Commissioning of PENELOPE and GATE Monte Carlo models for 6 and 18 MV photon beams from the Siemens Artiste linac
Résumé
Purpose: Monte Carlo (MC) simulations currently allow accurate calculation of the dose delivered to patients in radiotherapy treatments, provided that the linear accelerator can be modelled accurately. The commissioning of the linac model is thus an important task to ensure the overall accuracy of the calculated dose. This work aims at commissioning MC models developed with PENELOPE and GATE for the 6 and 18 MV photon beams of the Siemens ARTISTE linac, by comparing simulation to measurement.
Materials: Linac photon beam models at 6 and 18 MV were commissioned. Percentage depth doses (PDD) and lateral beam profiles at dmax, 5 and 10 cm were measured for open square fields from 5x5 to 30x30 cm at 100 SSD using a PTW large MP3 water tank with a PTW Semiflex 0.125 cm3 ionization chamber. MC simulations were performed using PENELOPE 2006/PENFAST and GATE v6.0 (based on GEANT4 9.3.p02) in two steps: phase space files storing particles exiting the linac head were first computed and then used as input data for dose computation. Both codes are optimized for radiotherapy purposes and include commonly used variance reduction techniques. The linac geometry was described in the simulation based on the manufacturer specifications. The main parameters defining the incident electron beam, i.e. its mean energy and its radial intensity distribution, were determined following the methodology suggested by Pena et al (Med. Phys. 2007), using PDDs and lateral profiles of the 5x5, 10x10 and 30x30 cm fields. The incident electron beam was assumed to be monoenergetic and monodirectional. Dose computations in the water phantom were performed using PENFAST and GATE 6.0.0, with a voxel size of 4 mm in all directions for all field sizes.
Results: At depth greater than dmax, simulated PDDs with GATE and PENELOPE both agreed with measurement to better than 1% for 6 and 18 MV. At shallower depths, both MC codes underestimated the measured dose. Lateral profiles matched measured ones within 1%/1 mm at 6 MV and within 2%/2 mm at 18 MV, for both codes. Larger discrepancies are however observed for field sizes larger than 25x25 cm. Work is ongoing to identify the reasons for these discrepancies; the impact of the initial electron energy distribution is particularly investigated.
Conclusions: Photon beam models of the Siemens ARTISTE linac at 6 and 18 MV were developed with PENELOPE and GATE and commissioned. Good agreement was obtained between measured and simulated dose distributions, confirming the validity of the models.
Origine : Fichiers produits par l'(les) auteur(s)
Loading...