Commissioning a BEAMnrc Monte-Carlo Model for a New Varian Linear Accelerator - Research Proposal Example

Commissioning a BEAMnrcMonte-Carlo Model for a New Varian Linear Accelerator Proposal Name Institution Commissioning a BEAMnrc Monte-Carlo Model for a New Varian Linear Accelerator Proposal 1.0 Introduction For the past 50 years, Monte Carlo (MC) method has been widely applied in medical field. And other related fields. The increased use of this technique is partially attributed to the escalating accessibility of several powerful software packages and partially attributed to the enormous growth in computing power for each unit cost in the past 50 years[Rog06]. The range of Monte Carlo methods applications in medical field is very broad. For instance, there are Monte Carlo techniques which are employed in commercial medical treatment planning systems for exterior beam radiotherapy (Heath, Scuntjens, & Sheikh-Bagheri, 2004; Cygler et al., 2004). In another differing application, Monte Carlo techniques were used by Flock et al. (1989) to stimulate the transport of near-infrared and optical photons with relevance to photo-dynamic therapy. MC methods have also been widely employed in photon transport problems including diagnostic x-ray applications (Boone &Scibert, 1988) or brachytherapy dosimetry. In a different context, the photo transport methods have been used to calculate the radiation protection quantities (Rogers, 2006). Electron beam planning programs are integrated by a number of commercial treatment planning systems. Nonetheless, several programs do not have comparable limitations or accuracy. Initially, electronic beam dose calculations were based on empirical function that used ray line geometries and acquired wide beam dose distributions in uniform media. A more advanced technique for electron beam dose calculations was developed in the early 1980s and was referred to as Pencil Beam algorithms, which was based on multiple scattering theories. One principle limitation of both the Pencil Beam algorithms and empirical methods is their incapability of predicting the depth dose distributions and accurately monitoring units for field sizes tinier than the magnitude of lateral scatter diagram. Currently, electron beam treatment planning can be performed by the available commercial dose calculation algorithm based on Monte Carlo technique in the Varian eclipse treatment planning system. MC electron dose calculation algorithm (eMC) has been implemented by Varian Medical Systems in the Eclipse treatment system (Xu et al., 2009). The MC technique accuracy in dose calculation for dosimetry verification and radiation therapy treatment is considered to be of high standard with substantial disputes indicated between dose distributions calculated using MC methods and dose calculated using conventional algorithms. Since Monte Carlo techniques may calculate dose distributions accurately for heterogeneous patient geometry and sophisticated beam delivery configurations, and since better computational algorithms and rapidly increasing computing power have minimized the time used to calculate the dose to a level accepted clinically, MC is anticipated to be broadly employed for dose calculation in radiation therapy treatment planning (Ma et al., 2008). This project will initially involve commissioning a BEAMnrcmonte-carlo model for a new Varian linear accelerator at the Royal Brisbane and Women’s Hospital (RBWH). The commissioning process involves optimising the monte-carlo model until the simulated radiation dose data it produces matches the measured radiation dose produced by the real accelerator. 2.0 Problem Statement and Justification Calculations of the radiation dose in the patient are routinely performed as part of the radiotherapy treatment planning process. Clinical treatment planning systems currently use analytical models for dose calculation in the patient using approximations to speed up the calculations. The approximations come with a loss of accuracy, particularly as the complexity of the treatment delivery increases, and for situations where the radiation passes through parts of the patient with significant variations in density e.g. the lung or bone. However, it is crucial for the success of a radiotherapy treatment that the dose distribution calculated or predicted as part of the treatment planning process matches the dose distribution that is actually delivered by the linear accelerator. Important clinical judgements are then made by the clinical team based on these dose calculations. Practically a compromise between speed and accuracy has to be found. Monte-Carlo is a mathematical technique that allows the physics of the radiation interactions in the treatment machine and patient to be modeled to a high accuracy (Reynaert et al., 2007; Verhaegen & Seuntjens, 2003). There are several general purpose MC computer software or codes that are freely available and can be applied to a wide range of mathematical and physical problems and some have more recently been adapted for medical physics applications. For example, the GEANT4 (Agostinelli et al., 2003, Poon & Verhaegen, 2005) and BEAMnrc/DOSXYZnrc (Rogers et al., 1995) Monte-Carlo computer codes will be used in this project; both have previously been successfully used to model radiotherapy treatment machines, dose calculation to the patient (Seco et al., 2005), and model radiation detectors including electronic portal imaging devices. Why then are Monte-Carlo simulations not applied in clinical treatment planning systems? The answer is that the computational times are currently prohibitive for routine interactive clinical use. They can however be extremely valuable in a research environment where time is less important and high accuracy and precision is required. This project will initially involve commissioning a BEAMnrcMonte-Carlo model for a new Varian linear accelerator at the Royal Brisbane and Women’s Hospital (RBWH). The commissioning process involves optimising the Monte-Carlo model until the simulated radiation dose data it produces matches the measured radiation dose produced by the real accelerator. 3.0 Project Objectives The main objective of the project is to evaluate the effectiveness of BEAMnrc MC model in dose calculations. To achieve this, the study will involve commissioning a BEAMnrc Monte-Carlo model for a new Varian linear accelerator at the Royal Brisbane and Women’s Hospital (RBWH). The commissioning process involves optimising the Monte-Carlo model until the simulated radiation dose data it produces matches the measured radiation dose produced by the real accelerator. 4.0 Literature Review This section will present an overview of work done previously that provides the required background for this research purposes. It will concentrate on various BEAMnrcmonte-carlo model topics and Varian linear accelerator issues. This section will begin with a thorough coverage BEAMnrcmonte-carlo model which will assist in setting the context of this research. In radiotherapy, Monte Carlo simulations present possibly the most accurate technique for patient dose calculations (Xu et al., 2009). The advancement of faster Monte Carlo simulation algorithms and the development of faster computational systems provide an incomparable opportunity for the utilization of MC calculations in radiation therapy treatment planning clinical environment. The development of BEAM code took place at NRC as section of a project to build a complete MC dose calculation algorithm intended for electron beam radiotherapy (Heath & Seuntjens, 2003). It was first released in 1995 with its associated software and its development has continued since then (Chetty et al., 2007; Shi et al., 2010). Several general purpose MC algorithms have been formulated to simulate transportion of photons and electrons. Perhaps the EGS code system is the most broadly used algorithm in the medical physics (Fix et al., 2004). There are a lot of other corresponding general purpose systems employed in medical field including the MCNP and ITS systems, which have integrated the electron algorithms from ETRAN (Chetty et al., 2007). Additional newer general purpose systems include GEANT4 and PENELOPE. The MCNP, EGS, and ITS/ETRAN systems are approximately of similar effectiveness in very simple geometries for calculations when variance reduction methods are not employed, while the other systems are likely to be much slower. BEAM is an EGS user code which is a significant special purpose code (Chetty et al., 2007). The BEAM program is optimised for the purpose of simulating the treatment head of radiotherapy methods and has several variance reduction methods to improve the effectiveness of the simulation. Even though the general-purpose codes accuracy may be approximately similar provided they are applied in a careful manner, these programs are regarded as excessively slow for routine treatment planning purposes (Heath, Scuntjens, & Sheikh-Bagheri, 2004). In relation to radiation therapy, there have been a number of Monte Carlo programs formulated to enhance the calculation effectiveness, particularly in simulation used in the patient. The PEREGRINE system has been a benchmarked against measurements and was founded at the Lawrence Livermore National Laboratory (Cygler et al., 2004). The PEREGRINE electron transport algorithm is considered to be modification version of the EGS4 (Cygler et al., 2004). A random hinge approach is used by PEREGRINE for electron transport mechanics. A number of efficiency enhancing and variance reduction methods are implemented in PEREGRINE, including range rejection, photon splitting, source particle reuse and Russian roulette (Rogers, 2006; Ma et al., 2008). With an aim of reducing the overall dose calculation time, calculation is parallelized on a number of computer processors. In PEREGRINE, source modeling is attained by carrying out a complete MC simulation of the accelerator head employing the BEAM code as well as utilizing the yield to develop a source model where regeneration of source particles takes place above the beam modifiers dependent on the patient. PERGRINE utilizes a number of approximations during beam transportation in the patient specified beam modifiers, accompanied by transportion in a patient’s Computer Tomography data set. The first Monte Carlo algorithm to receive approval by FDA 510-K was the PEREGRINE system and it was the earliest commercially existing photon beam treatment planning system in the US (Chetty et al., 2007). The Monte Carlo technique is currently broadly employed for modelling linear accelerators in medical field. Contrary to other commonly used methods, the Monte Carlo technique begins from initial tracks and principles specific particles histories; hence it puts into consideration the transportation of the secondary particles (Heath & Seuntjens, 2003). The Monte Carlo method employed for dose calculations gives accurate outcomes in areas of surface irregularities and tissue uniformities, and this provides the very accurate and convenient technique for the patient treatment dose distributions simulation. One limitation of this technique is the lengthy computing time required to acquire results for the dose with sensible statistical accuracy, particularly if it were photon beams (Flock et al., 1989; Poone & Verhaegen, 2005). Nonetheless, current advancement in Monte Carlo dose calculation algorithms as well as the increasing computing processing speed has caused MC dose calculation speed to be accepted for radiotherapy clinics. Despite the fact that there is availability of appropriate MC programs for patient dose calculation in radiotherapy, a drawback to their effectuation is the absence of an accurate and general model of the accelerator radiation source that will assist a utilizer with an impulsive accelerator to perpetrate a Monte Carlo algorithm for dose calculation that conforms to an accuracy requirement of 2mm or 2 percent for patient dose calculations (Fix et al., 2004). To overcome these limitations, a universal source model that may be adjusted to cope with the output of a common utilizer’s accelerator is required. As suggested by Jiang, Boyer, & Ma (2001) and Fippel et al (2003) the most excellent way of getting a source model is to analytically characterize the beam. Nevertheless, in the context of this style, it may be quite hard to extort the required input data, for example, to get the sources’ energy distribution. The distribution of energy referred here is being founded on a gamma distribution and it is assumed that, it is possible to calculate the off-axis energy spectrum using a scale on the central axis spectrum. Due to this, the scale factor can be set by using a correction method on basis of a measured half-value layer. This contributes to some estimation for the energy spectra. On the contrary, MC simulations may be used to obtain the information of the energy spectra directly (Fix et al., 2004). Additional approach is to carry out complete MC radiation transport simulations using the accelerator head and in process of this simulation, the PSD (phase-space-data) can be generated as it bears the essential data for every particle cutting through the phase-space (PS). The phase-space-data offers precise particle distributions within the phase-space plane, hence, in this context, phase-space files may be applied as source models directly (Fix et al., 2004). According to Fix et al. (2004), if particles are transported accurately to the PS plane through the beam-generating devices and the electron beam on top of the target is accurately modelled, then the dose calculation employing that PS file is capable of matching the measured dose. 5.0 Research Approach/Methodology This section will address how the research will be carried out, and methods to be used to gather the required information as well as how data will be analyzed. This section will also address the limit of the study and the data needed to collect. Justification of the methods used to collect the data will also be addressed in this section. Participants of this research will include people who have prior knowledge concerning Monte Carlo simulations and in particular BEAMnrc Monte Carlo model and Varian linear accelerator. Qualitative data collections methods will be applied in this research. The qualitative data for this research will be collected by studying the relevant literature concerning Monte Carlo simulation methods and any other relevant material that will assist in accomplishing the research. I also intend to use the interview method of qualitative data collection.Studying the relevant literature related to Monte Carlo simulations and their usage in dose calculations will set the background of the research. This step will give a clear picture of what research is intended to achieve. It will also set the direction of the research as the area of concern which is supposed reviewed will be well analyzed. The interview method will be employed as a backup of what will be collected from the first step. This method will also assist in gathering first-hand information from someone who has a hand experience on the Monte Carlo simulation methods and in particular BEAMnrc Monte-Carlo model. The people who will be interviewed will include those who have prior knowledge in Monte Carlo simulations including my supervisor. The qualitative data will be analyzed through an extensive review of the literature. This will help in understanding what others have done concerning areas related to the research. This will give me as researcher insight and knowledge required in making necessary decisions during data collection. The interview qualitative data will be analyzed using the discourse analysis.More research will be done on the internet sources to understand what is required for this research. Internet will assist in carrying out the investigation of this research. I believe there are a number of important sources in the internet which contain information concerning Monte Carlo simulation techniques. BEAMnrc Monte Carlo model is the main focus for this research and internet sources will assist in gathering information concerning this field. 6.0 Expected Outcome The expected outcome of the project is that the simulation radiation dose data produced by the BEAMnrc Monte Carlo model will match the radiation dose produced by the real accelerator. 7.0 Conclusion As the advancement of the MC simulations increases and as computer power continues to increase, MC methods for radiation transport simulation will proceed to gain importance in medical physics. In this project, the emphasis will be on the commissioning of a BEAMnrc Monte-Carlo model for a new Varian linear accelerator at the Royal Brisbane and Women’s Hospital (RBWH). The commissioning process involves optimising the Monte-Carlo model until the simulated radiation dose data it produces matches the measured radiation dose produced by the real accelerator. References Rog06: , (Rogers, 2006), Read More

This project will initially involve commissioning a BEAMnrcmonte-carlo model for a new Varian linear accelerator at the Royal Brisbane and Women’s Hospital (RBWH). The commissioning process involves optimising the monte-carlo model until the simulated radiation dose data it produces matches the measured radiation dose produced by the real accelerator. 2.0 Problem Statement and Justification Calculations of the radiation dose in the patient are routinely performed as part of the radiotherapy treatment planning process.

Clinical treatment planning systems currently use analytical models for dose calculation in the patient using approximations to speed up the calculations. The approximations come with a loss of accuracy, particularly as the complexity of the treatment delivery increases, and for situations where the radiation passes through parts of the patient with significant variations in density e.g. the lung or bone. However, it is crucial for the success of a radiotherapy treatment that the dose distribution calculated or predicted as part of the treatment planning process matches the dose distribution that is actually delivered by the linear accelerator.

Important clinical judgements are then made by the clinical team based on these dose calculations. Practically a compromise between speed and accuracy has to be found. Monte-Carlo is a mathematical technique that allows the physics of the radiation interactions in the treatment machine and patient to be modeled to a high accuracy (Reynaert et al., 2007; Verhaegen & Seuntjens, 2003). There are several general purpose MC computer software or codes that are freely available and can be applied to a wide range of mathematical and physical problems and some have more recently been adapted for medical physics applications.

For example, the GEANT4 (Agostinelli et al., 2003, Poon & Verhaegen, 2005) and BEAMnrc/DOSXYZnrc (Rogers et al., 1995) Monte-Carlo computer codes will be used in this project; both have previously been successfully used to model radiotherapy treatment machines, dose calculation to the patient (Seco et al., 2005), and model radiation detectors including electronic portal imaging devices. Why then are Monte-Carlo simulations not applied in clinical treatment planning systems? The answer is that the computational times are currently prohibitive for routine interactive clinical use.

They can however be extremely valuable in a research environment where time is less important and high accuracy and precision is required. This project will initially involve commissioning a BEAMnrcMonte-Carlo model for a new Varian linear accelerator at the Royal Brisbane and Women’s Hospital (RBWH). The commissioning process involves optimising the Monte-Carlo model until the simulated radiation dose data it produces matches the measured radiation dose produced by the real accelerator. 3.

0 Project Objectives The main objective of the project is to evaluate the effectiveness of BEAMnrc MC model in dose calculations. To achieve this, the study will involve commissioning a BEAMnrc Monte-Carlo model for a new Varian linear accelerator at the Royal Brisbane and Women’s Hospital (RBWH). The commissioning process involves optimising the Monte-Carlo model until the simulated radiation dose data it produces matches the measured radiation dose produced by the real accelerator. 4.0 Literature Review This section will present an overview of work done previously that provides the required background for this research purposes.

It will concentrate on various BEAMnrcmonte-carlo model topics and Varian linear accelerator issues. This section will begin with a thorough coverage BEAMnrcmonte-carlo model which will assist in setting the context of this research. In radiotherapy, Monte Carlo simulations present possibly the most accurate technique for patient dose calculations (Xu et al., 2009). The advancement of faster Monte Carlo simulation algorithms and the development of faster computational systems provide an incomparable opportunity for the utilization of MC calculations in radiation therapy treatment planning clinical environment.

Read More