Monte Carlo Calculated Dose Distributions Research Articles

TU‐A‐BRC‐01: Monte Carlo Treatment Planning: Implementation of Clinical Systems

Monte Carlo (MC) based treatment planning systems are already commercially available for both photon and electron beams and more users are implementing them in a clinical setting. Therefore it is important that strategies and paradigms for clinical commissioning and implementation of such systems be formulated and discussed. The purpose of commissioning tests for MC based treatment planning systems is not only to evaluate the accuracy of the system, but also to define the optimum calculation parameters, such as number of histories, calculation voxel size, etc., which will be applied in clinical use of the system. This in turn requires that medical physicists responsible for clinical implementation of such systems are well educated in principles of Monte Carlo algorithms.We provide a brief review of AAPM Task Group Report No. 105 (Med. Phys. 34 (2007) 4818–4853), a document which outlines the important aspects of a MC‐based dose calculation algorithm, from the basic aspects of the use of the MC method for radiation transport to the application of this approach in routine clinical photon and electron beam treatment planning. We also describe possible clinical implementation issues, including dose‐to‐medium vs. dose‐to‐water differences and give comparison of computation times for typical patient anatomies and calculation parameters.Learning Objectives:1. To provide an educational review of the physics of the MC method.2. To discuss the factors associated with MC dose calculation within the patient‐specific geometry, such as statistical uncertainties, approximations of the underlying physics model, CT‐number to material density assignments, and reporting of dose‐to‐medium versus dose‐to‐water.3. To briefly review the vendor transport codes currently used for clinical treatment planning.4. To discuss the issues associated with experimental verification of MC algorithms.5. To briefly review the potential clinical implications of MC calculated dose distributions.6. To provide example timing comparisons of the major vendor MC codes in the clinical setting.

Parametrization and application of scatter kernels for modelling scanned proton beam collimator scatter dose

Collimators are routinely used in proton radiotherapy to laterally confine the field and improve the penumbra. Collimator scatter contributes up to 15% of the local dose and is therefore important to include in treatment planning dose calculation. We present a method for reconstruction of the collimator scatter phase space based on the parametrization of pre-calculated scatter kernels. Collimator scatter distributions, generated by the Monte Carlo (MC) package GEANT4.8.2, were scored differential in direction and energy. The distributions were then parametrized so as to enable a fast reconstruction by sampling. MC calculated dose distributions in water based on the parametrized phase space were compared to full MC simulations that included the collimator in the simulation geometry, as well as to experimental data. The experiments were performed at the scanned proton beam line at the The Svedberg Laboratory (TSL) in Uppsala, Sweden. Dose calculations using the parametrization of this work and the full MC for isolated typical cases of collimator scatter were compared by means of the gamma index. The result showed that in total 96.7% (99.3%) of the voxels fulfilled the gamma 2.0%/2.0 mm (3.0%/3.0 mm) criterion. The dose distribution for a collimated field was calculated based on the phase space created by the collimator scatter model incorporated into the generation of the phase space of a scanned proton beam. Comparing these dose distributions to full MC simulations, including particle transport in the MLC, yielded that in total for 18 different collimated fields, 99.1% of the voxels satisfied the gamma 1.0%/1.0 mm criterion and no voxel exceeded the gamma 2.6%/2.6 mm criterion. The dose contribution of collimator scatter along the central axis as predicted by the model showed good agreement with experimental data.

TH-D-AUD B-08: Initial Experience with the Monaco IMRT Treatment Planning System

Purpose: To evaluate the commercial CMS Monaco IMRT treatment planning system which employs a Monte Carlo (MC) based dose calculation engine, biological motivated cost functions, multi-criteria optimization, and an efficient sequencing algorithm. Method and Materials: For a head and neck, a liver, a prostate and a rectal cancer patient, step-and-shoot IMRT plans were designed using Monaco. The plans were compared to ones generated by the established CMS XiO treatment planning system. The plans were optimized to achieve the same clinical objectives concerning dose to the tumor and to the relevant organs-at-risk. However, whereas the XiO plans were formulated using DVH and minimum/maximum dose constraints, the Monaco plans utilized the biological cost functions offered by the system. DVHs, EUD, mean- and maximum-doses were compared, as well as the number of beam segments and MUs. Finally the plans were delivered on a MapCheck device to verify the agreement between the MC calculated dose distributions and measurements to be less than 3% and 3 mm. Results: Plans optimized with Monaco achieved at least similar and in some cases superior dose distributions. The multi-criteria optimization tools and the sensitivity analysis helped to reduce the time needed to optimize the plan. The Monaco plans resulted in fewer segments and lower number of MUs and therefore reduced delivery time. All calculated dose distributions passed the dose verification with the MapCheck device. Conclusion: The commercially available Monaco system produces clinical relevant plans, which are dosimetrically equivalent or superior to plans from the conventional XiO system, feature shorter delivery times, and can easily be verified with normal QA procedures using MapCheck.

MO‐A‐AUD A‐01: Review of AAPM Task Group Report No. 105

A special class of Monte Carlo (MC)‐based dose calculation codes optimized for photon and electron beams in patient‐specific geometries has invigorated interest in the use of MC‐based dose calculations for radiotherapy treatment planning. In general, this class of “second generation” codes, including VMC++, XVMC, and DPM, among others, employ electron‐step algorithms that converge faster, i.e. fewer condensed‐history steps are required for the same precision versus “first‐generation” codes (such as EGS and MCNP). These advances, coupled with the use of sophisticated variance reduction techniques (e.g. directional bremsstrahlung splitting), have made it possible to perform MC‐based photon beam dose calculations, in some instances, of the entire linear accelerator treatment head and patient‐geometry, within minutes on a single processor. Consequently, several commercial vendors have released or are currently in the process of releasing MC algorithms for photon and/or electron beam treatment planning. With the impending availability of MC‐based dose calculation algorithms for routine clinical treatment planning, it is important that strategies and paradigms for clinical commissioning and implementation of these systems be formulated and discussed. We provide a review of AAPM Task Group Report No. 105 (Med. Phys. 34 (2007) 4818–4853), a document which outlines the important aspects of a MC‐based dose calculation algorithm, from the basic aspects of the use of the MC method for radiation transport to the application of this approach in routine clinical photon and electron beam treatment planning.Educational Objectives:1. To provide an educational review of the physics of the MC method including discussion of the approaches used for coupled photon and electron transport.2. To review the methods used to improve the MC simulation efficiency.3. To briefly review the vendor transport codes currently used for clinical treatment planning.4. To describe the development of beam models for clinical treatment planning.5. To discuss the factors associated with MC dose calculation within the patient‐specific geometry, such as statistical uncertainties, CT‐number to material density assignments, and reporting of dose‐to‐medium versus dose‐to‐water.6. To discuss the issues associated with experimental verification of MC algorithms.7. To briefly review the potential clinical implications of MC calculated dose distributions.8. To provide example timing comparisons of the major vendor MC codes in the clinical setting.

Optimization of deterministic transport parameters for the calculation of the dose distribution around a high dose‐rate brachytherapy source

The goal of this work was to calculate the dose distribution around a high dose-rate 192Ir brachytherapy source using a multi-group discrete ordinates code and then to compare the results with a Monte Carlo calculated dose distribution. The unstructured tetrahedral mesh discrete ordinates code Attila version 6.1.1 was used to calculate the photon kerma rate distribution in water around the Nucletron microSelectron mHDRv2 source. MCNPX 2.5.c was used to compute the Monte Carlo water photon kerma rate distribution. Two hundred million histories were simulated, resulting in standard errors of the mean of less than 3% overall. The number of energy groups, S(n) (angular order), P(n) (scattering order), and mesh elements were varied in addition to the method of analytic ray tracing to assess their effects on the deterministic solution. Water photon kerma rate matrices were exported from both codes into an in-house data analysis software. This software quantified the percent dose difference distribution, the number of points within +/- 3% and +/- 5%, and the mean percent difference between the two codes. The data demonstrated that a 5 energy-group cross-section set calculated results to within 0.5% of a 15 group cross-section set. S12 was sufficient to resolve the solution in angle. P2 expansion of the scattering cross-section was necessary to compute accurate distributions. A computational mesh with 55 064 tetrahedral elements in a 30 cm diameter phantom resolved the solution spatially. An efficiency factor of 110 with the above parameters was realized in comparison to MC methods. The Attila code provided an accurate and efficient solution of the Boltzmann transport equation for the mHDRv2 source.

The influence of air cavities within the PTV on Monte Carlo-based IMRT optimization

Integrating Monte Carlo calculated dose distributions into an iterative aperture-based IMRT optimization process can improve the final treatment plan. However, the influence of large air cavities in the planning target volume (PTV) on the outcome of the optimization process should not be underestimated. To study this influence, the treatment plan of an ethmoid sinus cancer patient, which has large air cavities included in the PTV, is iteratively optimized in two different situations, namely when the large air cavities are included in the PTV and when these air cavities are excluded from the PTV. Two optimization methods were applied to integrate the Monte Carlo calculated dose distributions into the optimization process, namely the 'Correction–method' and the 'Per Segment–method'. The 'Correction–method' takes the Monte Carlo calculated global dose distribution into account in the optimization process by means of a correction matrix, which is in fact a dose distribution that is equal to the difference between the Monte Carlo calculated global dose distribution and the global dose distribution calculated by a conventional dose calculation algorithm. The 'Per Segment–method' uses directly the Monte Carlo calculated dose distributions of the individual segments in the optimization process. Both methods tend to converge whether or not large air cavities are excluded from the PTV during the optimization process. However, the 'Per Segment–method' performs better than the 'Correction–method' in both situations and the 'Per Segment–method' in the case where the large air cavities are excluded from the PTV leads to a better treatment plan then when these air cavities are included. Therefore we advise to exclude large air cavities and to apply the 'Per Segment–method' to integrate the Monte Carlo dose calculations into an iterative aperture-based optimization process. Nevertheless, the 'Correction–method' provides a good alternative in the case when the external dose engine is not able to generate individual dose distributions for the individual segments.

SU-FF-T-299: Monte Carlo Based Dose Verification for Serial Tomotherapy

Purpose: To develop a Monte Carlo model to verify the final dose distributions and monitor units for serial tomotherapy plans developed and delivered using the Peacock system (Corvus Treatment Planning System and MIMiC collimator, Nomos Corp., Sewickley, PA). Materials and methods: The Peacock system delivers the dose to the patient using arc therapy. The treatment plan is created in Corvus were sinograms are created for each arc in order to dictate the state of each of the MIMiC leaves at different locations along the arc. In‐house functions were written in Matlab (Mathworks Inc., Natick, MA) to decode these sinograms. A simple three‐field plan (three gantry positions), as well as full patient treatment plans were simulated using our Monte Carlo model and the same plans were delivered using the Peacock system in solid water. Films were placed in the solid water phantom in order to measure the dose distribution for comparison against the Monte Carlo calculations. Matlab functions were written to convert the Monte Carlo output into a format RIT113 (RIT Inc., Colorado Springs, CO) could read. This allowed us to co‐register the calculated dose maps and the measured ones in order to compare the two. Results: The Monte Carlo calculated dose distribution from the complete arc therapy in solid water phantom was compared against film measurements. The agreement was within 2%. The comparison between Monte Carlo results and Corvus calculated dose distribution revealed that Corvus would fail to accurately compute the dose in the region where inhomogeneities were present. Conclusions: Based on the agreement between Monte Carlo and measurements we can use the Monte Carlo system as an independent quality assurance tool in order to verify dose distributions and MUs per arc computed by the Corvus.

TU‐C‐AUD‐06: Calculation of the Dose Distribution Around a Commercially Available 125I Brachytherapy Source Via a Multi‐Group Discrete Ordinates Method

Purpose: To calculate the dose distribution around a commercially available source by a multi‐group discrete ordinates code and compare with a Monte Carlo calculated dose distribution. Methods and Materials: The multi‐group discrete ordinates code, Attila™ version 6.1.1 (Transpire Inc., Gig Harbor, WA) was used to calculate the dose distribution around the IsoAid IAI‐125 (IsoAid LLC., 7824 Clark Moody Blvd, Port Richey, Florida 34668). MCNPX 2.4.k was used to benchmark the deterministic calculations. The source was constructed with Solidworks (Solidworks Corp., Concord, MA), a mechanical design software. The constructed geometry of the source, dose scoring plane and sphere were exported in Parasolid® file format so that it could be imported into Attila. MCNPX 2.4.k was used to compute the Monte Carlo dose distribution. 295 million histories were simulated resulting in standard errors of the mean of approximately 3% at a point 5 cm from the source. The source geometry was identical to that of the Attila run. Dose rate matrices were exported from both codes and imported into an in‐house data analysis software. Results: The qualitative agreement between Attila and MCNPX is excellent. The Attila calculations required approximately 13 minutes of CPU versus 1701minutes for MCNPX. Conclusions: The Attila code accurately captured the anisotropy from the IsoAid IAI‐125 source. Attila can calculated dose in an efficient manner for low energy photon sources.

Sci‐Thur PM Therapy‐04: Comparison of treatment planning system and Monte Carlo calculated dose distributions in an extreme water‐lung interface phantom

The study compares the performance of the analytical anisotropic algorithm (AAA), a new superposition‐convolution algorithm recently implemented in Eclipse™ Integrated Treatment Planning System (TPS), to that of the pencil beam convolution (PBC) algorithm for a variety of clinical beam configurations at nominal beam energies of 6 MV and 18 MV, respectively in an extreme water‐lung interface phantom using Monte Carlo (MC) calculated dose distributions as benchmarks. Dose profiles at specific depths, 2‐D dose difference (TPS — MC) maps and histograms as well as first order statistics were used to quantify the accuracy of both photon dose calculation algorithms. For each open beam configuration, the AAA algorithm yielded smaller means, standard deviations and confidence limits. On average, the standard deviation of the dose differences was reduced by half with the AAA dose calculation model. Confidence limits for the PBC algorithm were ⩾ 3%, ranging from 3.0% for the 6 MV, 10×10 cm2 beam to 11.2% for the 18 MV, 4×4 cm2 beam while for the AAA algorithm they were ⩽ 4% for all beam configurations. The worst agreement of AAA algorithm with MC results was observed for the highest energy and smallest field combination (18 MV, 4×4 cm2). The standard deviation of the differences (2.4%) and confidence limit (4.0%) were however right between the standard deviations and confidence limits provided by the current clinical PBC model of the 6MV 4×4 and 10×10 cm2 beams.

WE‐C‐224C‐09: Calculation of the Dose Distribution Around a High Dose‐Rate 192Ir Brachytherapy Source Via a Multi‐Group Discrete Ordinates Method

Purpose: To calculate the dose distribution around a high dose‐rate (HDR) 192Ir brachytherapy source in water by a multi‐group discrete ordinates code and compare with a benchmarked Monte Carlo calculated dose distribution. Methods and Materials: The multi‐group discrete ordinates code, Attila™ version 5.0.2 (Transpire Inc., Gig Harbor, WA) was used to calculate the dose distribution around the Nucletron microSelectron HDR source (Nucletron B.V., Veenendaal, The Netherlands). MCNPX 2.4.k was used to benchmark the deterministic calculations. The source was constructed with Solidworks (Solidworks Corp., Concord, MA), a mechanical design software. The constructed geometry of the source, dose scoring plane and sphere were exported in Parasolid® file format so that it could be imported into Attila.MCNPX 2.4.k was used to compute the Monte Carlo dose distribution. 50 million histories were simulated resulting in standard errors of the mean of less than 5% at a point 7 cm from the center of the source. The source geometry was identical to that of the Attila run except for minor differences in modeling the source due to limitations of the MCNPX code.Dose rate matrices were exported from both codes and imported into an in‐house data analysis software. This software overlaid the matrices and quantified the percent dose difference and distance‐to‐agreement for all points in the matrices. The number of points passing a 2%/2mm criterion were reported. Results: Attila calculated dose to within ±2% or 2mm for 99% of the points. The Attila calculations required approximately 38 minutes of CPU versus 568 minutes for MCNPX on the same CPU. Conclusions: The Attila multi‐group discrete ordinates code accurately calculated the anisotropy from the microselectron HDR source. Attila accurately calculated dose in an efficient manner.

SU-FF-T-198: Dosimetry of Small Lung Lesions with EGSnrc Monte Carlo and Treatment Planning Systems

Background: The ability to deliver a homogenous dose distribution to small lung lesions is usually affected by the surrounding less dense normal lung. In low‐density tissue, there is increased transmission of photons relative to that in tissue. The lateral scatter of electrons out of the beam can lead to loss of field flatness and increased penumbral width. The magnitude of these effects is known to be very significant at high energies. Some commonly used radiotherapy treatment planning systems have had limited success in predicting accurately the dose distribution under these inhomogeneous conditions. Purpose: The purpose of this work is to study the magnitude of these effects using the EGSnrc/BEAMnrc Monte Carlo code on a typical patient CT dataset and to assess the limitations in treatment planning of such lungtumours with three Treatment Planning Systems (TPS); ADAC Pinnacle III v7.4, Varian Eclipse IMX and MDS Nordion Theraplan Plus v3.8. Methods: Small lungtumours (GTV) of diameter (3cm and 5cm) were simulated as water equivalent volumes and graphically inserted into the CT dataset at three different locations (Anterior, Central and Posterior) in the right lung. Using 6MV and 15MV beams, modelled for a Varian 21EX with the BEAMnrc Monte Carlo dose distributions were simulated using a Parallel Opposed Pairs (POP) technique. The field sizes were varied such that the distance from GTV to field edge was 1.0cm, 1.5cm and 2.0cm. Results and Discussion: Compared to the Monte Carlo calculated dose distribution, considerable and unique differences were observed in predicted dose by the TPS around the periphery of tumour and in the lung. These differences can be attributed to changes in the electron transport in lung, which are not adequately taken into account. The severity of these uncertainties, increases with photon energy, and decreases with field size for all planning systems.

Denoising of Monte Carlo dose calculations: Smoothing capabilities versus introduction of systematic bias

In order to evaluate the performance of denoising algorithms applied to Monte Carlo calculated dose distributions, conventional evaluation methods (rms difference, 1% and 2% difference) can be used. However, it is illustrated that these evaluation methods sometimes underestimate the introduction of bias, since possible bias effects are averaged out over the complete dose distribution. In the present work, a new evaluation method is introduced based on a sliding window superimposed on a difference dose distribution (reference dose-noisy/denoised dose). To illustrate its importance, a new denoising technique (ANRT) is presented based upon a combination of the principles of bilateral filtering and Savitzky-Golay filters. This technique is very conservative in order to limit the introduction of bias in high dose gradient regions. ANRT is compared with IRON for three challenging cases, namely an electron and photon beam impinging on heterogeneous phantoms and two IMRT treatment plans of head-and-neck cancer patients to determine the clinical relevance of the obtained results. For the electron beam case, IRON outperforms ANRT concerning the smoothing capabilities, while no differences in systematic bias are observed. However, for the photon beam case, although ANRT and IRON perform equally well on the conventional evaluation tests (rms difference, 1% and 2% difference), IRON clearly introduces much more bias in the penumbral regions while ANRT seems to introduce no bias at all. When applied to the IMRT patient cases, both denoising methods perform equally well regarding smoothing and bias introduction. This is probably caused by the summation of a large set of different beam segments, decreasing dose gradients compared to a single beam. A reduction in calculation time without introducing large systematic bias can shorten a Monte Carlo treatment planning process considerably and is therefore very useful for the initial trial and error phase of the treatment planning process.

Development of a laser-driven proton accelerator for cancer therapy

Recent advances in laser technology have made proton (light ion) acceleration possible using laser-induced plasmas. In this work, we report our work for the last few years on the investigation of a new proton therapy system for radiation oncology, which employs laser-accelerated protons. If successfully developed, the new system will be compact, cost-effective, and capable of delivering energy-and intensity-modulated proton therapy (EIMPT). We have focused our research on three major aspects: (1) target design for laser-proton acceleration, (2) system design for particle/energy selection and beam collimation, and (3) dosimetric studies on the use of laser-accelerated protons for cancer therapy. We have performed particle-in-cell (PIC) simulations to investigate optimal target configurations for proton/ion acceleration. We also performed Monte Carlo simulations to study the beam characteristics and the feasibility of using such beams for cancer treatment. Since laser-accelerated protons have broad energy and angular distributions, which are not suitable for radiotherapy applications directly, we have designed a compact particle selection and beam collimating system for EIMPT beam delivery. We also proposed a new gantry design to make the whole system compact to retrofit existing linac vaults. We have compared Monte Carlo calculated dose distributions using X-ray IMRT and laser-proton EIMPT. Our results show that EIMPT using laser protons produces superior target coverage and much reduced critical structure dose and integral dose compared to X-ray IMRT.

EUD-based radiotherapy treatment plan evaluation: incorporating physical and Monte Carlo statistical dose uncertainties

The purpose of this work is to quantify the impact of dose uncertainty on radiobiologically based treatment plan evaluation. Dose uncertainties are divided into two categories: physical and statistical. Physical dose uncertainty is associated with the systematic and/or random errors incurred during treatment planning and/or delivery. The dose uncertainty associated with Monte Carlo calculated dose distributions is deemed statistical and noted as artificial with respect to the actual delivered dose. We will refer to all dose uncertainties that arise from either calculation or delivery as stochastic. Both physical and statistical dose uncertainties are considered at the intra- and inter-voxel levels. To account for voxel dose uncertainty, we calculate the mean survival fraction (SF) for the random dose deposition. Mathematically, the expression for the mean survival fraction is identical to that used by Niemierko (1997 Med. Phys. 24 103–10) in defining equivalent uniform dose (EUD). To distinguish between spatial and probabilistic dose variations, we define equivalent stochastic dose (ESD) as a voxel dose that gives the mean expected survival fraction for the randomly deposited dose. For a probability density function f(D), that represents the probabilistic voxel dose, SF(ESD) can be calculated by convolving SF(D) with f(D). In the case where the probability density function follows a Gaussian distribution, an analytic expression is derived for SF(ESD). The derived expression is verified using the Monte Carlo method and ESD values calculated with varied radiosensitivities for cases of 60 and 70 Gy at 2 Gy per fraction. The analytic expression is also extended to account for a multi-voxel dose distribution that incorporates a spatial dose heterogeneity. The results show that survival fraction increases with an increased dose uncertainty. This reduction depends on radiobiological parameters attributed to tissue and tumour. For tissue, ESD drops to 55% of the mean physical dose when the dose has a 10% intra- and inter-voxel dose uncertainty and inhomogeneity.

SU-FF-J-67: Monte Carlo Dose Calculation in Prostate Patients Aided by 3D Ultrasound Imaging

Purpose: To investigate the use of Monte Carlo (MC)dose calculations for improved assessment of the daily treatment dose for prostate patients using novel 3D ultrasound imaging.Method and Materials: Prostate displacements were measured in 39 prostate cancer patients undergoing external‐beam radiotherapy with a 3D ultrasoundimage guided radiotherapy system (RESTITU, Resonant Medical Inc). The system resides in both the CT‐Sim and treatment rooms. It acquires 3D ultrasound(US) data of the pelvic region just prior to the CT‐sim, with the 3D USimage registered in absolute space. Subsequent 3D USimages are also acquired before treatment delivery, resulting in a set of 3D USimages of the time‐evolving geometry. The system also calculates prostate displacement between the CT‐Sim and treatment data sets. An algorithm was developed to combine CTimages in a MCdose calculation system with 3D US information such as organ shifts and contours. Dose distributions for both shifted and unshifted CT data were calculated using both a conventional planning system (CADPLAN) and a MCdose calculation system. Results:MC calculated dose distributions in prostate patients were found to differ significantly from conventional dose calculations using the CADPLAN system, mainly due to the influence of the femoral heads. Incorporation of the 3D US measured prostate shifts in the MCdose calculations clearly show severe tumour underdosage would result if no correction was applied to the patient setup. When correcting the patient setup, calculations taking the time‐evolving 3D US contours into account show that significant differences in the dose distributions in the target and surrounding tissue occur compared to a single planning dose calculation. Conclusion:Ultrasound information from a 3D US system opens up the possibility to obtain improved dose assessments for daily treatment fractions using MCdose calculation techniques, leading eventually to image‐guided adaptive radiotherapy.

SU-FF-T-313: Impact On Tumor Dose Coverage Due to the Second Buildup for Lung Tumour Treatment Using An 18 MV Photon Beam

Purpose: Conformal radiotherapy requires accurate dose prediction in the planning target volume and organs at risk. This study is to investigate the target dose coverage in lungtumor treatment and to assess the impact of the second dose buildup as a function of irradiated field sizes when an 18 MV photon beam is used. Method and Materials: The Monte Carlo codes BEAMnrc/DOSXYZnrc are used to simulate photon beams from a Varian accelerator and to calculate dose distributions in a lung phantom. The dose distributions in lung and tumor were calculated for various field sizes to investigate the effect of the incident field size. The Monte Carlo calculated dose distributions are also compared with those predicted by commercial treatment planning systems. Results: There is no significant tumordose reduction ( 10×10cm2 for an 18 MV photon beam treated with parallel opposed beams. However there are about 4% to 12% dose reductions at the interface due to the second buildup when treatment fields are 7×7cm2 to 3×3cm2 respectively. The simple inhomogeneity correction methods (Batho and ETAR) employed in a commercial treatment planning system can give rise to inaccuracies up to 20% in the lung and tumordose for a small field. Conclusion: There is a benefit to using a higher energy 18 MV beam for larger treatment fields (>8×8 cm2) to reduce hot spots and there is negligible dose reduction at the interface due to the second buildup. However the second buildup can cause significant dose reduction at the interface for field sizes <7×7 cm2 and an 18 MV beam should be not be used for a small fields in the lung. Batho or ETAR correction methods result in large errors in predicting the dose in lung and tumor for a small field.

Modelling of electron contamination in clinical photon beams for Monte Carlo dose calculation

The purpose of this work is to model electron contamination in clinical photon beams and to commission the source model using measured data for Monte Carlo treatment planning. In this work, a planar source is used to represent the contaminant electrons at a plane above the upper jaws. The source size depends on the dimensions of the field size at the isocentre. The energy spectra of the contaminant electrons are predetermined using Monte Carlo simulations for photon beams from different clinical accelerators. A ‘random creep’ method is employed to derive the weight of the electron contamination source by matching Monte Carlo calculated monoenergetic photon and electron percent depth–dose (PDD) curves with measured PDD curves. We have integrated this electron contamination source into a previously developed multiple source model and validated the model for photon beams from Siemens PRIMUS accelerators. The EGS4 based Monte Carlo user code BEAM and MCSIM were used for linac head simulation and dose calculation. The Monte Carlo calculated dose distributions were compared with measured data. Our results showed good agreement (less than 2% or 2 mm) for 6, 10 and 18 MV photon beams.

Monitor unit calculation for Monte Carlo treatment planning

In this work, we investigate a formalism for monitor unit (MU) calculation in Monte Carlo based treatment planning. By relating MU to dose measured under reference calibration conditions (central axis, depth of dose maximum in water, 10 cm × 10 cm field defined at 100 cm source-to-surface distance) our formalism determines the MU required for a treatment plan based on the prescription dose and Monte Carlo calculated dose distribution. Detailed descriptions and formulae are given for various clinical situations including conventional treatments and advanced techniques such as intensity-modulated radiotherapy (IMRT) and modulated electron radiotherapy (MERT). Analysis is made of the effects of source modelling, beam modifier simulation and patient dose calculation accuracy, all of which are important factors for absolute dose calculations using Monte Carlo simulations. We have tested the formalism through phantom measurements and the predicted MU values were consistent with measured values to within 2%. The formalism has been used for MU calculation and plan comparison for advanced treatment techniques such as MERT, extracranial stereotactic IMRT, MRI-based treatment planning and intensity-modulated laser-proton therapy studies. It is also used for absolute dose calculations using Monte Carlo simulations for treatment verification, which has become part of our comprehensive IMRT quality assurance programme.

Comparison of measured and Monte Carlo calculated dose distributions in inhomogeneous phantoms in clinical electron beams

Calculations of dose distributions in heterogeneous phantoms in clinical electron beams, carried out using the fast voxel Monte Carlo (MC) system XVMC and the conventional MC code EGSnrc, were compared with measurements. Irradiations were performed using the 9 MeV and 15 MeV beams from a Varian Clinac-18 accelerator with a 10 × 10 cm2 applicator and an SSD of 100 cm. Depth doses were measured with thermoluminescent dosimetry techniques (TLD 700) in phantoms consisting of slabs of Solid Water™ (SW) and bone and slabs of SW and lung tissue-equivalent materials. Lateral profiles in water were measured using an electron diode at different depths behind one and two immersed aluminium rods. The accelerator was modelled using the EGS4/BEAM system and optimized phase-space files were used as input to the EGSnrc and the XVMC calculations. Also, for the XVMC, an experiment-based beam model was used. All measurements were corrected by the EGSnrc-calculated stopping power ratios. Overall, there is excellent agreement between the corrected experimental and the two MC dose distributions. Small remaining discrepancies may be due to the non-equivalence between physical and simulated tissue-equivalent materials and to detector fluence perturbation effect correction factors that were calculated for the 9 MeV beam at selected depths in the heterogeneous phantoms.

Smoothing Monte Carlo calculated dose distributions by iterative reduction of noise

A smoothing algorithm based on an optimization procedure is presented and evaluated for single electron and photon beams and a full intensity modulated radiation therapy (IMRT) delivery. The algorithm iteratively reduces the statistical noise of Monte Carlo (MC) calculated dose distributions. It is called IRON (iterative reduction of noise). By varying the dose in each voxel, the algorithm minimizes the second partial derivatives of dose with respect to X, Y and Z. An additional restoration term ensures that too large dose changes are prevented. IRON requires a MC calculated one-dimensional or three-dimensional dose distribution with or without known statistical uncertainties as input. The algorithm is tested using three different treatment plan examples, a photon beam dose distribution in water, an IMRT plan of a real patient and an electron beam dose distribution in a water phantom with inhomogeneities. It is shown that smoothing can lead to an additional reduction of MC calculation time by factors of 2 to 10. This is especially useful if MC dose calculation is part of an inverse treatment planning system. In addition to this, it is shown that smoothing a noisy dose distribution may introduce some bias into the final dose values by converting the statistical uncertainty of the dose distribution into a systematic deviation of the dose value.