Premium
A photon source model based on particle transport in a parameterized accelerator structure for Monte Carlo dose calculations
Author(s) -
Ishizawa Yoshiki,
Dobashi Suguru,
Kadoya Noriyuki,
Ito Kengo,
Chiba Takahito,
Takayama Yoshiki,
Sato Kiyokazu,
Takeda Ken
Publication year - 2018
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1002/mp.12976
Subject(s) - monte carlo method , truebeam , photon , physics , linear particle accelerator , monte carlo method for photon transport , imaging phantom , bremsstrahlung , computational physics , dosimetry , optics , beam (structure) , hybrid monte carlo , nuclear medicine , markov chain monte carlo , mathematics , statistics , medicine
Purpose An accurate source model of a medical linear accelerator is essential for Monte Carlo ( MC ) dose calculations. This study aims to propose an analytical photon source model based on particle transport in parameterized accelerator structures, focusing on a more realistic determination of linac photon spectra compared to existing approaches. Methods We designed the primary and secondary photon sources based on the photons attenuated and scattered by a parameterized flattening filter. The primary photons were derived by attenuating bremsstrahlung photons based on the path length in the filter. Conversely, the secondary photons were derived from the decrement of the primary photons in the attenuation process. This design facilitates these sources to share the free parameters of the filter shape and be related to each other through the photon interaction in the filter. We introduced two other parameters of the primary photon source to describe the particle fluence in penumbral regions. All the parameters are optimized based on calculated dose curves in water using the pencil‐beam–based algorithm. To verify the modeling accuracy, we compared the proposed model with the phase space data ( PSD ) of the Varian TrueBeam 6 and 15 MV accelerators in terms of the beam characteristics and the dose distributions. The EGS 5 Monte Carlo code was used to calculate the dose distributions associated with the optimized model and reference PSD in a homogeneous water phantom and a heterogeneous lung phantom. We calculated the percentage of points passing 1D and 2D gamma analysis with 1%/1 mm criteria for the dose curves and lateral dose distributions, respectively. Results The optimized model accurately reproduced the spectral curves of the reference PSD both on‐ and off‐axis. The depth dose and lateral dose profiles of the optimized model also showed good agreement with those of the reference PSD . The passing rates of the 1D gamma analysis with 1%/1 mm criteria between the model and PSD were 100% for 4 × 4, 10 × 10, and 20 × 20 cm 2 fields at multiple depths. For the 2D dose distributions calculated in the heterogeneous lung phantom, the 2D gamma pass rate was 100% for 6 and 15 MV beams. The model optimization time was less than 4 min. Conclusion The proposed source model optimization process accurately produces photon fluence spectra from a linac using valid physical properties, without detailed knowledge of the geometry of the linac head, and with minimal optimization time.