skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Effects of Hounsfield number conversion on CT based proton Monte Carlo dose calculations

Abstract

The Monte Carlo method provides the most accurate dose calculations on a patient computed tomography (CT) geometry. The increase in accuracy is, at least in part, due to the fact that instead of treating human tissues as water of various densities as in analytical algorithms, the Monte Carlo method allows human tissues to be characterized by elemental composition and mass density, and hence allows the accurate consideration of all relevant electromagnetic and nuclear interactions. On the other hand, the algorithm to convert CT Hounsfield numbers to tissue materials for Monte Carlo dose calculation introduces uncertainties. There is not a simple one to one correspondence between Hounsfield numbers and tissue materials. To investigate the effects of Hounsfield number conversion for proton Monte Carlo dose calculations, clinical proton treatment plans were simulated using the Geant4 Monte Carlo code. Three Hounsfield number to material conversion methods were studied. The results were compared in forms of dose volume histograms of gross tumor volume and clinical target volume. The differences found are generally small but can be dosimetrically significant. Further, different methods may cause deviations in the predicted proton beam range in particular for deep proton fields. Typically, slight discrepancies in mass density assignments playmore » only a minor role in the target region, whereas more significant effects are caused by different assignments in elemental compositions. In the presence of large tissue inhomogeneities, for head and neck treatments, treatment planning decisions could be affected by these differences because of deviations in the predicted tumor coverage. Outside the target area, differences in elemental composition and mass density assignments both may play a role. This can lead to pronounced effects for organs at risk, in particular in the spread-out Bragg peak penumbra or distal regions. In addition, the significance of the elemental composition effect (dose to water vs. dose to tissue) is tissue-type dependent and is also affected by nuclear reactions.« less

Authors:
; ;  [1]
  1. Department of Radiation Oncology, Massachusetts General Hospital, Boston, Massachusetts (United States)
Publication Date:
OSTI Identifier:
20951161
Resource Type:
Journal Article
Resource Relation:
Journal Name: Medical Physics; Journal Volume: 34; Journal Issue: 4; Other Information: DOI: 10.1118/1.2715481; (c) 2007 American Association of Physicists in Medicine; Country of input: International Atomic Energy Agency (IAEA)
Country of Publication:
United States
Language:
English
Subject:
62 RADIOLOGY AND NUCLEAR MEDICINE; ACCURACY; ALGORITHMS; ANIMAL TISSUES; BRAGG CURVE; COMPUTERIZED TOMOGRAPHY; DOSIMETRY; MONTE CARLO METHOD; NEOPLASMS; PROTON BEAMS; RADIATION DOSES; RADIOTHERAPY

Citation Formats

Jiang Hongyu, Seco, Joao, and Paganetti, Harald. Effects of Hounsfield number conversion on CT based proton Monte Carlo dose calculations. United States: N. p., 2007. Web. doi:10.1118/1.2715481.
Jiang Hongyu, Seco, Joao, & Paganetti, Harald. Effects of Hounsfield number conversion on CT based proton Monte Carlo dose calculations. United States. doi:10.1118/1.2715481.
Jiang Hongyu, Seco, Joao, and Paganetti, Harald. Sun . "Effects of Hounsfield number conversion on CT based proton Monte Carlo dose calculations". United States. doi:10.1118/1.2715481.
@article{osti_20951161,
title = {Effects of Hounsfield number conversion on CT based proton Monte Carlo dose calculations},
author = {Jiang Hongyu and Seco, Joao and Paganetti, Harald},
abstractNote = {The Monte Carlo method provides the most accurate dose calculations on a patient computed tomography (CT) geometry. The increase in accuracy is, at least in part, due to the fact that instead of treating human tissues as water of various densities as in analytical algorithms, the Monte Carlo method allows human tissues to be characterized by elemental composition and mass density, and hence allows the accurate consideration of all relevant electromagnetic and nuclear interactions. On the other hand, the algorithm to convert CT Hounsfield numbers to tissue materials for Monte Carlo dose calculation introduces uncertainties. There is not a simple one to one correspondence between Hounsfield numbers and tissue materials. To investigate the effects of Hounsfield number conversion for proton Monte Carlo dose calculations, clinical proton treatment plans were simulated using the Geant4 Monte Carlo code. Three Hounsfield number to material conversion methods were studied. The results were compared in forms of dose volume histograms of gross tumor volume and clinical target volume. The differences found are generally small but can be dosimetrically significant. Further, different methods may cause deviations in the predicted proton beam range in particular for deep proton fields. Typically, slight discrepancies in mass density assignments play only a minor role in the target region, whereas more significant effects are caused by different assignments in elemental compositions. In the presence of large tissue inhomogeneities, for head and neck treatments, treatment planning decisions could be affected by these differences because of deviations in the predicted tumor coverage. Outside the target area, differences in elemental composition and mass density assignments both may play a role. This can lead to pronounced effects for organs at risk, in particular in the spread-out Bragg peak penumbra or distal regions. In addition, the significance of the elemental composition effect (dose to water vs. dose to tissue) is tissue-type dependent and is also affected by nuclear reactions.},
doi = {10.1118/1.2715481},
journal = {Medical Physics},
number = 4,
volume = 34,
place = {United States},
year = {Sun Apr 15 00:00:00 EDT 2007},
month = {Sun Apr 15 00:00:00 EDT 2007}
}
  • The GEANT4 Monte Carlo code provides many powerful functions for conducting particle transport simulations with great reliability and flexibility. However, as a general purpose Monte Carlo code, not all the functions were specifically designed and fully optimized for applications in radiation therapy. One of the primary issues is the computational efficiency, which is especially critical when patient CT data have to be imported into the simulation model. In this paper we summarize the relevant aspects of the GEANT4 tracking and geometry algorithms and introduce our work on using the code to conduct dose calculations based on CT data. The emphasismore » is focused on modifications of the GEANT4 source code to meet the requirements for fast dose calculations. The major features include a quick voxel search algorithm, fast volume optimization, and the dynamic assignment of material density. These features are ready to be used for tracking the primary types of particles employed in radiation therapy such as photons, electrons, and heavy charged particles. Re-calculation of a proton therapy treatment plan generated by a commercial treatment planning program for a paranasal sinus case is presented as an example.« less
  • Purpose: Dose calculation based on pencil beam (PB) algorithms has its shortcomings predicting dose in tissue heterogeneities. The aim of this study was to compare dose distributions of clinically applied non-intensity-modulated radiotherapy 15-MV plans for stereotactic body radiotherapy between voxel Monte Carlo (XVMC) calculation and PB calculation for lung lesions. Methods and Materials: To validate XVMC, one treatment plan was verified in an inhomogeneous thorax phantom with EDR2 film (Eastman Kodak, Rochester, NY). Both measured and calculated (PB and XVMC) dose distributions were compared regarding profiles and isodoses. Then, 35 lung plans originally created for clinical treatment by PB calculationmore » with the Eclipse planning system (Varian Medical Systems, Palo Alto, CA) were recalculated by XVMC (investigational implementation in PrecisePLAN [Elekta AB, Stockholm, Sweden]). Clinically relevant dose-volume parameters for target and lung tissue were compared and analyzed statistically. Results: The XVMC calculation agreed well with film measurements (<1% difference in lateral profile), whereas the deviation between PB calculation and film measurements was up to +15%. On analysis of 35 clinical cases, the mean dose, minimal dose and coverage dose value for 95% volume of gross tumor volume were 1.14 {+-} 1.72 Gy, 1.68 {+-} 1.47 Gy, and 1.24 {+-} 1.04 Gy lower by XVMC compared with PB, respectively (prescription dose, 30 Gy). The volume covered by the 9 Gy isodose of lung was 2.73% {+-} 3.12% higher when calculated by XVMC compared with PB. The largest differences were observed for small lesions circumferentially encompassed by lung tissue. Conclusions: Pencil beam dose calculation overestimates dose to the tumor and underestimates lung volumes exposed to a given dose consistently for 15-MV photons. The degree of difference between XVMC and PB is tumor size and location dependent. Therefore XVMC calculation is helpful to further optimize treatment planning.« less
  • Computed tomography (CT) images of patients having metallic implants or dental fillings exhibit severe streaking artifacts. These artifacts may disallow tumor and organ delineation and compromise dose calculation outcomes in radiotherapy. We used a sinogram interpolation metal streaking artifact correction algorithm on several phantoms of exact-known compositions and on a prostate patient with two hip prostheses. We compared original CT images and artifact-corrected images of both. To evaluate the effect of the artifact correction on dose calculations, we performed Monte Carlo dose calculation in the EGSnrc/DOSXYZnrc code. For the phantoms, we performed calculations in the exact geometry, in the originalmore » CT geometry and in the artifact-corrected geometry for photon and electron beams. The maximum errors in 6 MV photon beam dose calculation were found to exceed 25% in original CT images when the standard DOSXYZnrc/CTCREATE calibration is used but less than 2% in artifact-corrected images when an extended calibration is used. The extended calibration includes an extra calibration point for a metal. The patient dose volume histograms of a hypothetical target irradiated by five 18 MV photon beams in a hypothetical treatment differ significantly in the original CT geometry and in the artifact-corrected geometry. This was found to be mostly due to miss-assignment of tissue voxels to air due to metal artifacts. We also developed a simple Monte Carlo model for a CT scanner and we simulated the contribution of scatter and beam hardening to metal streaking artifacts. We found that whereas beam hardening has a minor effect on metal artifacts, scatter is an important cause of these artifacts.« less
  • Purpose: The intensive use of Cone-Beam Computed Tomography (CBCT) during radiotherapy treatments raise some questions about the dose to healthy tissues delivered during image acquisitions. We hence developed a Monte Carlo (MC)-based tool to predict doses to organs delivered by the Elekta XVI kV-CBCT. This work aims at assessing the dosimetric accuracy of the MC tool, in all tissue types. Methods: The kV-CBCT MC model was developed using the PENELOPE code. The beam properties were validated against measured lateral and depth dose profiles in water, and energy spectra measured with a CdTe detector. The CBCT simulator accuracy then required verificationmore » in clinical conditions. For this, we compared calculated and experimental dose values obtained with OSL nanoDots and XRQA2 films inserted in CIRS anthropomorphic phantoms (male, female, and 5-year old child). Measurements were performed at different locations, including bone and lung structures, and for several acquisition protocols: lung, head-and-neck, and pelvis. OSLs and film measurements were corrected when possible for energy dependence, by taking into account for spectral variations between calibration and measurement conditions. Results: Comparisons between measured and MC dose values are summarized in table 1. A mean difference of 8.6% was achieved for OSLs when the energy correction was applied, and 89.3% of the 84 dose points were within uncertainty intervals, including those in bones and lungs. Results with XRQA2 are not as good, because incomplete information about electronic equilibrium in film layers hampered the application of a simple energy correction procedure. Furthermore, measured and calculated doses (Fig.1) are in agreement with the literature. Conclusion: The MC-based tool developed was validated with an extensive set of measurements, and enables the organ dose calculation with accuracy. It can now be used to compute and report doses to organs for clinical cases, and also to drive strategies to optimize imaging protocols.« less
  • Purpose: To establish a simple relation of backscatter dose enhancement around a high-Z dental alloy in head and neck radiation therapy to its average atomic number based on Monte Carlo calculations. Methods: The PHITS Monte Carlo code was used to calculate dose enhancement, which is quantified by the backscatter dose factor (BSDF). The accuracy of the beam modeling with PHITS was verified by comparing with basic measured data namely PDDs and dose profiles. In the simulation, a high-Z alloy of 1 cm cube was embedded into a tough water phantom irradiated by a 6-MV (nominal) X-ray beam of 10 cmmore » × 10 cm field size of Novalis TX (Brainlab). The ten different materials of high-Z alloys (Al, Ti, Cu, Ag, Au-Pd-Ag, I, Ba, W, Au, Pb) were considered. The accuracy of calculated BSDF was verified by comparing with measured data by Gafchromic EBT3 films placed at from 0 to 10 mm away from a high-Z alloy (Au-Pd-Ag). We derived an approximate equation to determine the relation of BSDF and range of backscatter to average atomic number of high-Z alloy. Results: The calculated BSDF showed excellent agreement with measured one by Gafchromic EBT3 films at from 0 to 10 mm away from the high-Z alloy. We found the simple linear relation of BSDF and range of backscatter to average atomic number of dental alloys. The latter relation was proven by the fact that energy spectrum of backscatter electrons strongly depend on average atomic number. Conclusion: We found a simple relation of backscatter dose enhancement around high-Z alloys to its average atomic number based on Monte Carlo calculations. This work provides a simple and useful method to estimate backscatter dose enhancement from dental alloys and corresponding optimal thickness of dental spacer to prevent mucositis effectively.« less