Measurement Based Source Model for On-Board Imager Monte Carlo Dosimetry Simulations
K McMillan*, A Michailian, M McNitt-Gray, D Ruan, University of California, Los Angeles, Los Angeles, CATH-C-BRA-6 Thursday 10:30:00 AM - 12:30:00 PM Room: Ballroom A
Purpose: The purpose of this study is to investigate and develop a systematic approach to quantify the dose impact of the radiographic and cone-beam computed tomography (CBCT) modes of operation of the on-board imager (OBI) on radiotherapy treatment machines, independent of manufacturer's data specifications.
Methods: In order to facilitate dose calculations using Monte Carlo (MC) methods, we have developed an equivalent source model for OBI systems consisting of an equivalent spectrum module incorporative of intrinsic filtration and an equivalent filter module characterizing the added bowtie filter. The equivalent spectrum was generated by iteratively updating the source spectrum estimate until the calculated beam behavior matched the measured half-value layer and quarter-value layer values. The equivalent filter was established by obtaining a series of exposure readings along the bowtie profile and calculating the equivalent path lengths of a particular filter material. The equivalent spectrum module was validated by assessing the consistency of the estimation results obtained from half- and full- bowtie measurements. The end-to-end system was validated by comparing dose distributions from MC simulations based on an equivalent spectrum and an equivalent filter with dose distributions from film measurements for both the full- and half- bowtie scenario.
Results: Results were generated using a 95 kVp beam. Agreement was observed between the equivalent source spectrums derived from measurements using either full- or half- bowtie filters. This consistency indirectly validates the consistency and accuracy of the equivalent source approach, as the experimental data are mutually independent, with source and internal filtration being the intrinsic common factor. A root-mean-square error of approximately 3% was observed between the MC and film measurement dose distributions for both bowtie filters.
Conclusions: We have developed a source model that generates valid input for OBI MC dosimetry simulations. The method uses relatively simple measurement data, instead of manufacturer's specifications.