Monte Carlo Dose Distributions Research Articles

TH-C-T-617-10: Monte Carlo Investigation of Heterogeneity Effect for Head and Neck IMRT

Purpose: The purpose of this work is to verify the accuracy of the dose distribution calculated by the CORVUS treatment planning system and to investigate the heterogeneity effect on head and neck IMRT plans using Monte Carlo simulations. Method and Materials: A Monte Carlo dose calculation tool, MCSIM, was used to carry out patient-specific dose calculations. Patient-specific CT data and the IMRT RTP files generated by the CORVUS treatment planning system were used for Monte Carlo dose calculations. Ten head and neck IMRT treatment plans were re-calculated and compared with the original plans. The isodose distributions and the dose-volume histograms were used for comparison. In order to investigate the heterogeneity effect for head and neck patient, IMRT plans generated by CORVUS with and without heterogeneity correction were compared with Monte Carlo dose distributions re-calculated under the same conditions. Results: The mean target dose and the dose at the isocenter predicted by CORVUS agreed to within ±5% of those predicted by Monte Carlo for all cases. Hot spots and cold spots were observed in the target volume due to the heterogeneity effect in some patients with beams going through bones, teeth and/or air cavities. D95 (dose received by 95% of the target volume) recalculated by the Monte Carlo method could be 8% lower than the original plan, which is outside the clinical acceptance criterion. More than 10% differences in the critical structure dose were also observed. Conclusion: In general, dose distributions from CORVUS with heterogeneity correction agreed to within ±5% with Monte Carlo calculations. Cold spots in the target volume due to inaccurate heterogeneity correction may compromise the local control of head and neck IMRT. The dosimetry differences in critical structures are another aspect worthy to be further investigated.

SU-FF-T-83: Partial Breast Treatment Using Energy- and Intensity- Modulated Photon and Electron Beams

Purpose: A new treatment technique has been developed for partial breast treatment using energy‐ and intensity‐modulated photon and electron beams.Treatment plans for partial breast irradiation using this technique are investigated in this work. Method and Materials: Two tangential photonbeams with several electron beams of different energies and field shapes were used for partial breast irradiation. The patient setup and the orientations of the two opposed tangential photonbeams were the same as conventional treatment for the whole breast. The electron beams were perpendicular to the patient skin surface. The dose‐distributions for the photon beamlets and electron apertures were calculated by a Monte Carlo code, EGS4/MCSIM. The optimization was performed based on the Monte Carlodose distributions employing a gradient search method. The photonbeams were delivered using a multi‐leaf collimator with a step‐and‐shot technique and the electron beams were delivered using cutouts. The effects of bremsstrahlung leakage and electron scattering by the cutout were taken into account in the Monte Carlo simulations.Results:Treatment plans using this new technique for five breast patients were generated for partial breast treatment. Compared with multi‐field photon IMRT, this new technique required no change for patient setup from the conventional technique and resulted in less dose to the ipsilateral lung and other normal tissues and less dose to the contralateral breast. Our results showed that the dose to the contralateral breast was reduced by about 50%. The volumes of normal tissues exposed to high and moderate radiation dose were reduced up to 50%. The doses to critical structures, such as the lung and the heart, were also reduced significantly. Conclusion: The new treatment technique integrates intensity‐ and energy‐modulated photon and electron beams and combines the advantages from both treatment modalities for partial breast treatment.

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.

Monte Carlo calculation of dose to spinal structures from intrathecally administered Yittrium-90

The treatment of cerebrospinal fluid (CSF) malignancies by intrathecal administration of Iodine-131 radiolabeled monoclonal antibodies has been previously studied. Based on the assumption that more healthy tissue will be spared when a pure beta-emitter is employed, Yittrium-90 has been proposed as a preferable radioimmunoconjugate to I-131. It is believed that the shorter range of beta particles will offer better localization of the dose to metastases found within the subarachnoid space. However, such assumptions have not been accurately verified and validated. The purpose of this study is to quantitatively evaluate the dose distribution to the CSF space and its surrounding spinal structures following intrathecal administration of Y-90 based upon Monte Carlo modeling of the Y-90 source coupled with a 3-D digital phantom of a region of the spinal cord. A 3-D digital phantom of a section of the T-spine was constructed from the Visible Human Project series of images. The original color photographs of the female cadaver were used to produce an 8-bit gray-scale series of images. This series of images was further processed so that only certain anatomical features of interest were represented. These features of interest include the spinal cord, central canal, subarachnoid space, pia mater, arachnoid, dura mater, vertebral bone, and intervertebral disc. Monte Carlo N-Particle (MCNP) version 4C was used to model the Y-90 radiation distribution. The size of the cells modeled with MCNP was made to coincide with the voxel size of the 3-D phantom. A set of images of just the CSF compartment was convolved with the radiation distribution obtained from the MCNP calculations to determine the dose distribution within the subarachnoid space and surrounding tissues. A fast Fourier transform (FFT) was utilized to transform the two sets of data into frequency space where they were multiplied together. The inverse Fourier transform was performed on their product in frequency space to obtain the convolution. The cumulated activity of Y-90 in the CSF compartment was calculated based on previously-published I-131 pharmacokinetic biodistribution data to obtain the overall dose due to the presence of Y-90 in the CSF. Dose calculations show that a significant dose is delivered to the subarachnoid space compared to the surrounding structures including vital tissues such as the spinal cord and bone marrow. Specifically, the average doses to tissues in Gy per mCi of initial activity are calculated to be: subarachnoid space = 1.4; central canal = 0.9; spinal cord = 0.6; pia mater = 0.9; arachnoid = 0.6; dura mater = 0.3. The vertebral bone (including marrow) and disc receive <0.02 and <0.005 Gy/mCi respectively. Monte Carlo dose distribution calculations confirm that Y-90 delivers localized radiation dose, which is confined largely to the subarachnoid space as compared to the surrounding vital tissues. Based upon our dose distribution calculations, Y-90 is well suited for the treatment of CSF malignancies

Adaptive anisotropic diffusion filtering of Monte Carlo dose distributions

The Monte Carlo method is the most accurate method for radiotherapy dose calculations, if used correctly. However, any Monte Carlo dose calculation is burdened with statistical noise. In this paper, denoising of Monte Carlo dose distributions with a three-dimensional adaptive anisotropic diffusion method was investigated. The standard anisotropic diffusion method was extended by changing the filtering parameters adaptively according to the local statistical noise. Smoothing of dose distributions with different noise levels in an inhomogeneous phantom, a conventional and an IMRT treatment case is shown. The resultant dose distributions were analysed using several evaluating criteria. It is shown that the adaptive anisotropic diffusion method can reduce statistical noise significantly (two to five times, corresponding to the reduction of simulation time by a factor of up to 20), while preserving important gradients of the dose distribution well. The choice of free parameters of the method was found to be fairly robust.

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.

Monte Carlo simulation and dose distribution of low energy electron irradiation of an apple

The most difficult technical challenge for surface irradiation of fruits and vegetables is the need to achieve a uniform dose over the entire surface of convoluted shapes. The main goal of this research was to calculate the dose distribution produced by low energy electron irradiation of a typical complex shape, an apple, using Monte Carlo simulation. In this study, Monte Carlo simulation was used to determine the dose distribution at the surface of an apple irradiated with e-beam generated by a Van de Graaf accelerator (1–2 MeV). The dose distribution was used to develop the best irradiation angle while rotating an irregularly shaped food material (the apple) for uniform surface irradiation. A software package, Monte Carlo N-particle, was used to simulate an electron beam irradiation with a 1.0, 1.5 and 2.0 MeV sources on an apple. The apple structure was simulated by creating 3-D apple-like shape geometry by joining two spheres of 3.2 cm of radius each. The entire body is centered at the origin and the spheres are centered at −1.2 and 1.2 cm on the y-axis in the z– y-plane. As the apple is tilted towards the source, the dose gets distributed mainly in the front top region of the apple. Likewise, as the apple is tilted against the source, the dose gets distributed primarily in the lower front portion of the apple. A combination irradiation treatment with the axis of the apple tilted 30° against and 30° towards the source irradiation position after one full revolution resulted in the most uniform dose distribution for the three energies source tested. The average dose at the surface of the front of the apple was in the range from 1.2 to 1.6 kGy and 3.6 to 4.0 kGy at the surface of the top and bottom.

Monte Carlo dose calculations and radiobiological modelling: analysisof the effect of the statistical noise of the dose distribution on theprobability of tumour control

The aim of this work is to investigate the influence of the statisticalfluctuations of Monte Carlo (MC) dose distributions on the dose volumehistograms (DVHs) and radiobiological models, in particular the Poisson modelfor tumour control probability (tcp). The MC matrix is characterized by a meandose in each scoring voxel, d, and a statistical error on the mean dose,σd; whilst the quantities d and σd depend on manystatistical and physical parameters, here we consider only their dependence onthe phantom voxel size and the number of histories from the radiation source.Dose distributions from high-energy photon beams have been analysed. It hasbeen found that the DVH broadens when increasing the statistical noise of thedose distribution, and the tcp calculation systematically underestimates thereal tumour control value, defined here as the value of tumour control whenthe statistical error of the dose distribution tends to zero. When increasingthe number of energy deposition events, either by increasing the voxeldimensions or increasing the number of histories from the source, the DVHbroadening decreases and tcp converges to the `correct' value. It is shownthat the underestimation of the tcp due to the noise in the dose distributiondepends on the degree of heterogeneity of the radiobiological parameters overthe population; in particular this error decreases with increasing thebiological heterogeneity, whereas it becomes significant in the hypothesis ofa radiosensitivity assay for single patients, or for subgroups of patients.It has been found, for example, that when the voxel dimension is changed froma cube with sides of 0.5 cm to a cube with sides of 0.25 cm (with a fixednumber of histories of 108 from the source), the systematic error in thetcp calculation is about 75% in the homogeneous hypothesis, and it decreasesto a minimum value of about 15% in a case of high radiobiologicalheterogeneity. The possibility of using the error on the tcp to decide howmany histories to run for a given voxel size is also discussed.

Monte Carlo dose computation for IMRT optimization*

A method which combines the accuracy of Monte Carlo dose calculation with afinite size pencil-beam based intensity modulation optimization is presented.The pencil-beam algorithm is employed to compute the fluence element updatesfor a converging sequence of Monte Carlo dose distributions. The combinationis shown to improve results over the pencil-beam based optimization in a lungtumour case and a head and neck case. Inhomogeneity effects like a broaderpenumbra and dose build-up regions can be compensated for by intensitymodulation.

Denoising of electron beam Monte Carlo dose distributions usingdigital filtering techniques

The Monte Carlo (MC) method has long been viewed as the ultimate dosedistribution computational technique. The inherent stochastic dosefluctuations (i.e. noise), however, have several important disadvantages:noise will affect estimates of all the relevant dosimetric and radiobiologicalindices, and noise will degrade the resulting dose contour visualizations. Wesuggest the use of a post-processing denoising step to reduce statisticalfluctuations and also improve dose contour visualization. We report theresults of applying four different two-dimensional digital smoothing filtersto two-dimensional dose images. The Integrated Tiger Series MC code was usedto generate 10 MeV electron beam dose distributions at various depths in twodifferent phantoms. The observed qualitative effects of filtering include:(a) the suppression of voxel-to-voxel (high-frequency) noise and(b) the resulting contour plots are visually more comprehensible. Drawbacksinclude, in some cases, slight blurring of penumbra near the surface andslight blurring of other very sharp real dosimetric features. Of the fourdigital filters considered here, one, a filter based on a local least-squaresprinciple, appears to suppress noise with negligible degradation of realdosimetric features. We conclude that denoising of electron beam MC dosedistributions is feasible and will yield improved dosimetric reliability andimproved visualization of dose distributions.

Electron beam modeling and commissioning for Monte Carlo treatment planning.

A hybrid approach for commissioning electron beam Monte Carlo treatment planning systems has been studied. The approach is based on the assumption that accelerators of the same type have very similar electron beam characteristics and the major difference comes from the on-site tuning of the electron incident energy at the exit window. For one type of accelerator, a reference machine can be selected and simulated with the Monte Carlo method. A multiple source model can be built on the full Monte Carlo simulation of the reference beam. When commissioning electron beams from other accelerators of the same type, the energy spectra in the source model are tuned to match the measured dose distributions. A Varian Clinac 2100C accelerator was chosen as the reference machine and a four-source beam model was established based on the Monte Carlo simulations. This simplified beam model can be used to generate Monte Carlo dose distributions accurately (within 2%/2 mm compared to those calculated with full phase space data) for electron beams from the reference machine with various nominal energies, applicator sizes, and SSDs. Three electron beams were commissioned by adjusting the energy spectra in the source model. The dose distributions calculated with the adjusted source model were compared with the dose distributions calculated using the phase space data for these beams. The agreement is within 1% in most of cases and 2% in all situations. This preliminary study has shown the capability of the commissioning approach for handling large variation in the electron incident energy. The possibility of making the approach more versatile is also discussed.

A pencil beam model for photon dose calculation.

A method for photon dose calculation in radio therapy planning using pencil beam energy deposition kernels is presented. It is designed to meet the requirements of an algorithm for 3-D treatment planning that is general enough to handle irregularly shaped radiation fields incident on a heterogeneous patient. It is point oriented and thus faster than a full 3-D convolution algorithm and uses the same physical data base to characterize a clinical beam as a full 3-D convolution algorithm. It is shown that photon therapy beams can be characterized with great accuracy from a combination of precalculated Monte Carlo energy deposition kernels and dose distributions measured in a water phantom. The data are used to derive analytical pencil beam kernels that are approximately partitionated into the dose from (i) primary released electrons and positrons, (ii) scattered, bremsstrahlung, and annihilation photons, (iii) contaminating photons, and (iv) charged particles from the collimator head. A semianalytical integration method, based on triangulation of the field, is developed for dose calculation using the analytical kernels. Dose is calculated in units normalized to the incident energy fluence which facilitates output factor calculation. For application in heterogeneous media, a scatter correction factor is derived using monodirectional convolution along the ray path. In homogeneous media results are compared with measurements and in heterogeneous media with Monte Carlo calculations and the Batho method.

In vivo measurement of lead in bone using X-ray fluorescence

The factors affecting the accuracy and minimum detectable concentration of in vivo tibia lead measurement are discussed, and it is demonstrated that the use of a 109Cd source in a backscatter geometry and using the 88 keV coherently scattered photon for normalisation optimises both criteria. The measurement is shown to be independent of variations in source-sample distance, thickness of overlying tissue and tibia size and shape. Applying the same technique in vitro to samples of human tibia and metatarsals, it is shown that the results are not significantly different (p approximately=0.9) from atomic absorption spectrometry results from another laboratory. The results of Monte Carlo dose distribution calculations are presented and compared with measurements using thermoluminescent dosemeters: the mean absorbed dose to a 20 cm leg section is less than 0.1 mGy (10 mrad) and the maximum absorbed skin dose is 0.45 mGy (45 mrad). For this dose the minimum detectable lead concentration is 10 mu g g-1. Finally, the technique has been applied to groups of normals and occupationally exposed workers, and the means have been shown to be significantly different, namely 10 and 31 mu g g-1 respectively. In the normal subjects tibia lead correlated strongly with age (r=0.63, p<0.001).