Technical notes Monte Carlo photon beam modeling and commissioning for radiotherapy dose calculation algorithm A. Toutaoui a, * , S. Ait chikh b , N. Khelassi-Toutaoui a , B. Hattali c a D epartement de Physique M edicale, Centre de Recherche Nucl eaire d'Alger, 2 Bd Frantz Fanon BP399 Alger RP, Algiers, Algeria b Service de Radioth erapie, Centre Pierre et Marie Curie, 1 Rue Battandier Place du 1er Mai, Algiers, Algeria c Service de Radioth erapie, Centre Anti-Cancer d'Oran, Oran, Algeria article info Article history: Received 11 February 2014 Received in revised form 21 May 2014 Accepted 22 May 2014 Available online xxx Keywords: Monte Carlo TPS commissioning EGSnrc abstract The aim of the present work was a Monte Carlo verification of the Multi-grid superposition (MGS) dose calculation algorithm implemented in the CMS XiO (Elekta) treatment planning system and used to calculate the dose distribution produced by photon beams generated by the linear accelerator (linac) Siemens Primus. The BEAMnrc/DOSXYZnrc (EGSnrc package) Monte Carlo model of the linac head was used as a benchmark. In the first part of the work, the BEAMnrc was used for the commissioning of a 6 MV photon beam and to optimize the linac description to fit the experimental data. In the second part, the MGS dose distributions were compared with DOSXYZnrc using relative dose error comparison and g- index analysis (2%/2 mm, 3%/3 mm), in different dosimetric test cases. Results show good agreement between simulated and calculated dose in homogeneous media for square and rectangular symmetric fields. The g-index analysis confirmed that for most cases the MGS model and EGSnrc doses are within 3% or 3 mm. © 2014 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved. Introduction Dose calculation algorithms in TPS can be broadly classified into measurement-based, such as the Clarkson algorithm [1], and model-based approaches [2]. Most of the model-based algorithms are based on the convolution of the primary energy distribution with a kernel describing the energy spread by secondary particles; the kernel is usually pre-calculated from Monte Carlo data, allow- ing a combination of Monte Carlo calculation and analytical de- scriptions [3]. There are two classes of model-based algorithms: One is based on pencil-beam kernels, the other is based on point- spread functions. The latter class better approaches a full 3D algorithm. The usual approach in the evaluation of the accuracy of dose calculation algorithms is to compare their results with experi- mental measurements [4,5]. Monte Carlo (MC) simulation is considered as the most accurate method of radiation dose calcu- lation in radiotherapy [6]. It is usually used as a benchmarking tool in predicting dose distributions in phantoms, especially in cases where the experimental dose measurement is very difficult, or reaches its limitations. The aim of this work is to use an MC simulated model of a Pri- mus (Siemens AG, Germany) linear accelerator (linac) for the verification of a photon dose calculation algorithm in a commercial TPS (CMS Xio, Elekta) [7]. This algorithm has been validated against measurement with radiographic films and ion chambers [5]. Comparisons off the multigrid superposition (MGS) against MC simulations in phantoms with different densities have also been published [8]. Even if some of the experiments and results of the studies mentioned above will be partly replicated and thus confirmed by the data presented here, this study presents a detailed dosimetric evaluation of the performance of the MGS model, using two eval- uation approaches, over a range of fundamental clinically relevant irradiation geometries. It will also allow the establishment of a basic evaluation tool for the treatment planning algorithms used in clinical practice today. In this work, The EGSnrc MC code [9], including user codes BEAMnrc and DOSXYZnrc [10,11], was employed to model a Siemens Primus linac working in 6 MV photon mode and to calculate the dose distributions in various phantoms. These results were compared to those calculated by the model-based MGS algorithm. * Corresponding author. Tel.: þ213 21 43 44 44, þ213 554771849; fax: þ213 21 43 42 80. E-mail address: toutaoui.aek@gmail.com (A. Toutaoui). Contents lists available at ScienceDirect Physica Medica journal homepage: http://www.physicamedica.com http://dx.doi.org/10.1016/j.ejmp.2014.05.007 1120-1797/© 2014 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved. Physica Medica xxx (2014) 1e5 Please cite this article in press as: Toutaoui A, et al., Monte Carlo photon beam modeling and commissioning for radiotherapy dose calculation algorithm, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.007