Monte Carlo source model for photon beam radiotherapy: photon source characteristics

A phase space model of a Versa HD linear accelerator for application to Monte Carlo dose calculation in a real-time adaptive workflow

PURPOSE

This study aims to develop and validate a simple geometric model of the accelerator head, from which a particle phase space can be calculated for application to fast Monte Carlo dose calculation in real-time adaptive photon radiotherapy. With this objective in view, the study investigates whether the phase space model can facilitate dose calculations which are compatible with those of a commercial treatment planning system, for convenient interoperability.

MATERIALS AND METHODS

A dual-source model of the head of a Versa HD accelerator (Elekta AB, Stockholm, Sweden) was created. The model used parameters chosen to be compatible with those of 6-MV flattened and 6-MV flattening filter-free photon beams in the RayStation treatment planning system (RaySearch Laboratories, Stockholm, Sweden). The phase space model was used to calculate a photon phase space for several treatment plans, and the resulting phase space was applied to the Dose Planning Method (DPM) Monte Carlo dose calculation algorithm. Simple fields and intensity-modulated radiation therapy (IMRT) treatment plans for prostate and lung were calculated for benchmarking purposes and compared with the convolution-superposition dose calculation within RayStation.

RESULTS

For simple square fields in a water phantom, the calculated dose distribution agrees to within ±2% with that from the commercial treatment planning system, except in the buildup region, where the DPM code does not model the electron contamination. For IMRT plans of prostate and lung, agreements of ±2% and ±6%, respectively, are found, with slightly larger differences in the high dose gradients.

CONCLUSIONS

The phase space model presented allows convenient calculation of a phase space for application to Monte Carlo dose calculation, with straightforward translation of beam parameters from the RayStation beam model. This provides a basis on which to develop dose calculation in a real-time adaptive setting.

Development of a virtual source model for Monte Carlo-based independent dose calculation for varian linac

Monte Carlo (MC) independent dose calculations are often based on phase-space files (PSF), as they can accurately represent particle characteristics. PSF generally are large and create a bottleneck in computation time. In addition, the number of independent particles is limited by the PSF, preventing further reduction of statistical uncertainty. The purpose of this study is to develop and validate a virtual source model (VSM) to address these limitations. Particles from existing PSF for the Varian TrueBeam medical linear accelerator 6X, 6XFFF, 10X, and 10XFFF beam configurations were tallied, analyzed, and used to generate a dual-source photon VSM that includes electron contamination. The particle density distribution, kinetic energy spectrum, particle direction, and the correlations between characteristics were computed. The VSM models for each beam configuration were validated with water phantom measurements as well as clinical test cases against the original PSF. The new VSM requires 67 MB of disk space for each beam configuration, compared to 50 GB for the PSF from which they are based and effectively remove the bottleneck set by the PSF. At 3% MC uncertainty, the VSM approach reduces the calculation time by a factor of 14 on our server. MC doses obtained using the VSM approach were compared against PSF-generated doses in clinical test cases and measurements in a water phantom using a gamma index analysis. For all tests, the VSMs were in excellent agreement with PSF doses and measurements (>90% passing voxels between doses and measurements). Results of this study indicate the successful derivation and implementation of a VSM model for Varian Linac that significantly saves computation time without sacrificing accuracy for independent dose calculation.

A structural analytic method on the phase space data of Linac 4 MV photons based on the real world

PURPOSE

It was given that the characteristics of the fluence distribution and the energy spectrum structure of 4MV photons on the Phase Space (PhSp) plane and a method to analyzing the characteristics.

METHODS

After the PhSp file of 4 MV photons was acquired by the method of Monte Carlo (MC) calculation, the photons recorded by PhSp file were grouped based on the energy bin, and it was analyzed that the spatial distribution and energy spectrum structure of the photons. The photons in each energy group were continually grouped to sub-files according to momentum bin, and the primary and scattered photons could be separated according to the character of the fluence distribution of the photons in the sub-files.

RESULTS

The energy of 4 MV beam is a continuous spectrum. The energy constituent on a pixel at different distances from the center point is different, and the average energy on the center axis of the field is the highest; The photons with 0-1.0 MeV had 42.6% of all; that with energy more than 3.0 MeV had 11.7%; greater than 4 MeV, just 1.5%. The primary and scattered photons were easy collected according to the distribution characteristics of sub-groups.

CONCLUSIONS

The work to acquire and analyze the PhSp file of the 4 MV beam is significant. 4 MV, 6 MV, 8 MV, 10 MV and 15 MV energy beams basically cover the beams of radiotherapy, and a database of the energy beams could be built for the MC related research of other scholars.

Monte carlo model and output factors of elekta infinity™ 6 and 10 MV photon beam

Aim

This study aimed to commission the Elekta Infinity™ working in 6 and 10 MV photon beam installed in Concord International Hospital, Singapore, and compare the OFs between MC simulation and measurement using PTW semiflex and microDiamond detector for small field sizes.

Material and Methods

There are two main steps in this study: modelling of Linac 6 and 10 MV photon beam and analysis of the output factors for field size 2 × 2-10 × 10 cm². The EGSnrc/BEAMnrc-DOSXYZnrc code was used to model and characterize the Linac and to calculate the dose distributions in a water phantom. The dose distribution and OFs were compared to the measurement data in the same condition.

Results

The commissioning process was only conducted for a 10 × 10 cm² field size. The PDD obtained from MC simulation showed a good agreement with the measurement. The local dose difference of PDDs was less than 2% for 6 and 10 MV. The initial electron energy was 5.2 and 9.4 MeV for 6 and 10 MV photon beam, respectively. This Linac model can be used for dose calculation in other situations and different field sizes because this Linac has been commissioned and validated using Monte Carlo simulation. The 10 MV Linac produces higher electron contamination than that of 6 MV.

Conclusions

The Linac model in this study was acceptable. The most important result in this work comes from OFs resulted from MC calculation. This value was more significant than the OFs from measurement using semiflex and microDiamond for all beam energy and field sizes because of the CPE phenomenon.

Beam modeling and beam model commissioning for Monte Carlo dose calculation-based radiation therapy treatment planning: Report of AAPM Task Group 157

Dose calculation plays an important role in the accuracy of radiotherapy treatment planning and beam delivery. The Monte Carlo (MC) method is capable of achieving the highest accuracy in radiotherapy dose calculation and has been implemented in many commercial systems for radiotherapy treatment planning. The objective of this task group was to assist clinical physicists with the potentially complex task of acceptance testing and commissioning MC-based treatment planning systems (TPS) for photon and electron beam dose calculations. This report provides an overview on the general approach of clinical implementation and testing of MC-based TPS with a specific focus on models of clinical photon and electron beams. Different types of beam models are described including those that utilize MC simulation of the treatment head and those that rely on analytical methods and measurements. The trade-off between accuracy and efficiency in the various source-modeling approaches is discussed together with guidelines for acceptance testing of MC-based TPS from the clinical standpoint. Specific recommendations are given on methods and practical procedures to commission clinical beam models for MC-based TPS.

Hybrid Monte Carlo source model: Advantages and deficiencies

Monte Carlo (MC) simulation is the gold standard for dose calculation. An accurate mathematical source model can be used for the radiation beams. Source models can consist of sub-sources or fewer sources with data that need to be measured. This can speed up treatment plan verification without the need for a full simulation of the radiation treatment machine.

Aims: This study aimed to construct a novel hybrid source model for 6 MV photon beams for an Elekta Synergy accelerator and to commission it against measured beam data and treatments plans.

Methods and Material: The model comprised of a circular photon and planar electron contamination source. The modified Schiff formula provided off-axis variable bremsstrahlung spectra. Collimation and scatter were modelled with error functions. An exponential function modelled the transmitted fluence through the collimators. The source model was commissioned by comparing simulated and measured MC data. Dose data included profiles, depth dose and film measurements in a Rando phantom. Field sizes ranged from 1 × 1 cm2 to 40 × 40 cm2.

Results: Regular, wedged and asymmetrical fields could be modelled within 1.5% or 1.5 mm. More than 95% of all points lie within 3% or 3 mm for the multi-leaf collimators contours data. A gamma criterion of 3% or 3 mm was met for a complex treatment case.

Conclusions: The two sub-source model replicated clinical 6 MV Elekta Synergy photons beams and could calculate the dose accurately for conformal treatments in complex geometries such as a head-and-neck case.

A fast jaw-tracking model for VMAT and IMRT Monte Carlo simulations

Modern radiotherapy techniques involve routine use of volumetric arc therapy (VMAT) and intensity modulated radiotherapy (IMRT) with jaw‐tracking – dynamic motion of the secondary collimators (jaws) in tandem with multi‐leaf collimators (MLCs). These modalities require accurate dose calculations for the purposes of treatment planning and dose verification. Monte Carlo (MC) methods for radiotherapy dose calculation are widely accepted as capable of achieving high accuracy. This paper presents an efficiency‐enhancement method for secondary collimator modeling, presented in the context of a tool for MC‐based dose second checks. The model constitutes an accuracy trade‐off in the source model for the sake of efficiency enhancement, but maintains the advantages of MC transport in patient heterogeneities. The secondary collimator model is called Flat‐Absorbing‐Jaw‐Tracking (FAJT). Transmission through and scatter from the secondary collimators is neglected, and jaws are modeled as perfectly absorbing planes. To couple the motion of secondary collimators with MLCs for jaw‐tracking, the FAJT model was built into the VCU‐MLC model. Gamma‐index analysis of the dose distributions from FAJT against the full BEAMnrc MC simulations showed over 99% pass rate for a range of open fields, two clinical IMRT, and one VMAT treatment plan, for 2%/2 mm criteria above 10%. Using FAJT, the simulation speed of the secondary collimators for open fields increased by a factor of 237, 1489, and 1395 for 4 × 4, 10 × 10, and 30 × 30 cm², respectively. In general, clinically oriented simulation times are reduced from “hours” to “minutes” on identical hardware. Results for nine representative clinical cases (seven with jaw‐tracking) are presented. The average 2%/2 mm γ‐test success rate above the 80% isodose was 96.8% when tested against the EPIDose electronic portal image‐based dose reconstruction method and 97.3% against the Eclipse analytical anisotropic algorithm.

Evaluation of latent variances in Monte Carlo dose calculations with Varian TrueBeam photon phase-spaces used as a particle source

The purpose of this study was to evaluate the latent variance (LV) of Varian TrueBeam photon phase-space files (PSF) for open 10  ×  10 cm² and small stereotactic fields and estimate the number of phase spaces required to be summed up in order to maintain sub-percent LV in Monte Carlo (MC) dose calculations. BEAMnrc/DOSXYZnrc software was used to transport particles from Varian phase-space files (PSF_A) through the secondary collimators. Transported particles were scored into another phase-space located under the jaws (PSF_B), or transported further through the cone collimators and scored straight below, forming PSF_C. Phase-space files (PSF_B) were scored for 6 MV-FFF, 6 MV, 10 MV-FFF, 10 MV and 15 MV beams with 10  ×  10 cm² field size, and PSF_C were scored for 6 MV beam under circular cones of 0.13, 0.25, 0.35, and 1 cm diameter. Both PSF_B and PSF_C were transported into a water phantom with particle recycling number ranging from 10 to 1000. For 10  ×  10 cm² fields 0.5  ×  0.5  ×  0.5 cm³ voxels were used to score the dose, whereas the dose was scored in 0.1  ×  0.1  ×  0.5 cm³ voxels for beams collimated with small cones. In addition, for small 0.25 cm diameter cone-collimated 6 MV beam, phantom voxel size varied as 0.02  ×  0.02  ×  0.5 cm³, 0.05  ×  0.05  ×  0.5 cm³ and 0.1  ×  0.1  ×  0.5 cm³. Dose variances were scored in all cases and LV evaluated as per Sempau et al. For the 10  ×  10 cm² fields calculated LVs were greatest at the phantom surface and decreased with depth until they reached a plateau at 5 cm depth. LVs were found to be 0.54%, 0.96%, 0.35%, 0.69% and 0.57% for the 6 MV-FFF, 6 MV, 10 MV-FFF, 10 MV and 15 MV energies, respectively at the depth of 10 cm. For the 6 MV phase-space collimated with cones of 0.13, 0.25, 0.35, 1.0 cm diameter, the LVs calculated at 1.5 cm depth were 75.6%, 25.4%, 17.6% and 8.0% respectively. Calculated LV for the 0.25 cm cone-collimated 6 MV beam were 61.2%, 40.7%, 22.5% in 0.02  ×  0.02  ×  0.5 cm³, 0.05  ×  0.05  ×  0.5 cm³ and 0.1  ×  0.1  ×  0.5 cm³ voxels respectively. In order to achieve sub-percent LV in open 10  ×  10 cm² field MC simulations a single PSF can be used, whereas for small SRS fields (0.13-1.0 cm) more PSFs (66-8 PSFs) would have to be summed.

A single-source photon source model of a linear accelerator for Monte Carlo dose calculation

PURPOSE

To introduce a new method of deriving a virtual source model (VSM) of a linear accelerator photon beam from a phase space file (PSF) for Monte Carlo (MC) dose calculation.

MATERIALS AND METHODS

A PSF of a 6 MV photon beam was generated by simulating the interactions of primary electrons with the relevant geometries of a Synergy linear accelerator (Elekta AB, Stockholm, Sweden) and recording the particles that reach a plane 16 cm downstream the electron source. Probability distribution functions (PDFs) for particle positions and energies were derived from the analysis of the PSF. These PDFs were implemented in the VSM using inverse transform sampling. To model particle directions, the phase space plane was divided into a regular square grid. Each element of the grid corresponds to an area of 1 mm2 in the phase space plane. The average direction cosines, Pearson correlation coefficient (PCC) between photon energies and their direction cosines, as well as the PCC between the direction cosines were calculated for each grid element. Weighted polynomial surfaces were then fitted to these 2D data. The weights are used to correct for heteroscedasticity across the phase space bins. The directions of the particles created by the VSM were calculated from these fitted functions. The VSM was validated against the PSF by comparing the doses calculated by the two methods for different square field sizes. The comparisons were performed with profile and gamma analyses.

RESULTS

The doses calculated with the PSF and VSM agree to within 3% /1 mm (>95% pixel pass rate) for the evaluated fields.

CONCLUSION

A new method of deriving a virtual photon source model of a linear accelerator from a PSF file for MC dose calculation was developed. Validation results show that the doses calculated with the VSM and the PSF agree to within 3% /1 mm.

An analytic linear accelerator source model for GPU-based Monte Carlo dose calculations

Recently, there has been a lot of research interest in developing fast Monte Carlo (MC) dose calculation methods on graphics processing unit (GPU) platforms. A good linear accelerator (linac) source model is critical for both accuracy and efficiency considerations. In principle, an analytical source model should be more preferred for GPU-based MC dose engines than a phase-space file-based model, in that data loading and CPU-GPU data transfer can be avoided. In this paper, we presented an analytical field-independent source model specifically developed for GPU-based MC dose calculations, associated with a GPU-friendly sampling scheme. A key concept called phase-space-ring (PSR) was proposed. Each PSR contained a group of particles that were of the same type, close in energy and reside in a narrow ring on the phase-space plane located just above the upper jaws. The model parameterized the probability densities of particle location, direction and energy for each primary photon PSR, scattered photon PSR and electron PSR. Models of one 2D Gaussian distribution or multiple Gaussian components were employed to represent the particle direction distributions of these PSRs. A method was developed to analyze a reference phase-space file and derive corresponding model parameters. To efficiently use our model in MC dose calculations on GPU, we proposed a GPU-friendly sampling strategy, which ensured that the particles sampled and transported simultaneously are of the same type and close in energy to alleviate GPU thread divergences. To test the accuracy of our model, dose distributions of a set of open fields in a water phantom were calculated using our source model and compared to those calculated using the reference phase-space files. For the high dose gradient regions, the average distance-to-agreement (DTA) was within 1 mm and the maximum DTA within 2 mm. For relatively low dose gradient regions, the root-mean-square (RMS) dose difference was within 1.1% and the maximum dose difference within 1.7%. The maximum relative difference of output factors was within 0.5%. Over 98.5% passing rate was achieved in 3D gamma-index tests with 2%/2 mm criteria in both an IMRT prostate patient case and a head-and-neck case. These results demonstrated the efficacy of our model in terms of accurately representing a reference phase-space file. We have also tested the efficiency gain of our source model over our previously developed phase-space-let file source model. The overall efficiency of dose calculation was found to be improved by ~1.3-2.2 times in water and patient cases using our analytical model.

A virtual source model for Monte Carlo simulation of helical tomotherapy

The purpose of this study was to present a Monte Carlo (MC) simulation method based on a virtual source, jaw, and MLC model to calculate dose in patient for helical tomotherapy without the need of calculating phase‐space files (PSFs). Current studies on the tomotherapy MC simulation adopt a full MC model, which includes extensive modeling of radiation source, primary and secondary jaws, and multileaf collimator (MLC). In the full MC model, PSFs need to be created at different scoring planes to facilitate the patient dose calculations. In the present work, the virtual source model (VSM) we established was based on the gold standard beam data of a tomotherapy unit, which can be exported from the treatment planning station (TPS). The TPS‐generated sinograms were extracted from the archived patient XML (eXtensible Markup Language) files. The fluence map for the MC sampling was created by incorporating the percentage leaf open time (LOT) with leaf filter, jaw penumbra, and leaf latency contained from sinogram files. The VSM was validated for various geometry setups and clinical situations involving heterogeneous media and delivery quality assurance (DQA) cases. An agreement of <1% was obtained between the measured and simulated results for percent depth doses (PDDs) and open beam profiles for all three jaw settings in the VSM commissioning. The accuracy of the VSM leaf filter model was verified in comparing the measured and simulated results for a Picket Fence pattern. An agreement of <2% was achieved between the presented VSM and a published full MC model for heterogeneous phantoms. For complex clinical head and neck (HN) cases, the VSM‐based MC simulation of DQA plans agreed with the film measurement with 98% of planar dose pixels passing on the 2%/2 mm gamma criteria. For patient treatment plans, results showed comparable dose‐volume histograms (DVHs) for planning target volumes (PTVs) and organs at risk (OARs). Deviations observed in this study were consistent with literature. The VSM‐based MC simulation approach can be feasibly built from the gold standard beam model of a tomotherapy unit. The accuracy of the VSM was validated against measurements in homogeneous media, as well as published full MC model in heterogeneous media.

PACS numbers: 87.53.‐j, 87.55.K‐

A hybrid phase-space and histogram source model for GPU-based Monte Carlo radiotherapy dose calculation

In GPU-based Monte Carlo simulations for radiotherapy dose calculation, source modelling from a phase-space source can be an efficiency bottleneck. Previously, this has been addressed using phase-space-let (PSL) sources, which provided significant efficiency enhancement. We propose that additional speed-up can be achieved through the use of a hybrid primary photon point source model combined with a secondary PSL source. A novel phase-space derived and histogram-based implementation of this model has been integrated into gDPM v3.0. Additionally, a simple method for approximately deriving target photon source characteristics from a phase-space that does not contain inheritable particle history variables (LATCH) has been demonstrated to succeed in selecting over 99% of the true target photons with only ~0.3% contamination (for a Varian 21EX 18 MV machine). The hybrid source model was tested using an array of open fields for various Varian 21EX and TrueBeam energies, and all cases achieved greater than 97% chi-test agreement (the mean was 99%) above the 2% isodose with 1% / 1 mm criteria. The root mean square deviations (RMSDs) were less than 1%, with a mean of 0.5%, and the source generation time was 4-5 times faster. A seven-field intensity modulated radiation therapy patient treatment achieved 95% chi-test agreement above the 10% isodose with 1% / 1 mm criteria, 99.8% for 2% / 2 mm, a RMSD of 0.8%, and source generation speed-up factor of 2.5.

Automatic commissioning of a GPU-based Monte Carlo radiation dose calculation code for photon radiotherapy

Monte Carlo (MC) simulation is commonly considered as the most accurate method for radiation dose calculations. Commissioning of a beam model in the MC code against a clinical linear accelerator beam is of crucial importance for its clinical implementation. In this paper, we propose an automatic commissioning method for our GPU-based MC dose engine, gDPM. gDPM utilizes a beam model based on a concept of phase-space-let (PSL). A PSL contains a group of particles that are of the same type and close in space and energy. A set of generic PSLs was generated by splitting a reference phase-space file. Each PSL was associated with a weighting factor, and in dose calculations the particle carried a weight corresponding to the PSL where it was from. Dose for each PSL in water was pre-computed, and hence the dose in water for a whole beam under a given set of PSL weighting factors was the weighted sum of the PSL doses. At the commissioning stage, an optimization problem was solved to adjust the PSL weights in order to minimize the difference between the calculated dose and measured one. Symmetry and smoothness regularizations were utilized to uniquely determine the solution. An augmented Lagrangian method was employed to solve the optimization problem. To validate our method, a phase-space file of a Varian TrueBeam 6 MV beam was used to generate the PSLs for 6 MV beams. In a simulation study, we commissioned a Siemens 6 MV beam on which a set of field-dependent phase-space files was available. The dose data of this desired beam for different open fields and a small off-axis open field were obtained by calculating doses using these phase-space files. The 3D γ-index test passing rate within the regions with dose above 10% of dmax dose for those open fields tested was improved averagely from 70.56 to 99.36% for 2%/2 mm criteria and from 32.22 to 89.65% for 1%/1 mm criteria. We also tested our commissioning method on a six-field head-and-neck cancer IMRT plan. The passing rate of the γ-index test within the 10% isodose line of the prescription dose was improved from 92.73 to 99.70% and from 82.16 to 96.73% for 2%/2 mm and 1%/1 mm criteria, respectively. Real clinical data measured from Varian, Siemens, and Elekta linear accelerators were also used to validate our commissioning method and a similar level of accuracy was achieved.

A comprehensive procedure for characterizing arbitrary azimuthally symmetric photon beams

PURPOSE

A new Monte Carlo (MC) source model (SM) has been developed for azimuthally symmetric photon beams.

METHODS

The MC simulation tallied phase space file (PSF) is divided into two categories depending on the relationship of the particle track line to the beam central axis: multiple point source (MPS) and spatial mesh based surface source (SMBSS). To validate this SM, MCNPX2.6 was used to generate two PSFs for a 6 MV photon beam from a Varian 2100C/D linear accelerator.

RESULTS

PDDs and profiles were calculated using the SM and original PSF for different field sizes from 5 × 5 to 40 × 40 cm2. Agreement was within 2% of the maximum dose at 100 cm SSD and 2.5% of the maximum dose at 200 cm SSD for beam profiles at depths of 3.5 cm and 15 cm with respect to the original PSF. Differences between the source model and the PSF in the out-of-field regions were less than 0.5% of the profile maximum value at 100 cm SSD. Differences between measured and calculated points were less than 2% of the maximum dose or 2 mm distance to agreement (DTA) at 100 cm SSD.

CONCLUSIONS

This SM is unique in that it accounts for a higher level of energy dependence on the particle's direction and it is independent from accelerator components, unlike other published SMs. The model can be applied to any arbitrary azimuthally symmetric beam and has source biasing capabilities that significantly increase the simulation speed up to 3300 for certain field sizes.

GPU-based Monte Carlo radiotherapy dose calculation using phase-space sources

A novel phase-space source implementation has been designed for graphics processing unit (GPU)-based Monte Carlo dose calculation engines. Short of full simulation of the linac head, using a phase-space source is the most accurate method to model a clinical radiation beam in dose calculations. However, in GPU-based Monte Carlo dose calculations where the computation efficiency is very high, the time required to read and process a large phase-space file becomes comparable to the particle transport time. Moreover, due to the parallelized nature of GPU hardware, it is essential to simultaneously transport particles of the same type and similar energies but separated spatially to yield a high efficiency. We present three methods for phase-space implementation that have been integrated into the most recent version of the GPU-based Monte Carlo radiotherapy dose calculation package gDPM v3.0. The first method is to sequentially read particles from a patient-dependent phase-space and sort them on-the-fly based on particle type and energy. The second method supplements this with a simple secondary collimator model and fluence map implementation so that patient-independent phase-space sources can be used. Finally, as the third method (called the phase-space-let, or PSL, method) we introduce a novel source implementation utilizing pre-processed patient-independent phase-spaces that are sorted by particle type, energy and position. Position bins located outside a rectangular region of interest enclosing the treatment field are ignored, substantially decreasing simulation time with little effect on the final dose distribution. The three methods were validated in absolute dose against BEAMnrc/DOSXYZnrc and compared using gamma-index tests (2%/2 mm above the 10% isodose). It was found that the PSL method has the optimal balance between accuracy and efficiency and thus is used as the default method in gDPM v3.0. Using the PSL method, open fields of 4 × 4, 10 × 10 and 30 × 30 cm(2) in water resulted in gamma passing rates of 99.96%, 99.92% and 98.66%, respectively. Relative output factors agreed within 1%. An intensity modulated radiation therapy patient plan using the PSL method resulted in a passing rate of 97%, and was calculated in 50 s (per GPU) compared to 8.4 h (per CPU) for BEAMnrc/DOSXYZnrc.

An integral conservative gridding--algorithm using Hermitian curve interpolation

The problem of re-sampling spatially distributed data organized into regular or irregular grids to finer or coarser resolution is a common task in data processing. This procedure is known as 'gridding' or 're-binning'. Depending on the quantity the data represents, the gridding-algorithm has to meet different requirements. For example, histogrammed physical quantities such as mass or energy have to be re-binned in order to conserve the overall integral. Moreover, if the quantity is positive definite, negative sampling values should be avoided. The gridding process requires a re-distribution of the original data set to a user-requested grid according to a distribution function. The distribution function can be determined on the basis of the given data by interpolation methods. In general, accurate interpolation with respect to multiple boundary conditions of heavily fluctuating data requires polynomial interpolation functions of second or even higher order. However, this may result in unrealistic deviations (overshoots or undershoots) of the interpolation function from the data. Accordingly, the re-sampled data may overestimate or underestimate the given data by a significant amount. The gridding-algorithm presented in this work was developed in order to overcome these problems. Instead of a straightforward interpolation of the given data using high-order polynomials, a parametrized Hermitian interpolation curve was used to approximate the integrated data set. A single parameter is determined by which the user can control the behavior of the interpolation function, i.e. the amount of overshoot and undershoot. Furthermore, it is shown how the algorithm can be extended to multidimensional grids. The algorithm was compared to commonly used gridding-algorithms using linear and cubic interpolation functions. It is shown that such interpolation functions may overestimate or underestimate the source data by about 10-20%, while the new algorithm can be tuned to significantly reduce these interpolation errors. The accuracy of the new algorithm was tested on a series of x-ray CT-images (head and neck, lung, pelvis). The new algorithm significantly improves the accuracy of the sampled images in terms of the mean square error and a quality index introduced by Wang and Bovik (2002 IEEE Signal Process. Lett. 9 81-4).

VMC++ versus BEAMnrc: a comparison of simulated linear accelerator heads for photon beams

BEAMnrc, a code for simulating medical linear accelerators based on EGSnrc, has been bench-marked and used extensively in the scientific literature and is therefore often considered to be the gold standard for Monte Carlo simulations for radiotherapy applications. However, its long computation times make it too slow for the clinical routine and often even for research purposes without a large investment in computing resources. VMC++ is a much faster code thanks to the intensive use of variance reduction techniques and a much faster implementation of the condensed history technique for charged particle transport. A research version of this code is also capable of simulating the full head of linear accelerators operated in photon mode (excluding multileaf collimators, hard and dynamic wedges). In this work, a validation of the full head simulation at 6 and 18 MV is performed, simulating with VMC++ and BEAMnrc the addition of one head component at a time and comparing the resulting phase space files. For the comparison, photon and electron fluence, photon energy fluence, mean energy, and photon spectra are considered. The largest absolute differences are found in the energy fluences. For all the simulations of the different head components, a very good agreement (differences in energy fluences between VMC++ and BEAMnrc <1%) is obtained. Only a particular case at 6 MV shows a somewhat larger energy fluence difference of 1.4%. Dosimetrically, these phase space differences imply an agreement between both codes at the <1% level, making VMC++ head module suitable for full head simulations with considerable gain in efficiency and without loss of accuracy.

Total scatter factors of small beams: a multidetector and Monte Carlo study

The scope of this study was to estimate total scatter factors (S(c,p)) of the three smallest collimators of the Cyberknife radiosurgery system (5-10 mm in diameter), combining experimental measurements and Monte Carlo simulation. Two microchambers, a diode, and a diamond detector were used to collect experimental data. The treatment head and the detectors were simulated by means of a Monte Carlo code in order to calculate correction factors for the detectors and to estimate total scatter factors by means of a consistency check between measurement and simulation. Results for the three collimators were: S(c,p) (5 mm) = 0.677 +/- 0.004, S(c,p) (7.5 mm) = 0.820 +/- 0.008, S(c,p) (10 mm) = 0.871 +/- 0.008, all relative to the 60 mm collimator at 80 cm source-to-detector distance. The method also allows the full width at half maximum of the electron beam to be estimated; estimations made with different collimators and different detectors were in excellent agreement and gave a value of 2.1 mm. Correction factors to be applied to the detectors for the measurement of S(c,p) were consistent with a prevalence of volume effect for the microchambers and the diamond and a prevalence of scattering from high-Z material for the diode detector. The proposed method is more sensitive to small variations of the electron beam diameter with respect to the conventional method used to commission Monte Carlo codes, i.e., by comparison with measured percentage depth doses (PDD) and beam profiles. This is especially important for small fields (less than 10 mm diameter), for which measurements of PDD and profiles are strongly affected by the type of detector used. Moreover, this method should allow S(c,p) of Cyberknife systems different from the unit under investigation to be estimated without the need for further Monte Carlo calculation, provided that one of the microchambers or the diode detector of the type used in this study are employed. The results for the diamond are applicable only to the specific detector that was investigated due to excessive variability in manufacturing.

An efficient framework for photon Monte Carlo treatment planning

Currently photon Monte Carlo treatment planning (MCTP) for a patient stored in the patient database of a treatment planning system (TPS) can usually only be performed using a cumbersome multi-step procedure where many user interactions are needed. This means automation is needed for usage in clinical routine. In addition, because of the long computing time in MCTP, optimization of the MC calculations is essential. For these purposes a new graphical user interface (GUI)-based photon MC environment has been developed resulting in a very flexible framework. By this means appropriate MC transport methods are assigned to different geometric regions by still benefiting from the features included in the TPS. In order to provide a flexible MC environment, the MC particle transport has been divided into different parts: the source, beam modifiers and the patient. The source part includes the phase-space source, source models and full MC transport through the treatment head. The beam modifier part consists of one module for each beam modifier. To simulate the radiation transport through each individual beam modifier, one out of three full MC transport codes can be selected independently. Additionally, for each beam modifier a simple or an exact geometry can be chosen. Thereby, different complexity levels of radiation transport are applied during the simulation. For the patient dose calculation, two different MC codes are available. A special plug-in in Eclipse providing all necessary information by means of Dicom streams was used to start the developed MC GUI. The implementation of this framework separates the MC transport from the geometry and the modules pass the particles in memory; hence, no files are used as the interface. The implementation is realized for 6 and 15 MV beams of a Varian Clinac 2300 C/D. Several applications demonstrate the usefulness of the framework. Apart from applications dealing with the beam modifiers, two patient cases are shown. Thereby, comparisons are performed between MC calculated dose distributions and those calculated by a pencil beam or the AAA algorithm. Interfacing this flexible and efficient MC environment with Eclipse allows a widespread use for all kinds of investigations from timing and benchmarking studies to clinical patient studies. Additionally, it is possible to add modules keeping the system highly flexible and efficient.

Azimuthal particle redistribution for the reduction of latent phase-space variance in Monte Carlo simulations

It is well known that the use of a phase space in Monte Carlo simulation introduces a baseline level of variance that cannot be suppressed through the use of standard particle recycling techniques. This variance (termed latent phase-space variance by Sempau et al) can be a significant limiting factor in achieving accurate, low-uncertainty dose scoring results, especially near the surface of a phantom. A BEAMnrc component module (MCTWIST) has been developed to reduce the presence of latent variance in phase-space-based Monte Carlo simulations by implementing azimuthal particle redistribution (APR). For each recycled use of a phase-space particle a random rotation about the beam's central axis is applied, effectively utilizing cylindrical symmetry of the particle fluence and therefore providing a more accurate representation of the source. The MCTWIST module is unique in that no physical component is actually added to the accelerator geometry. Beam modifications are made by directly transforming particle characteristics outside of BEAMnrc/EGSnrc particle transport. Using MCTWIST, we have demonstrated a reduction in latent phase-space variance by more than a factor of 20, for a 10 x 10 cm(2) field, when compared to standard phase-space particle recycling techniques. The reduction in latent variance has enabled the achievement of dramatically smoother in-water dose profiles. This paper outlines the use of MCTWIST in Monte Carlo simulation and quantifies for the first time the latent variance reduction resulting from exploiting cylindrical phase-space symmetry.

Quantification of the impact of MLC modeling and tissue heterogeneities on dynamic IMRT dose calculations

This study quantifies the dose prediction errors (DPEs) in dynamic IMRT dose calculations resulting from (a) use of an intensity matrix to estimate the multi-leaf collimator (MLC) modulated photon fluence (DPE(IGfluence) instead of an explicit MLC particle transport, and (b) handling of tissue heterogeneities (DPE(hetero)) by superposition/convolution (SC) and pencil beam (PB) dose calculation algorithms. Monte Carlo (MC) computed doses are used as reference standards. Eighteen head-and-neck dynamic MLC IMRT treatment plans are investigated. DPEs are evaluated via comparing the dose received by 98% of the GTV (GTV D 98%), the CTV D 95%, the nodal D 90%, the cord and the brainstem D 02%, the parotid D 50%, the parotid mean dose (D (Mean)), and generalized equivalent uniform doses (gEUDs) for the above structures. For the MC-generated intensity grids, DPE(IGfluence) is within +/- 2.1% for all targets and critical structures. The SC algorithm DPE(hetero) is within +/- 3% for 98.3% of the indices tallied, and within +/- 3.4% for all of the tallied indices. The PB algorithm DPE(hetero) is within +/- 3% for 92% of the tallied indices. Statistical equivalence tests indicate that PB DPE(hetero) requires a +/- 3.6% interval to state equivalence with the MC standard, while the intervals are < 1.5% for SC DPE(hetero) and DPE(IGfluence). Overall, these results indicate that SC and MC IMRT dose calculations which use MC-derived intensity matrices for fluence prediction do not introduce significant dose errors compared with full Monte Carlo dose computations; however, PB algorithms may result in clinically significant dose deviations.

Determination of parameters for a multiple-source model of megavoltage photon beams using optimization methods

Accurate modelling of the radiation output of a medical linear accelerator is important for radiotherapy treatment planning. The major challenge is the adjustment of the model to a specific treatment unit. One approach is to use a multiple-source model containing a set of physical parameters. In this work, the parameters were derived from standard beam data measurements using optimization methods. The source model used includes sub-sources for bremsstrahlung radiation from the target, extra-focal photon radiation and electron contamination. The cost function includes a gamma error measure between measurements and current dose calculations. The procedure was applied to six beam data sets (6 MV to 23 MV) measured with accelerators from three vendors, but the results focus primarily on Varian accelerators. The obtained average gamma error (1%, 1 mm) between dose calculations and measurements used in optimization was smaller than 0.7 for each studied treatment beam and field size, and a minimum of 83% of measurement points passed the gamma < 1 criterion. For experiments made at different SSDs and for asymmetric fields, the average gamma errors were smaller than 1.1. For irregularly shaped MLC apertures, the differences in point doses were smaller than 1.0%. This work demonstrates that the source model parameters can be automatically derived from simple measurements using optimization methods. The developed procedure is applicable to a wide range of accelerators, and has an acceptable accuracy and processing time.

A Monte Carlo based three-dimensional dose reconstruction method derived from portal dose images

The verification of intensity-modulated radiation therapy (IMRT) is necessary for adequate quality control of the treatment. Pretreatment verification may trace the possible differences between the planned dose and the actual dose delivered to the patient. To estimate the impact of differences between planned and delivered photon beams, a three-dimensional (3-D) dose verification method has been developed that reconstructs the dose inside a phantom. The pretreatment procedure is based on portal dose images measured with an electronic portal imaging device (EPID) of the separate beams, without the phantom in the beam and a 3-D dose calculation engine based on the Monte Carlo calculation. Measured gray scale portal images are converted into portal dose images. From these images the lateral scattered dose in the EPID is subtracted and the image is converted into energy fluence. Subsequently, a phase-space distribution is sampled from the energy fluence and a 3-D dose calculation in a phantom is started based on a Monte Carlo dose engine. The reconstruction model is compared to film and ionization chamber measurements for various field sizes. The reconstruction algorithm is also tested for an IMRT plan using 10 MV photons delivered to a phantom and measured using films at several depths in the phantom. Depth dose curves for both 6 and 10 MV photons are reconstructed with a maximum error generally smaller than 1% at depths larger than the buildup region, and smaller than 2% for the off-axis profiles, excluding the penumbra region. The absolute dose values are reconstructed to within 1.5% for square field sizes ranging from 5 to 20 cm width. For the IMRT plan, the dose was reconstructed and compared to the dose distribution with film using the gamma evaluation, with a 3% and 3 mm criterion. 99% of the pixels inside the irradiated field had a gamma value smaller than one. The absolute dose at the isocenter agreed to within 1% with the dose measured with an ionization chamber. It can be concluded that our new dose reconstruction algorithm is able to reconstruct the 3-D dose distribution in phantoms with a high accuracy. This result is obtained by combining portal dose images measured prior to treatment with an accurate dose calculation engine.

Analytic IMRT dose calculations utilizing Monte Carlo to predict MLC fluence modulation

A hybrid dose-computation method is designed which accurately accounts for multileaf collimator (MLC)-induced intensity modulation in intensity modulated radiation therapy (IMRT) dose calculations. The method employs Monte Carlo (MC) modeling to determine the fluence modulation caused by the delivery of dynamic or multisegmental (step-and-shoot) MLC fields, and a conventional dose-computation algorithm to estimate the delivered dose to a phantom or a patient. Thus, it determines the IMRT fluence prediction accuracy achievable by analytic methods in the limit that the analytic method includes all details of the MLC leaf transport and scatter. The hybrid method is validated and benchmarked by comparison with in-phantom film dose measurements, as well as dose calculations from two in-house, and two commercial treatment planning system analytic fluence estimation methods. All computation methods utilize the same dose algorithm to calculate dose to a phantom, varying only in the estimation of the MLC modulation of the incident photon energy fluence. Gamma analysis, with respect to measured two-dimensional (2D) dose planes, is used to benchmark each algorithm's performance. The analyzed fields include static and dynamic test patterns, as well as fields from ten DMLC IMRT treatment plans (79 fields) and five SMLC treatment plans (29 fields). The test fields (fully closed MLC, picket fence, sliding windows of different size, and leaf-tip profiles) cover the extremes of MLC usage during IMRT, while the patient fields represent realistic clinical conditions. Of the methods tested, the hybrid method most accurately reproduces measurements. For the hybrid method, 79 of 79 DMLC field calculations have gamma < 1 (3%/3 mm) for more than 95% of the points (per field) while for SMLC fields, 27 of 29 pass the same criteria. The analytic energy fluence estimation methods show inferior pass rates, with 76 of 79 DMLC and 24 of 29 SMLC fields having more than 95% of the test points with gamma < or = 1 (3%/3 mm). Paired one-way ANOVA tests of the gamma analysis results found that the hybrid method better predicts measurements in terms of both the fraction of points with gamma < or = 1 and the average gamma for both 2%/2 mm and 3%/3 mm criteria. These results quantify the enhancement in accuracy in IMRT dose calculations when MC is used to model the MLC field modulation.

Photon-beam subsource sensitivity to the initial electron-beam parameters

One limitation to the widespread implementation of Monte Carlo (MC) patient dose-calculation algorithms for radiotherapy is the lack of a general and accurate source model of the accelerator radiation source. Our aim in this work is to investigate the sensitivity of the photon-beam subsource distributions in a MC source model (with target, primary collimator, and flattening filter photon subsources and an electron subsource) for 6- and 18-MV photon beams when the energy and radial distributions of initial electrons striking a linac target change. For this purpose, phase-space data (PSD) was calculated for various mean electron energies striking the target, various normally distributed electron energy spread, and various normally distributed electron radial intensity distributions. All PSD was analyzed in terms of energy, fluence, and energy fluence distributions, which were compared between the different parameter sets. The energy spread was found to have a negligible influence on the subsource distributions. The mean energy and radial intensity significantly changed the target subsource distribution shapes and intensities. For the primary collimator and flattening filter subsources, the distribution shapes of the fluence and energy fluence changed little for different mean electron energies striking the target, however, their relative intensity compared with the target subsource change, which can be accounted for by a scaling factor. This study indicates that adjustments to MC source models can likely be limited to adjusting the target subsource in conjunction with scaling the relative intensity and energy spectrum of the primary collimator, flattening filter, and electron subsources when the energy and radial distributions of the initial electron-beam change.