOp Medical Corporation, Sunnyvale, CA) IOERT accelerator was employed. PDD and profiles for five cylindrical plastic applicators with 50–90 mm diameter and 0°, 30° beveling were measured in a water phantom, at nominal energies of 6, 9 and 12 MeV. Additional PDD measurements were performed for all the energies without applicator. MC modeling of the MOBETRON was performed with the user code BEAMnrc and egs_chamber of the MC simulation toolkit EGSnrc. The generated phase space files of the two 0°-bevel applicators (50 mm, 80 mm) and three energies in both RADIANCE and BEAMnrc, were used to determine PDD and profiles in various set-ups of virtual water phantoms with air and bone inhomogeneities. 3D dose distributions were also calculated in image data sets of an anthropomorphic tissue-equivalent pelvis phantom. Image acquisitions were realized with a CT scanner (Philips Big Bore CT, Netherlands). Gamma analysis was applied to quantify the deviations of the RADIANCE calculations to the measurements and EGSnrc calculations. Gamma criteria normalized to the global maximum were investigated between 2%, 2 mm and 3%, 3 mm. ResultsRADIANCE MC calculations satisfied the gamma criteria of 3%, 3 mm with a tolerance limit of 85% passing rate compared to in- water phantom measurements, except for the dose profiles of the 30° beveled applicators. Mismatches lay in surface doses, in umbra regions and in the beveled end of the 30° applicators. A very good agreement to the EGSnrc calculations in heterogeneous media was observed. Deviations were more pronounced for the larger applicator diameter and higher electron energy. In 3D dose comparisons in the anthropomorphic phantom, gamma passing rates were higher than 96 % for both simulated applicators. ConclusionsRADIANCE MC algorithm agrees within 3%, 3 mm criteria with in-water phantom measurements and EGSnrc MC dose distributions in heterogeneous media for 0°-bevel applicators. The user should be aware of missing scattering components and the 30° beveled applicators should be used with attention.

A fast analytical dose calculation approach for MRI-guided proton therapy.

Objective.Magnetic resonance (MR) is an innovative technology for online image guidance in conventional radiotherapy and is also starting to be considered for proton therapy as well. For MR-guided therapy, particularly for online plan adaptations, fast dose calculation is essential. Monte Carlo (MC) simulations, however, which are considered the gold standard for proton dose calculations, are very time-consuming. To address the need for an efficient dose calculation approach for MRI-guided proton therapy, we have developed a fast GPU-based modification of an analytical dose calculation algorithm incorporating beam deflections caused by magnetic fields.Approach.Proton beams (70-229 MeV) in orthogonal magnetic fields (0.5/1.5 T) were simulated using TOPAS-MC and central beam trajectories were extracted to generate look-up tables (LUTs) of incremental rotation angles as a function of water-equivalent depth. Beam trajectories are then reconstructed using these LUTs for the modified ray casting dose calculation. The algorithm was validated against MC in water, different materials and for four example patient cases, whereby it has also been fully incorporated into a treatment plan optimisation regime.Main results.Excellent agreement between analytical and MC dose distributions could be observed with sub-millimetre range deviations and differences in lateral shifts <2 mm even for high densities (1000 HU). 2%/2 mm gamma pass rates were comparable to the 0 T scenario and above 94.5% apart for the lung case. Further, comparable treatment plan quality could be achieved regardless of magnetic field strength.Significance.A new method for accurate and fast proton dose calculation in magnetic fields has been developed and successfully implemented for treatment plan optimisation.

Deep learning-based fast denoising of Monte Carlo dose calculation in carbon ion radiotherapy.

Plan verification is one of the important steps of quality assurance (QA) in carbon ion radiotherapy. Conventional methods of plan verification are based on phantom measurement, which is labor-intensive and time-consuming. Although the plan verification method based on Monte Carlo (MC) simulation provides a more accurate modeling of the physics, it is also time-consuming when simulating with a large number of particles. Therefore, how to ensure the accuracy of simulation results while reducing simulation time is the current difficulty and focus. The purpose of this work was to evaluate the feasibility of using deep learning-based MC denoising method to accelerate carbon-ion radiotherapy plan verification. Three models, including CycleGAN, 3DUNet and GhostUNet with Ghost module, were used to denoise the 1×106 carbon ions-based MC dose distribution to the accuracy of 1×108 carbon ions-based dose distribution. The CycleGAN's generator, 3DUNet and GhostUNet were all derived from the 3DUNet network. A total of 59 cases including 29 patients with head-and-neck cancers and 30 patients with lung cancers were collected, and 48 cases were randomly selected as the training set of the CycleGAN network and six cases as the test set. For the 3DUNet and GhostUNet models, the numbers of training set, validation set, and test set were 47, 6, and 6, respectively. Finally, the three models were evaluated qualitatively and quantitatively using RMSE and three-dimensional gamma analysis (3mm, 3%). The three end-to-end trained models could be used for denoising the 1×106 carbon ions-based dose distribution, and their generalization was proved. The GhostUNet obtained the lowest RMSE value of 0.075, indicating the smallest difference between its denoised and 1×108 carbon ions-based dose distributions. The average gamma passing rate (GPR) between the GhostUNet denoising-based versus 1×108 carbon ions-based dose distributions was 99.1%, higher than that of the CycleGAN at 94.3% and the 3DUNet at 96.2%. Among the three models, the GhostUNet model had the fewest parameters (4.27 million) and the shortest training time (99s per epoch) but achieved the best denoising results. The end-to-end deep network GhostUNet outperforms the CycleGAN, 3DUNet models in denoising MC dose distributions for carbon ion radiotherapy. The network requires less than 5s to denoise a sample of MC simulation with few particles to obtain a qualitative and quantitative result comparable to the dose distribution simulated by MC with relatively large number particles, offering a significant reduction in computation time.

AAPMWGDCAB Report 372:A joint AAPM, ESTRO, ABG, and ABS report oncommissioning of model-based dose calculation algorithms in brachytherapy.

The introduction of model-based dose calculation algorithms (MBDCAs) in brachytherapy provides an opportunity for a more accurate dose calculation and opens the possibility for novel, innovative treatment modalities. The joint AAPM, ESTRO, and ABG Task Group 186 (TG-186) report provided guidance to early adopters. However, the commissioning aspect of these algorithms was described only in general terms with no quantitative goals. This report, from the Working Group on Model-Based Dose Calculation Algorithms in Brachytherapy, introduced a field-tested approach to MBDCA commissioning. It is based on a set of well-characterized test cases for which reference Monte Carlo (MC) and vendor-specific MBDCA dose distributions are available in a Digital Imaging and Communications in Medicine-Radiotherapy (DICOM-RT) format to the clinical users. The key elements of the TG-186 commissioning workflow are now described in detail, and quantitative goals are provided. This approach leverages the well-known Brachytherapy Source Registry jointly managed by the AAPM and the Imaging and Radiation Oncology Core (IROC) Houston Quality Assurance Center (with associated links at ESTRO) to provide open access to test cases as well as step-by-step user guides. While the current report is limited to the two most widely commercially available MBDCAs and only for 192 Ir-based afterloading brachytherapy at this time, this report establishes a general framework that can easily be extended to other brachytherapy MBDCAs and brachytherapy sources. The AAPM, ESTRO, ABG, and ABS recommend that clinical medical physicists implement the workflow presented in this report to validate both the basic and the advanced dose calculation features of their commercial MBDCAs. Recommendations are also given to vendors to integrate advanced analysis tools into their brachytherapy treatment planning system to facilitate extensive dose comparisons. The use of the test cases for research and educational purposes is further encouraged.

Dose Profiles Distribution of 6 MV Linear Accelerator Using Monte Carlo Code System for Optimum Treatment planning in Radiotherapy Department

Monte Carlo (MC) radiotherapy treatment planning has become practical in medical physics, particularly with the rapid development of computer technology. Requires detailed knowledge of linear medical accelerators (Linacs) beam radiation. In the present work the basic dosimetric properties of the 6-MV photon beam and the Varian Clinic 600CLinac were examined. The dosimetric features of interest are, the absorbed dose from the central axis, beam profiles, photon blockers, electron and positron, mean energy, spectral distribution, and isodose curves. Varian Clinic 600C Linac is simulated using the OMEGA BEAMnrc MC code system. Detailed spectra of phase space profiles of a photon beam with a field of view of 10 cm × 10 cm were modeled, simulated and finally analyzed using BEAMdp. The central axis depth dose curves, dose profiles and isodose curves for photon beams in the aqueous phantom were also recorded and analyzed using the DOSXYZnrc, STATDOSE and DOSXYZ_SHOW code. This study demonstrates that MC can generate phase space data files that can be used to generate accurate MC dose distributions for photon beams produced using a high-powered clinical linear accelerator in a water phantom or in patients for radiotherapy.

Fast D M,M calculation in LDR brachytherapy using deep learning methods

Objective. The Monte Carlo (MC) method provides a complete solution to the tissue heterogeneity effects in low-energy low-dose rate (LDR) brachytherapy. However, long computation times limit the clinical implementation of MC-based treatment planning solutions. This work aims to apply deep learning (DL) methods, specifically a model trained with MC simulations, to predict accurate dose to medium in medium (D M,M) distributions in LDR prostate brachytherapy. Approach. To train the DL model, 2369 single-seed configurations, corresponding to 44 prostate patient plans, were used. These patients underwent LDR brachytherapy treatments in which 125I SelectSeed sources were implanted. For each seed configuration, the patient geometry, the MC dose volume and the single-seed plan volume were used to train a 3D Unet convolutional neural network. Previous knowledge was included in the network as an r 2 kernel related to the first-order dose dependency in brachytherapy. MC and DL dose distributions were compared through the dose maps, isodose lines, and dose-volume histograms. Features enclosed in the model were visualized. Main results. Model features started from the symmetrical kernel and finalized with an anisotropic representation that considered the patient organs and their interfaces, the source position, and the low- and high-dose regions. For a full prostate patient, small differences were seen below the 20% isodose line. When comparing DL-based and MC-based calculations, the predicted CTV D 90 metric had an average difference of −0.1%. Average differences for OARs were −1.3%, 0.07%, and 4.9% for the rectum D 2cc, the bladder D 2cc, and the urethra D 0.1cc. The model took 1.8 ms to predict a complete 3D D M,M volume (1.18 M voxels). Significance. The proposed DL model stands for a simple and fast engine which includes prior physics knowledge of the problem. Such an engine considers the anisotropy of a brachytherapy source and the patient tissue composition.

A MC-based anthropomorphic test case for commissioning model-based dose calculation in interstitial breast 192-Ir HDR brachytherapy.

To provide the first clinical test case for commissioning of 192 Ir brachytherapy model-based dose calculation algorithms (MBDCAs) according to the AAPM TG-186 report workflow. A computational patient phantom model was generated from a clinical multi-catheter 192 Ir HDR breast brachytherapy case. Regions of interest (ROIs) were contoured and digitized on the patient CT images and the model was written to a series of DICOM CT images using MATLAB. The model was imported into two commercial treatment planning systems (TPSs) currently incorporating an MBDCA. Identical treatment plans were prepared using a generic 192 Ir HDR source and the TG-43-based algorithm of each TPS. This was followed by dose to medium in medium calculations using the MBDCA option of each TPS. Monte Carlo (MC) simulation was performed in the model using three different codes and information parsed from the treatment plan exported in DICOM radiation therapy (RT) format. Results were found to agree within statistical uncertainty and the dataset with the lowest uncertainty was assigned as the reference MC dose distribution. The dataset is available online at http://irochouston.mdanderson.org/rpc/BrachySeeds/BrachySeeds/index.html,https://doi.org/10.52519/00005. Files include the treatment plan for each TPS in DICOM RT format, reference MC dose data in RT Dose format, as well as a guide for database users and all files necessary to repeat the MC simulations. The dataset facilitates the commissioning of brachytherapy MBDCAs using TPS embedded tools and establishes a methodology for the development of future clinical test cases. It is also useful to non-MBDCA adopters for intercomparing MBDCAs and exploring their benefits and limitations, as well as to brachytherapy researchers in need of a dosimetric and/or a DICOM RT information parsing benchmark. Limitations include specificity in terms of radionuclide, source model, clinical scenario, and MBDCA version used for its preparation.

A plan verification platform for online adaptive proton therapy using deep learning-based Monte–Carlo denoising

Background PurposeThis study focused on developing a fast Monte Carlo (MC) plan verification platform via a deep learning (DL)-based denoising approach. It can maintain the MC dose calculation accuracy while significantly reducing the computation time. We also investigated its potential applications for online adaptive proton therapy (APT). MethodsFirst, we modeled an MC platform for proton therapy using the beam data library (BDL) required for treatment planning systems and then tested it with measured data. To accelerate the dose calculation, a dl-based denoising model with deep ResNet-deconvolution networks was developed. It was trained on the MC dose distribution of tumor sites obtained from 52 patients. The input MC dose distribution was with 1 × 106 simulated protons and the reference was 1 × 108. Fivefold cross-validation was performed. ResultsComparing the MC model with measured data, the range agreement (point-to-point difference) was better than 0.85 mm, and the lateral dose profile difference was below 2.41 %. For the denoising approach, we found a significant improvement in the dose volume histogram (DVH) for predicted images compared with input images. The root mean squared error (RMSE) for predicted versus reference images was 3.94 times lower than that of the input versus reference images. Moreover, for the gamma passing rate (3 mm/3%), the predicted versus reference images have an average of 99 %, much higher than the 82 % of the input versus reference images. The MC model successfully denoised the test dose map (high noise) to approach the reference (low noise). The elapsed time can be reduced to < 60 s (simulation time [high noise] + predicted time), much lower than the simulation time of a low noise dose map (e.g., >100 min of simulation of 1 + E8 particles). ConclusionsWe propose an analogous end-to-end fast plan verification platform using the combination of MC and DL methods. The platform yields dose calculation accuracy similar to MC codes while significantly reducing the elapsed time and can be used for online APT as an alternative to online plan verification.

A novel multichannel deep learning model for fast denoising of Monte Carlo dose calculations: preclinical applications

Objective. In preclinical radiotherapy with kilovolt (kV) x-ray beams, accurate treatment planning is needed to improve the translation potential to clinical trials. Monte Carlo based radiation transport simulations are the gold standard to calculate the absorbed dose distribution in external beam radiotherapy. However, these simulations are notorious for their long computation time, causing a bottleneck in the workflow. Previous studies have used deep learning models to speed up these simulations for clinical megavolt (MV) beams. For kV beams, dose distributions are more affected by tissue type than for MV beams, leading to steep dose gradients. This study aims to speed up preclinical kV dose simulations by proposing a novel deep learning pipeline. Approach. A deep learning model is proposed that denoises low precision (∼106 simulated particles) dose distributions to produce high precision (109 simulated particles) dose distributions. To effectively denoise the steep dose gradients in preclinical kV dose distributions, the model uses the novel approach to use the low precision Monte Carlo dose calculation as well as the Monte Carlo uncertainty (MCU) map and the mass density map as additional input channels. The model was trained on a large synthetic dataset and tested on a real dataset with a different data distribution. To keep model inference time to a minimum, a novel method for inference optimization was developed as well. Main results. The proposed model provides dose distributions which achieve a median gamma pass rate (3%/0.3 mm) of 98% with a lower bound of 95% when compared to the high precision Monte Carlo dose distributions from the test set, which represents a different dataset distribution than the training set. Using the proposed model together with the novel inference optimization method, the total computation time was reduced from approximately 45 min to less than six seconds on average. Significance. This study presents the first model that can denoise preclinical kV instead of clinical MV Monte Carlo dose distributions. This was achieved by using the MCU and mass density maps as additional model inputs. Additionally, this study shows that training such a model on a synthetic dataset is not only a viable option, but even increases the generalization of the model compared to training on real data due to the sheer size and variety of the synthetic dataset. The application of this model will enable speeding up treatment plan optimization in the preclinical workflow.

Elaboration and experimental validation of a Monte Carlo source model for linac 6 MV photon beams with and without Flattening Filter

The concept of virtual source model has been developed to provide an efficient and precise particle generator mainly involved in standard photon beam treatment planing and Monte Carlo dosimetry simulations. In this work, we propose a source model for 6 MV linac photon beam for both irradiation modes, with and without flattening filter. Two sources are considered to represent primary and a scattered radiations. Energy spectra taking into account the correlation between energy and diffusion angle have been used. Model parameters were tuned through a dose distribution iterative optimisation approach. The convergence condition relies on γindex evaluation. The simulations were performed in two steps, phase space files were generated using Geant4 toolkit, then energy deposit in water were computed using DPM code. Monte Carlo dose distributions of the model were compared to experimental data of two linac systems, an Elekta Synergy® with flattening filter and a Varian TrueBeam® with and without flattening filter. Good results were obtained in terms of γindex 2%/2 mm criterion, dose difference and distance to agreement (DTA) analysis for field size range from 3 × 3 cm2 to 30 × 30 cm2.

Integration of an Independent Monitor Unit Check for High-Magnetic-Field MR-Guided Radiation Therapy System.

PurposeCommercial independent monitor unit (IMU) check systems for high-magnetic-field MR-guided radiation therapy (RT) systems are lacking. We investigated the feasibility of adopting an existing treatment planning system (TPS) as an IMU check for online adaptive radiotherapy using 1.5-Tesla MR-Linac.MethodsThe 7-MV flattening filter free (FFF) beam and multi-leaf collimator (MLC) models of a 1.5-T Elekta Unity MR-Linac within Monte Carlo-based Monaco TPS were used to generate an optimized beam model in Eclipse TPS. The MLC dosimetric leaf gap of the beam in Eclipse was determined by matching the dose distribution of Eclipse-generated intensity-modulated radiation therapy (IMRT) plans using the Analytical Anisotropic Algorithm (AAA) algorithm to Monaco plans. The plans were automatically adjusted for different source-to-axis distances (SADs) between the two systems. For IMU check, the treatment plans developed in Monaco were transferred to Eclipse to recalculate the dose using AAA. A plug-in within Eclipse was created to perform a 2D gamma analysis of the AAA and Monte Carlo dose distribution on a beam’s eye view parallel plane. Monaco dose distribution was shifted laterally by 2 mm during gamma analysis to account for the impact of magnetic field on electron trajectories. Eclipse doses for posterior beams were corrected for both the Unity couch and the posterior MR coil attenuation. Thirteen patients, each with 4–5 fractions for a variety of tumor sites (pancreas, rectum, and prostate), were tested.ResultsAfter thorough commissioning, the method was implemented as part of the standard clinical workflow. A total of 62 online plans, each with approximately 15 beams, were evaluated. The average per-beam gamma (3%/3 mm) pass rate for plans was 97.9% (range, 95.9% to 98.8%). The average pass rate per beam for all 932 beams used in these plans was 97.9% ± 1.9%, with the lowest per-beam gamma pass rate at 88.4%. The time for the process was within 3.2 ± 0.9 min.ConclusionThe use of a second planning system provides an efficient way to perform IMU checks with clinically acceptable accuracy for online adaptive plans on Unity MR-Linac. This is essential for meeting the safety requirements for second checks as outlined in American Association of Physicists in Medicine Task Group (AAPM TG) reports 114 and 219.

Abstract P3-19-10: Subcutaneous layer dosimetry of the breast and chest wall at clinical beam energies without bolus: A Monte Carlo and analytical anisotropic algorithm (AAA) calculation study

Abstract PURPOSE: Breast skin thickness is 3 mm (1). The subcutaneous layer (SL) lies immediately beneath the skin and is at risk for local recurrence after breast cancer surgery and adjuvant radiotherapy. As the SL lies within the buildup region for megavoltage radiation treatment, placement of bolus for patients receiving chest wall (CW) radiotherapy (RT) is routine at many centers. The dermal toxicity of bolus is well known: in 12 studies of CW RT, the pooled risk of Grade 3 acute toxicity is 9.6% with bolus and 1.2% without bolus (2). Meanwhile, bolus is rarely used after breast-conserving surgery (BCS) RT for patients with similar cancers. This study examines variation in tangential RT dose coverage of the SL in the intact breast and bolus-free CW with clinically relevant photon beam energies using Monte Carlo (MC) calculations and a commercial treatment planning system (TPS). METHODS: Thirty CT datasets from patients without skin involvement were identified. There were two groups of patients: 15 treated with BCS RT and 15 treated with CW RT. In each group, 5-patient subgroups had tangent-beam RT planned without bolus with 6, 10, and 15 MV photons, respectively, using the Analytical Anisotropic Algorithm (AAA) algorithm (v.13.6.23) in Eclipse v.15.6 (Varian Medical Systems Inc., Palo Alto, CA). On each CT, the SL was segmented as a high-resolution shell from 3 to 5 mm below the body contour in the Eclipse TPS v.15.6 (Varian Medical Systems Inc., Palo Alto, CA). A 1x1x1 mm MC dose simulation was performed using EGSnrc code (BEAMnrc/DOSXYZnrc). The MC dose distributions were imported back into the TPS for comparison with AAA calculations. The V95% and V90% for the SL were calculated for each case and the mean V95% and V90% were reported for each subgroup. A t-test was used with a two-sided alpha = 0.05 for statistical analysis. RESULTS: The mean separation increased with use of higher energies for both BCS and CW RT. The MC-calculated mean SL V90% and V95% were higher for CW RT than for BCS RT at each energy. The V90% coverage was 91.5% for CW and 74.4 % for BCS at 6 MV (p&amp;lt;0.001), 89.3% for CW and 61.3% for BCS (p&amp;lt;0.001) at 10 MV and 87.1% for CW and 60.9% for BCS (p&amp;lt;0.001) at 15 MV (Table 1). For SL V95% the CW coverage was higher than the BCS coverage for every energy. For SL V90% at 6 MV, the AAA and MC calculations agreed within 2.5%, with the MC being slightly higher. The agreement between AAA and MC decreased for higher energies with MC reporting higher SL V90% coverage by up to 16.3%. The higher MC-calculated dose to the SL is consistent with the literature (3). CONCLUSION: MC and AAA SL dose calculations agreed well for 6 MV, but AAA underestimated the dose for 10 and 15 MV. For 6-15 MV photons, the MC-calculated dosimetric coverage of the SL is higher for CW RT than BCS RT. Since radiation oncologists are satisfied with the SL coverage by BCS RT, bolus is not needed for CW RT, because, without bolus, CW RT delivers a higher SL dose than BCS RT. REFERENCES:. 1.Pope TL Jr, et al.J Can Assoc Radiol. 1984 Dec;35(4):365-8. PMID: 6526847. 2.Dahn HM, et al. Crit Rev Oncol Hematol. 2021 Jun 5;163:10339. PMID: 34102286. 3.Panettieri V, et al., Radiother Oncol 2009; 93: 94-101 Table 1.Monte Carlo calculated mean V90% and V95% for Breast and Chest Wall for each energyV95%V95%V95%V90%V90%V90%Energy(MV)Chest Wall(Mean)Breast. (Mean)p-valueChest Wall(Mean)Breast. (Mean)p-value661.9%35.1%&amp;lt;0.00191.5%74.4%&amp;lt;0.0011065.2%39.7%&amp;lt;0.00189.3%61.3%&amp;lt;0.0011561.7%38.9%0.01287.1%60.9%&amp;lt;0.001 Citation Format: Dylan Narinesingh, Alan Nichol, Alanah Bergman, Tony Popescu. Subcutaneous layer dosimetry of the breast and chest wall at clinical beam energies without bolus: A Monte Carlo and analytical anisotropic algorithm (AAA) calculation study [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P3-19-10.

Denoising proton therapy Monte Carlo dose distributions in multiple tumor sites: A comparative neural networks architecture study.

Monte Carlo (MC) algorithms provide accurate modeling of dose calculation by simulating the delivery and interaction of many particles through patient geometry. Fast MC simulations using large number of particles are desirable as they can lead to reliable clinical decisions. In this work, we assume that faster simulations with fewer particles can approximate slower ones by denoising them with deep learning. We use mean squared error (MSE) as loss function to train networks (sNet and dUNet), with 2.5D and 3D setups considering volumes of 7 and 24 slices. Our models are trained on proton therapy MC dose distributions of six different tumor sites acquired from 50 patients. We provide networks with input MC dose distributions simulated using 1×106 particles while keeping 1×109 particles as reference. On average over 10 new patients with different tumor sites, in 2.5D and 3D, our models recover relative residual error on target volume, ΔD95TV of 0.67±0.43% and 1.32±0.87% for sNet vs. 0.83±0.53% and 1.66±0.98% for dUNet, compared to the noisy input at 12.40±4.06%. Moreover, the denoising time for a dose distribution is: < 9s and <1s for sNet vs. < 16s and <1.5s for dUNet in 2.5D and 3D, in comparison to about 100min (MC simulation using 1×109 particles). We propose a fast framework that can successfully denoise MC dose distributions. Starting from MC doses with 1×106 particles only, the networks provide comparable results as MC doses with1×109 particles, reducing simulation time significantly.

Radiation dose calculation in 3D heterogeneous media using artificial neural networks.

External beam radiotherapy (EBRT) treatment planning requires a fast and accurate method of calculating the dose delivered by a clinical treatment plan. However, existing methods of calculating dose distributions have limitations. Monte Carlo (MC) methods are accurate but can take too long to be clinically viable. Deterministic approaches are quicker but can be inaccurate under certain conditions, particularly near heterogeneities and air interfaces. Neural networks trained on MC-derived data have the potential to reproduce dose distributions that agree closely with the MC method while being significantly quicker to deploy. In this work we present a framework for training machine learning models capable of directly calculating the dose delivered to a point in three-dimensional (3D) heterogeneous media given only spatially local information. The framework consists of three parts. First, we describe a novel method of randomly generating 3D heterogeneous geometries using simplex noise. Dose distributions for training were obtained by importing these geometries into a MC simulation. The second and third parts of the framework are precalculated data channels, aligned with the patient computed tomography (CT) image, to be used as input to the model. These data channels are a computationally efficient way of encoding the parameters of an incident radiation beam while also allowing the model to learn from data that would otherwise be outside of its receptive field. We demonstrate the viability of the framework by a training small, fully connected neural network model to reproduce dose distributions from megavoltage photon beams. The trained network displayed excellent agreement with MC dose distributions in randomly generated geometries with an average gamma index (3%/3mm) pass rate of 94.7% and an average error of 1.45% of peak dose. Finally, the network was used to calculate the dose in a patient CT image, on which the network was not trained, producing similarly impressive results. A novel method of generating training data for learned radiation dosimetry models has been introduced, along with preprocessing steps that allow even simple models to reproduce accurate dose distributions for EBRT. More importantly, we have demonstrated that a model trained using the proposed framework can generalize from the training data to predicting the therapeutic dose in realistic media.

Monte Carlo calculations of radiotherapy dose in "homogeneous" anatomy.

Given the substantial literature on the use of Monte Carlo (MC) simulations to verify treatment planning system (TPS) calculations of radiotherapy dose in heterogeneous regions, such as head and neck and lung, this study investigated the potential value of running MC simulations of radiotherapy treatments of nominally homogeneous pelvic anatomy. A pre-existing in-house MC job submission and analysis system, built around BEAMnrc and DOSXYZnrc, was used to evaluate the dosimetric accuracy of a sample of 12 pelvic volumetric arc therapy (VMAT) treatments, planned using the Varian Eclipse TPS, where dose was calculated with both the Analytical Anisotropic Algorithm (AAA) and the Acuros (AXB) algorithm. In-house TADA (Treatment And Dose Assessor) software was used to evaluate treatment plan complexity, in terms of the small aperture score (SAS), modulation index (MI) and a novel exposed leaf score (ELS/ELA). Results showed that the TPS generally achieved closer agreement with the MC dose distribution when treatments were planned for smaller (single-organ) targets rather than larger targets that included nodes or metastases. Analysis of these MC results with reference to the complexity metrics indicated that while AXB was useful for reducing dosimetric uncertainties associated with density heterogeneity, the residual TPS dose calculation uncertainties resulted from treatment plan complexity and TPS model simplicity. The results of this study demonstrate the value of using MC methods to recalculate and check the dose calculations provided by commercial radiotherapy TPSs, even when the treated anatomy is assumed to be comparatively homogeneous, such as in the pelvic region.

Artifact-free CT images for electron beam therapy using a patient-specific non metallic shield

Patient’s CT images taken with metallic shields for radiotherapy suffer from artifacts. Furthermore, the treatment planning system (TPS) has a limitation on accurate dose calculations for high density materials. In this study, a Monte Carlo (MC)-based method was developed to accurately evaluate the dosimetric effect of the metallic shield. Two patients with a commercial tungsten shield of lens and two patients with a custom-made lead shield of lip were chosen to produce their non-metallic dummy shields using 3D scanner and printer. With these dummy shields, we generated artifact-free CT images. The maximum CT number allowed in TPS was assigned to metallic shields. MC simulations with real material information were carried out. In addition, clinically relevant dose-volumetric parameters were calculated for the comparison between MC and TPS. Relative dosimetry was performed using radiochromic films. The dose reductions below metallic structures were shown on MC dose distributions, but not evident on TPS dose distributions. The differences in dose-volumetric parameters of PTV between TPS and MC for eye shield cases were not clearly shown. However, the mean dose of lens from TPS and MC was different. The MC results were in superior agreement with measured data in relative dosimetry. The lens dose could be overestimated by TPS. The differences in dose-volumetric parameters of PTV between TPS and MC were generally larger in lip cases than in eye cases. The developed method is useful in predicting the realistic dose distributions around the organs blocked by the metallic shields.

Mitigating inherent noise in Monte Carlo dose distributions using dilated U-Net.

Monte Carlo (MC) algorithms offer accurate modeling of dose calculation by simulating the transport and interactions of many particles through the patient geometry. However, given their random nature, the resulting dose distributions have statistical uncertainty (noise), which prevents making reliable clinical decisions. This issue is partly addressable using a huge number of simulated particles but is computationally expensive as it results in significantly greater computation times. Therefore, there is a trade-off between the computation time and the noise level in MC dose maps. In this work, we address the mitigation of noise inherent to MC dose distributions using dilated U-Net - an encoder-decoder-styled fully convolutional neural network, which allows fast and fully automated denoising of whole-volume dose maps. We use mean squared error (MSE) as loss function to train the model, where training is done in 2D and 2.5D settings by considering a number of adjacent slices. Our model is trained on proton therapy MC dose distributions of different tumor sites (brain, head and neck, liver, lungs, and prostate) acquired from 35 patients. We provide the network with input MC dose distributions simulated using particles while keeping particles as reference. After training, our model successfully denoises new MC dose maps. On average (averaged over five patients with different tumor sites), our model recovers of 55.99Gy from the noisy MC input of 49.51Gy, whereas the low noise MC (reference) offers 56.03Gy. We observed a significant reduction in average RMSE (thresholded >10% max ref) for reference vs denoised (1.25Gy) than reference vs input (16.96Gy) leading to an improvement in signal-to-noise ratio (ISNR) by 18.06dB. Moreover, the inference time of our model for a dose distribution is less than 10s vs 100min (MC simulation using particles). We propose an end-to-end fully convolutional network that can denoise Monte Carlo dose distributions. The networks provide comparable qualitative and quantitative results as the MC dose distribution simulated with particles, offering a significant reduction in computation time.

Preclinical dose verification using a 3D printed mouse phantom for radiobiology experiments.

Dose verification in preclinical radiotherapy is often challenged by a lack of standardization in the techniques and technologies commonly employed along with the inherent difficulty of dosimetry associated with small-field kilovoltage sources. As a consequence, the accuracy of dosimetry in radiobiological research has been called into question. Fortunately, the development and characterization of realistic small-animal phantoms has emerged as an effective and accessible means of improving dosimetric accuracy and precision in this context. The application of three-dimensional (3D) printing, in particular, has enabled substantial improvements in the conformity of representative phantoms with respect to the small animals they are modeled after. In this study, our goal was to evaluate a fully 3D printed mouse phantom for use in preclinical treatment verification of sophisticated therapies for various anatomical targets of therapeutic interest. An anatomically realistic mouse phantom was 3D printed based on segmented microCT data of a tumor-bearing mouse. The phantom was modified to accommodate both laser-cut EBT3 radiochromic film within the mouse thorax and a plastic scintillator dosimeter (PSD), which may be placed within the brain, abdomen, or 1-cm flank subcutaneous tumor. Various treatments were delivered on an image-guided small-animal irradiator in order to determine the doses to isocenter using a PSD and validate lateral- and depth-dose distributions using film dosimeters. On-board cone-beam CT imaging was used to localize isocenter to the film plane or PSD active element prior to irradiation. The PSD irradiations comprised a 3×3mm2 brain arc, 5×5mm2 parallel-opposed pair (POP), and 5-beam 10×10mm2 abdominal coplanar arrangement while two-dimensional (2D) film dose distributions were acquired using a 3×3mm2 arc and both 5×5 and 10×10mm2 3-beam coplanar plans. A validated Monte Carlo (MC) model of the source was used as to verify the accuracy of the film and PSD dose measurements. computer-aided design (CAD) geometries for the mouse phantom and dosimeters were imported directly into the MC code to allow for highly accurate reproduction of the physical experiment conditions. Experimental and MC-derived film data were co-registered and film dose profiles were compared for points above 90% of the dose maximum. Point dose measurements obtained with the PSD were similarly compared for each of the candidate (brain, abdomen, and tumor) treatment sites. For each treatment configuration and anatomical target, the MC-calculated and measured doses met the proposed 5% agreement goal for dose accuracy in radiobiology experiments. The 2D film and MC dose distributions were successfully registered and mean doses for lateral profiles were found to agree to within 2.3% in all cases. Isocentric point-dose measurements taken with the PSD were similarly consistent, with a maximum percentage deviation of 3.2%. Our study confirms the utility of 3D printed phantom design in providing accurate dose estimates for a variety of preclinical treatment paradigms. As a tool for pretreatment dose verification, the phantom may be of particular interest to researchers for its ability to facilitate precise dosimetry while fostering a reduction in cost for radiobiology experiments.