Experimental verification and clinical implementation of a commercial Monte Carlo electron beam dose calculation algorithm

This study describes the modeling and the experimental verification and clinical implementation of the alpha release of Pinnacle3 Monte Carlo (MC) electron beam dose calculation algorithm for patient-specific treatment planning. The MC electron beam modeling was performed for beam energies ranging from 6 to 18 MeV from a Siemens (Primus) linear accelerator using standard-shaped electron applicators and 100 cm source-to-surface distance (SSD). The agreement between MC calculations and measurements was, on average, within 2% and 2 mm for all applicator sizes. However, differences of the order of 3%-4% were noted in the off-axis dose profiles for the largest applicator modeled and for all energies. Output factors were calculated for standard electron cones and square cutouts inserted in the 10 x 10 cm2 applicator for different SSDs and were found to be within 4% of measured data. Experimental verification of the MC electron beam model was carried out using an ionization chamber and film in solid-water slab and anthropomorphic phantoms containing bone and lung materials. Agreement between calculated and measured dose distributions was within +/-3%. Clinical comparison was performed in four patient treatment plans with lesions in highly irregular anatomies, such as the ear, face, and breast, where custom-designed bolus and field shaping blocks were used in the patient treatments. For comparison purposes, treatment planning was also performed using the conventional pencil beam (PB) algorithm with the Pinnacle3 treatment planning system. Differences between MC and PB dose calculations for the patient treatment plans were significant, particularly in anatomies where the target was in close proximity to low density tissues, such as lung and air cavities. Concerning monitor unit calculations, the largest differences obtained between MC and PB algorithms were between 4.0% and 5.0% for two patients treated with oblique beams and involving highly irregular surfaces, i.e., breast and cheek. Clinical results are reported for overall uncertainty values (averaged over voxels with doses >50% dosemax) ranging from 2% to 0.3% and calculations were performed using cubic voxels with side 0.3 cm. Timing values ranged from 2 min to 24.5 h, depending on the field size, beam energy, number, and thickness of computed tomography slices used to define the patient's anatomy for the overall uncertainty values mentioned above.     

Evaluation of the first commercial Monte Carlo dose calculation engine for electron beam treatment planning

The purpose of this study is to perform a clinical evaluation of the first commercial (MDS Nordion, now Nucletron) treatment planning system for electron beams incorporating Monte Carlo dose calculation module. This software implements Kawrakow's VMC++ voxel-based Monte Carlo calculation algorithm. The accuracy of the dose distribution calculations is evaluated by direct comparisons with extensive sets of measured data in homogeneous and heterogeneous phantoms at different source-to-surface distances (SSDs) and gantry angles. We also verify the accuracy of the Monte Carlo module for monitor unit calculations in comparison with independent hand calculations for homogeneous water phantom at two different SSDs. All electron beams in the range 6-20 MeV are from a Siemens KD-2 linear accelerator. We used 10,000 or 50,000 histories/cm2 in our Monte Carlo calculations, which led to about 2.5% and 1% relative standard error of the mean of the calculated dose. The dose calculation time depends on the number of histories, the number of voxels used to map the patient anatomy, the field size, and the beam energy. The typical run time of the Monte Carlo calculations (10,000 histories/cm2) is 1.02 min on a 2.2 GHz Pentium 4 Xeon computer for a 9 MeV beam, 10 x 10 cm2 field size, incident on the phantom 15 x 15 x 10 cm3 consisting of 31 CT slices and voxels size of 3 x 3 x 3 mm3 (total of 486,720 voxels). We find good agreement (discrepancies smaller than 5%) for most of the tested dose distributions. We also find excellent agreement (discrepancies of 2.5% or less) for the monitor unit calculations relative to the independent manual calculations. The accuracy of monitor unit calculations does not depend on the SSD used, which allows the use of one virtual machine for each beam energy for all arbitrary SSDs. In some cases the test results are found to be sensitive to the voxel size applied such that bigger systematic errors (>5%) occur when large voxel sizes interfere with the extensions of heterogeneities or dose gradients because of differences between the experimental and calculated geometries. Therefore, user control over voxelization is important for high accuracy electron dose calculations.     

Monte Carlo calculation of helical tomotherapy dose delivery

Helical tomotherapy delivers intensity modulated radiation therapy using a binary multileaf collimator (MLC) to modulate a fan beam of radiation. This delivery occurs while the linac gantry and treatment couch are both in constant motion, so the beam describes, from a patient/phantom perspective, a spiral or helix of dose. The planning system models this continuous delivery as a large number (51) of discrete gantry positions per rotation, and given the small jaw/fan width setting typically used (1 or 2.5 cm) and the number of overlapping rotations used to cover the target (pitch often <0.5), the treatment planning system (TPS) potentially employs a very large number of static beam directions and leaf opening configurations to model the modulated fields. All dose calculations performed by the system employ a convolution/superposition model. In this work the authors perform a full Monte Carlo (MC) dose calculation of tomotherapy deliveries to phantom computed tomography (CT) data sets to verify the TPS calculations. All MC calculations are performed with the EGSnrc-based MC simulation codes, BEAMnrc and DOSXYZnrc. Simulations are performed by taking the sinogram (leaf opening versus time) of the treatment plan and decomposing it into 51 different projections per rotation, as does the TPS, each of which is segmented further into multiple MLC opening configurations, each with different weights that correspond to leaf opening times. Then the projection is simulated by the summing of all of the opening configurations, and the overall rotational treatment is simulated by the summing of all of the projection simulations. Commissioning of the source model was verified by comparing measured and simulated values for the percent depth dose and beam profiles shapes for various jaw settings. The accuracy of the MLC leaf width and tongue and groove spacing were verified by comparing measured and simulated values for the MLC leakage and a picket fence pattern. The validated source and MLC configuration were then used to simulate a complex modulated delivery from fixed gantry angle. Further, a preliminary rotational treatment plan to a delivery quality assurance phantom (the "cheese" phantom) CT data set was simulated. Simulations were compared with measured results taken with an A1SL ionization chamber or EDR2 film measurements in a water tank or in a solid water phantom, respectively. The source and MLC MC simulations agree with the film measurements, with an acceptable number of pixels passing the 2%/1 mm gamma criterion. 99.8% of voxels of the MC calculation in the planning target volume (PTV) of the preliminary plan passed the 2%/2 mm gamma value test. 87.0% and 66.2% of the voxels in two organs at risk (OARs) passed the 2%/2 mm tests. For a 3%/3 mm criterion, the PTV and OARs show 100%, 93.2%, and 86.6% agreement, respectively. All voxels passed the gamma value test with a criterion of 5%/3 mm. The Tomo-Therapy TPS showed comparable results.     

Comprehensive evaluation of a commercial macro Monte Carlo electron dose calculation implementation using a standard verification data set

A commercial electron dose calculation software implementation based on the macro Monte Carlo algorithm has recently been introduced. We have evaluated the performance of the system using a standard verification data set comprised of two-dimensional (2D) dose distributions in the transverse plane of a 15 X 15 cm2 field. The standard data set was comprised of measurements performed for combinations of 9-MeV and 20-MeV beam energies and five phantom geometries. The phantom geometries included bone and air heterogeneities, and irregular surface contours. The standard verification data included a subset of the data needed to commission the dose calculation. Additional required data were obtained from a dosimetrically equivalent machine. In addition, we performed 2D dose measurements in a water phantom for the standard field sizes, a 4 cm X 4 cm field, a 3 cm diameter circle, and a 5 cm X 13 cm triangle for the 6-, 9-, 12-, 15-, and 18-MeV energies of a Clinac 21EX. Output factors were also measured. Synthetic CT images and structure contours duplicating the measurement configurations were generated and transferred to the treatment planning system. Calculations for the standard verification data set were performed over the range of each of the algorithm parameters: statistical precision, grid-spacing, and smoothing. Dose difference and distance-to-agreement were computed for the calculation points. We found that the best results were obtained for the highest statistical precision, for the smallest grid spacing, and for smoothed dose distributions. Calculations for the 21EX data were performed using parameters that the evaluation of the standard verification data suggested would produce clinically acceptable results. The dose difference and distance-to-agreement were similar to that observed for the standard verification data set except for the portion of the triangle field narrower than 3 cm for the 6- and 9-MeV electron beams. The output agreed with measurements to within 2%, with the exception of the 3-cm diameter circle and the triangle for 6 MeV, which were within 5%. We conclude that clinically acceptable results may be obtained using a grid spacing that is no larger than approximately one-tenth of the distal falloff distance of the electron depth dose curve (depth from 80% to 20% of the maximum dose) and small relative to the size of heterogeneities. For judicious choices of parameters, dose calculations agree with measurements to better than 3% dose difference and 3-mm distance-to-agreement for fields with dimensions no less than about 3 cm.     

A comparison of Monte Carlo dose calculation denoising techniques

Recent studies have demonstrated that Monte Carlo (MC) denoising techniques can reduce MC radiotherapy dose computation time significantly by preferentially eliminating statistical fluctuations ('noise') through smoothing. In this study, we compare new and previously published approaches to MC denoising, including 3D wavelet threshold denoising with sub-band adaptive thresholding, content adaptive mean-median-hybrid (CAMH) filtering, locally adaptive Savitzky-Golay curve-fitting (LASG), anisotropic diffusion (AD) and an iterative reduction of noise (IRON) method formulated as an optimization problem. Several challenging phantom and computed-tomography-based MC dose distributions with varying levels of noise formed the test set. Denoising effectiveness was measured in three ways: by improvements in the mean-square-error (MSE) with respect to a reference (low noise) dose distribution; by the maximum difference from the reference distribution and by the 'Van Dyk' pass/fail criteria of either adequate agreement with the reference image in low-gradient regions (within 2% in our case) or, in high-gradient regions, a distance-to-agreement-within-2% of less than 2 mm. Results varied significantly based on the dose test case: greater reductions in MSE were observed for the relatively smoother phantom-based dose distribution (up to a factor of 16 for the LASG algorithm); smaller reductions were seen for an intensity modulated radiation therapy (IMRT) head and neck case (typically, factors of 2-4). Although several algorithms reduced statistical noise for all test geometries, the LASG method had the best MSE reduction for three of the four test geometries, and performed the best for the Van Dyk criteria. However, the wavelet thresholding method performed better for the head and neck IMRT geometry and also decreased the maximum error more effectively than LASG. In almost all cases, the evaluated methods provided acceleration of MC results towards statistically more accurate results.     

IVBTMC,a Monte Carlo dose calculation tool for intravascular brachytherapy

A new Monte Carlo code (IVBTMC) is developed for accurate dose calculations in intravascular brachytherapy (IVBT). IVBTMC calculates the dose distribution of a brachytherapy source with arbitrary size and curvature in a general three-dimensional heterogeneous medium. Both beta and gamma sources are considered. IVBTMC is based on a modified version of the EGSNRC code. A voxel-based geometry is used to describe the target medium incorporating heterogeneities with arbitrary composition and shape. The source term is modeled using appropriate phase-space data. The phase-space data are calculated for three widely used sources (32P, 90Sr/90Y, and 192Ir). To speed up dose calculations for gamma sources, a special version of IVBTMC based on the kerma approximation is developed. The accuracy of the phase-space data model is verified and IVBTMC is validated against other Monte Carlo codes and against reported measurements using radio-chromic films. To illustrate the IVBTMC capabilities, a variety of examples are treated. 32P, 90Sr/90Y, and 192Ir sources with different lengths and degrees of curvature are considered. Calcified plaques with regular and irregular shapes are modeled. The dose distributions are calculated with a spatial resolution ranging between 0.1 and 0.5 mm. They are presented in terms of isodose contour plots. The dosimetric effects of the source curvature and/or the presence of calcified plaques are discussed. In conclusion, IVBTMC has the capability to perform high-precision IVBT dose calculations taking into account the realistic configurations of both the source and the target medium.     

Head-and-neck IMRT treatments assessed with a Monte Carlo dose calculation engine

IMRT is frequently used in the head-and-neck region, which contains materials of widely differing densities (soft tissue, bone, air-cavities). Conventional methods of dose computation for these complex, inhomogeneous IMRT cases involve significant approximations. In the present work, a methodology for the development, commissioning and implementation of a Monte Carlo (MC) dose calculation engine for intensity modulated radiotherapy (MC-IMRT) is proposed which can be used by radiotherapy centres interested in developing MC-IMRT capabilities for research or clinical evaluations. The method proposes three levels for developing, commissioning and maintaining a MC-IMRT dose calculation engine: (a) development of a MC model of the linear accelerator, (b) validation of MC model for IMRT and (c) periodic quality assurance (QA) of the MC-IMRT system. The first step, level (a), in developing an MC-IMRT system is to build a model of the linac that correctly predicts standard open field measurements for percentage depth-dose and off-axis ratios. Validation of MC-IMRT, level (b), can be performed in a rando phantom and in a homogeneous water equivalent phantom. Ultimately, periodic quality assurance of the MC-IMRT system is needed to verify the MC-IMRT dose calculation system, level (c). Once the MC-IMRT dose calculation system is commissioned it can be applied to more complex clinical IMRT treatments. The MC-IMRT system implemented at the Royal Marsden Hospital was used for IMRT calculations for a patient undergoing treatment for primary disease with nodal involvement in the head-and-neck region (primary treated to 65 Gy and nodes to 54 Gy), while sparing the spinal cord, brain stem and parotid glands. Preliminary MC results predict a decrease of approximately 1-2 Gy in the median dose of both the primary tumour and nodal volumes (compared with both pencil beam and collapsed cone). This is possibly due to the large air-cavity (the larynx of the patient) situated in the centre of the primary PTV and the approximations present in the dose calculation.     

GPU-based fast Monte Carlo simulation for radiotherapy dose calculation

Monte Carlo (MC) simulation is commonly considered to be the most accurate dose calculation method in radiotherapy. However, its efficiency still requires improvement for many routine clinical applications. In this paper, we present our recent progress toward the development of a graphics processing unit (GPU)-based MC dose calculation package, gDPM v2.0. It utilizes the parallel computation ability of a GPU to achieve high efficiency, while maintaining the same particle transport physics as in the original dose planning method (DPM) code and hence the same level of simulation accuracy. In GPU computing, divergence of execution paths between threads can considerably reduce the efficiency. Since photons and electrons undergo different physics and hence attain different execution paths, we use a simulation scheme where photon transport and electron transport are separated to partially relieve the thread divergence issue. A high-performance random number generator and a hardware linear interpolation are also utilized. We have also developed various components to handle the fluence map and linac geometry, so that gDPM can be used to compute dose distributions for realistic IMRT or VMAT treatment plans. Our gDPM package is tested for its accuracy and efficiency in both phantoms and realistic patient cases. In all cases, the average relative uncertainties are less than 1%. A statistical t-test is performed and the dose difference between the CPU and the GPU results is not found to be statistically significant in over 96% of the high dose region and over 97% of the entire region. Speed-up factors of 69.1 ～ 87.2 have been observed using an NVIDIA Tesla C2050 GPU card against a 2.27 GHz Intel Xeon CPU processor. For realistic IMRT and VMAT plans, MC dose calculation can be completed with less than 1% standard deviation in 36.1 ～ 39.6 s using gDPM.     

A Monte Carlo dose calculation algorithm for proton therapy

A Monte Carlo (MC) code (VMCpro) for treatment planning in proton beam therapy of cancer is introduced. It is based on ideas of the Voxel Monte Carlo algorithm for photons and electrons and is applicable to human tissue for clinical proton energies. In the present paper the implementation of electromagnetic and nuclear interactions is described. They are modeled by a Class II condensed history algorithm with continuous energy loss, ionization, multiple scattering, range straggling, delta-electron transport, nuclear elastic proton nucleus scattering and inelastic proton nucleus reactions. VMCpro is faster than the general purpose MC codes FLUKA by a factor of 13 and GEANT4 by a factor of 35 for simulations in a phantom with inhomogeneities. For dose calculations in patients the speed improvement is larger, because VMCpro has only a weak dependency on the heterogeneity of the calculation grid. Dose distributions produced with VMCpro are in agreement with GEANT4 results. Integrated or broad beam depth dose curves show maximum deviations not larger than 1% or 0.5 mm in regions with large dose gradients for the examples presented here.     

Radiation dose to the breast during mammography: a comprehensive, realistic Monte Carlo calculation

Certain assumptions have been made regarding the composition of the breast in different age groups: these are believed to be more realistic than existing assumptions and have been used in comprehensive Monte Carlo calculations of radiation dose to the breast and of appropriate indicators of risk as follows: (i) Mean dose to the sensitive tissues in an average breast. This may be used to compare risk in different mammographic techniques, and has been calculated for a wide variety of techniques, including some not previously explicitly studied. (ii) Integral dose to the sensitive tissues in a breast, which is a good indicator of risk in an individual case, has been calculated for two commonly used mammographic techniques and for four different breast compositions and thicknesses. The results have been compared with results obtained using other approaches.     

Monte Carlo modeling of radiation dose distributions in intravascular radiation therapy

Radiation dose distributions are developed for balloon and wire sources of radioactivity within coronary arteries. The Monte Carlo codes MCNP 4B and EGS4 were used to calculate dose distributions for photons and electrons at discrete energies around such sources, with and without the presence of a high-density atherosclerotic plaque. An interactive computer program was developed which then calculates dose distributions for many radionuclides by applying the emission spectra to the discrete energy grids calculated by the Monte Carlo codes, weighting appropriately for electron energy and abundance. Results for Re-186 and Re-188 balloon sources are shown in comparison to an Ir-192 wire source. The program provides dose distributions as well as estimates of activity levels needed to deliver prescribed doses to the vessel wall at selected distances from the lumen in a selected time interval. In addition, dose calculations are presented in this paper for other organs in the body, from photon radiation as well as from possible loss of liquid activity into the bloodstream in the case of a balloon rupture. These results, especially the interactive computer program permitting easy comparison of various radionuclides and their physical characteristics, will greatly facilitate the comparison process and aid in the selection of the best candidate(s) for clinical use.     

Comprehensive Monte Carlo calculation of the point spread function for a commercial a-Si EPID

Images produced by commercial amorphous silicon electronic portal imaging devices (a-Si EPIDs) are subject to multiple blurring processes. Implementation of these devices for fluence measurement requires that the blur be removed from the images. A standard deconvolution operation can be performed to accomplish this assuming the blur kernel is spatially invariant and accurately known. This study determines a comprehensive blur kernel for the Varian aS500 EPID. Monte Carlo techniques are used to derive a dose kernel and an optical kernel, which are then combined to yield an overall blur kernel for both 6 and 15 MV photon beams. Experimental measurement of the line spread function (LSF) is used to verify kernel shape. Kernel performance is gauged by comparing EPID image profiles with in-air dose profiles measured using a diamond detector (approximating fluence) both before and after the EPID images have been deconvolved. Quantitative comparisons are performed using the chi metric, an extension of the well-known y metric, using acceptance criteria of 0.0784 cm (1 pixel width) distance-to-agreement (deltad) and 2% of the relative central axis fluence (deltaD). Without incorporating any free parameters, acceptance was increased from 49.0% of pixels in a cross-plane profile for a 6 MV 10 x 10 cm2 open field to 92.0%. For a 10 x 10 cm2 physically wedged field, acceptance increased from 40.3% to 73.9%. The effect of the optical kernel was found to be negligible for these chi acceptance parameters, however for (deltaD= 1%, deltad = 0.0784 cm) we observed an improvement from 66.1% (without) to 78.6% (with) of chi scores <1 (from 20.6% before deconvolution). It is demonstrated that an empirical kernel having a triple exponential form or a semiempirical kernel based on a simplified model of the detector stack can match the performance of the comprehensive kernel.     

Validation of a Monte Carlo dose calculation tool for radiotherapy treatment planning

A new EGS4/PRESTA Monte Carlo user code, MCDOSE, has been developed as a routine dose calculation tool for radiotherapy treatment planning. It is suitable for both conventional and intensity modulated radiation therapy. Two important features of MCDOSE are the inclusion of beam modifiers in the patient simulation and the implementation of several variance reduction techniques. Before this tool can be used reliably for clinical dose calculation, it must be properly validated. The validation for beam modifiers has been performed by comparing the dose distributions calculated by MCDOSE and the well-benchmarked EGS4 user codes BEAM and DOSXYZ. Various beam modifiers were simulated. Good agreement in the dose distributions was observed. The differences in electron cutout factors between the results of MCDOSE and measurements were within 2%. The accuracy of MCDOSE with various variance reduction techniques was tested by comparing the dose distributions in different inhomogeneous phantoms with those calculated by DOSXYZ without variance reduction. The agreement was within 1.0%. Our results demonstrate that MCDOSE is accurate and efficient for routine dose calculation in radiotherapy treatment planning, with or without beam modifiers.     

Experimental investigation of a fast Monte Carlo photon beam dose calculation algorithm

An experimental verification of the recently developed XVMC code, a fast Monte Carlo algorithm to calculate dose distributions of photon beams in treatment planning, is presented. The treatment head is modelled by a point source with energy distribution (primary photons) and an additional head scatter contribution. Utility software is presented, allowing the determination of the parameters for this model using a single measured depth dose curve in water. The simple beam model is considered to be a starting point for more complex models being planned for future versions of the code. This paper is mainly focused on the influence of the different techniques on variance reduction and material property determination for dose distributions. It is demonstrated that XVMC and the simple beam model reproduce measured (by a diamond detector) relative dose distributions with an accuracy of better than +/-2% in various homogeneous and inhomogeneous phantoms. Furthermore, relative dose distributions in solid state phantoms have been measured by film. Also for these cases, measured and calculated dose distributions agree within experimental uncertainty. The short calculation time (depending on voxel resolution, statistical accuracy, field size and energy, a span of 1 min to 1 h using a present-day personal computer) and an interface to a commercial planning system will allow the implementation of the code for routine treatment planning of clinical electron and photon beams.     

MCPI: a sub-minute Monte Carlo dose calculation engine for prostate implants

An accelerated Monte Carlo code [Monte Carlo dose calculation for prostate implant (MCPI)] is developed for dose calculation in prostate brachytherapy. MCPI physically simulates a set of radioactive seeds with arbitrary positions and orientations, merged in a three-dimensional (3D) heterogeneous phantom representing the prostate and surrounding tissue. MCPI uses a phase space data source-model to account for seed self-absorption and seed anisotropy. A "hybrid geometry" model (full 3D seed geometry merged in 3D mesh of voxels) is used for rigorous treatment of the interseed attenuation and tissue heterogeneity effects. MCPI is benchmarked against the MCNP5 code for idealized and real implants, for 103Pd and 125I seeds. MCPI calculates the dose distribution (2-mm voxel mesh) of a 103Pd implant (83 seeds) with 2% average statistical uncertainty in 59 s using a single Pentium 4 PC (2.4 GHz). MCPI is more than 10(3) and 10(4) times faster than MCNP5 for prostate dose calculations using 2- and 1-mm voxels, respectively. To illustrate its usefulness, MCPI is used to quantify the dosimetric effects of interseed attenuation, tissue composition, and tissue calcifications. Ignoring the interseed attenuation effect or slightly varying the prostate tissue composition may lead to 6% decreases of D100, the dose delivered to 100% of the prostate. The presence of calcifications, covering 1%-5% of the prostate volume, decreases D80, D90, and D100 by up to 32%, 37%, and 58%, respectively. In conclusion, sub-minute dose calculations, taking into account all dosimetric effects, are now possible for more accurate dose planning and dose assessment in prostate brachytherapy.     

A Monte Carlo dose calculation tool for radiotherapy treatment planning

A Monte Carlo user code, MCDOSE, has been developed for radiotherapy treatment planning (RTP) dose calculations. MCDOSE is designed as a dose calculation module suitable for adaptation to host RTP systems. MCDOSE can be used for both conventional photon/electron beam calculation and intensity modulated radiotherapy (IMRT) treatment planning. MCDOSE uses a multiple-source model to reconstruct the treatment beam phase space. Based on Monte Carlo simulated or measured beam data acquired during commissioning, source-model parameters are adjusted through an automated procedure. Beam modifiers such as jaws, physical and dynamic wedges, compensators, blocks, electron cut-outs and bolus are simulated by MCDOSE together with a 3D rectilinear patient geometry model built from CT data. Dose distributions calculated using MCDOSE agreed well with those calculated by the EGS4/DOSXYZ code using different beam set-ups and beam modifiers. Heterogeneity correction factors for layered-lung or layered-bone phantoms as calculated by both codes were consistent with measured data to within 1%. The effect of energy cut-offs for particle transport was investigated. Variance reduction techniques were implemented in MCDOSE to achieve a speedup factor of 10-30 compared to DOSXYZ.