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

Division of Medical Radiation Physics, Insel Hospital, University of Berne, Berne 3010, Switzerland.
Medical Physics (Impact Factor: 2.91). 04/2008; 35(4):1521-31. DOI: 10.1118/1.2885372
Source: PubMed

ABSTRACT 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.

1 Bookmark
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The purpose of this work is to revisit the impediments and characteristics of fast Monte Carlo techniques for applications in radiation therapy treatment planning using new methods of utilizing pregenerated electron tracks. The limitations of various techniques for the improvement of speed and accuracy of electron transport have been evaluated. A method is proposed that takes advantage of large available memory in current computer hardware for extensive generation of precalculated data. Primary tracks of electrons are generated in the middle of homogeneous materials (water, air, bone, lung) and with energies between 0.2 and 18 MeV using the EGSnrc code. Secondary electrons are not transported, but their position, energy, charge, and direction are saved and used as a primary particle. Based on medium type and incident electron energy, a track is selected from the precalculated set. The performance of the method is tested in various homogeneous and heterogeneous configurations and the results were generally within 2% compared to EGSnrc but with a 40-60 times speed improvement. In a second stage the authors studied the obstacles for further increased speed-ups in voxel geometries by including ray-tracing and particle fluence information in the pregenerated track information. The latter method leads to speed increases of about a factor of 500 over EGSnrc for voxel-based geometries. In both approaches, no physical calculation is carried out during the runtime phase after the pregenerated data has been stored even in the presence of heterogeneities. The precalculated data are generated for each particular material and this improves the performance of the precalculated Monte Carlo code both in terms of accuracy and speed. Precalculated Monte Carlo codes are accurate, fast, and physics independent and therefore applicable to different radiation types including heavy-charged particles.
    Medical Physics 03/2009; 36(2):530-40. · 2.91 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The aim of this work was a Monte Carlo (MC) based investigation of the impact of different radiation transport methods in collimators of a linear accelerator on photon beam characteristics, dose distributions, and efficiency. Thereby it is investigated if it is possible to use different simplifications in the radiation transport for some clinical situations in order to save calculation time. Within the Swiss Monte Carlo Plan, a GUI-based framework for photon MC treatment planning, different MC methods are available for the radiation transport through the collimators [secondary jaws and multileaf collimator (MLC)]: EGSnrc (reference), VMC++, and Pin (an in-house developed MC code). Additional nonfull transport methods were implemented in order to provide different complexity levels for the MC simulation: Considering collimator attenuation only, considering Compton scatter only or just the firstCompton process, and considering the collimators as totally absorbing. Furthermore, either a simple or an exact geometry of the collimators can be selected for the absorbing or attenuation method. Phasespaces directly above and dose distributions in a water phantom are analyzed for academic and clinical treatment fields using 6 and 15 MV beams, including intensity modulated radiation therapy with dynamic MLC. For all MC transport methods, differences in the radial mean energy and radial energy fluence are within 1% inside the geometric field. Below the collimators, the energy fluence is underestimated for nonfull MC transport methods ranging from 5% for Compton to 100% for Absorbing. Gamma analysis using EGSnrc calculated doses as reference shows that the percentage of voxels fulfilling a 1% /1 mm criterion is at least 98% when using VMC++, Compton, or firstCompton transport methods. When using the methods Pin, Transmission, Flat-Transmission, Flat-Absorbing or Absorbing, the mean value of points fulfilling this criterion over all tested cases is 97%, 88%, 74%, 68%, or 65%, respectively. However, compared to EGSnrc calculations, the gain in efficiency is a factor of up to 10 for VMC++ and up to 48 for the absorbing method. The results of this investigation suggest that it is an option to use a simple transport technique in the initial treatment planning process and use more accurate transport methods for the final dose calculation accepting longer calculation times.
    Medical Physics 02/2010; 37(2):492-504. · 2.91 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: For superficial lesions, the electrons may be used for radiation therapy. The high energy photons and electrons are produced by a Linear accelerator (Linac). Many of electron fields need the shielding of normal or critical organs. The electron shields are usually lead slabs with few millimeter thicknesses which should be placed near the skin, less than 1 cm away from skin. In the inspection of patients setting in a clinic by a physicist, it was noted that in some cases the technician places the shields far away from skin in the way that the shadow of the field still matches the shielded area. This is due to a conceptual mistake in which one assumes that electrons travel in a straight line and matching the shadow of lead slab is enough for the shielding. This project is about Monte Carlo simulation of this case and dosimetry in which the excess dose to the tissue under the shield is calculated. In this study, BEAMnrc and DOSXYZnrc are used for simulation of the Linac and the electron shields. The water phantom as well as the Linac head (NEPTON Linac) is simulated in the electron mode. The simulation is performed in three various cases in which the lead shield is placed in distances of 1, 20, 40 cm from the surface of the phantom. In all cases, the edge of the shield is matched with the light field, so the shields get smaller as they move from the surface because of the divergence of the light field. The simulations were done in two energies, 6 and 13 MeV. The experiments also were done with EDR2 film dosimetry and the simulation results were validated using the experimental results. In all cases, the dose under the shield was normalized to the dose in the center. The dose of the normal organ under the shield was 2, 38, 43% with respect to the center for shield distances of 1, 20, 40 cm, respectively. So there is a considerable increasing of the dose due to the distanced shielding. In this work exact amount of the dose from this mistake (distanced shielding) is calculated and simulated.
    Journal of medical signals and sensors. 01/2012; 2(3):144-148.


Available from