Utilization of MCNP for the development of spectral stripping algorithms for orthovoltage radiation therapy systems

We are developing a Monte-Carlo based method for spectral stripping of a low-voltage X-ray detector spectrum for calibrating and performing quality assurance with orthovoltage (X-ray) radiation therapy systems. The technique involves a tungsten-based X-ray generator with limited energy spectral information of at the most two half value layers of shielding available from the manufacturer. The goal is to generate a complete energy spectrum of the source from a measured detector spectrum, by correcting for radiation losses and misplaced counts due to Compton scattering and the photoelectric effect. In this work, energy spectra were measured at generator voltages of 150–300 kVp with HPGe and Si detectors, with several optional shielding configurations. The source to detector geometry were modeled in MCNP, and the reaction rates and spatial locations of Compton scattering and photoelectric events in the detectors have been investigated as a function of source energy to gain information on where interactions occur for the given beam collimation and geometry. Photon leakage and energy flux out of the detector were also simulated, and initial spectral stripping has been performed on the HPGe and Si detector spectra.