Monte Carlo computer simulation of a camera system for proton beam range verification in cancer treatment

Purpose: This paper investigated a framework to efficiently compute and merge Monte Carlo simulations with new technique development in cancer treatment. A newly designed 3-D prompt gamma (PG) camera system was developed using Monte Carlo computer simulation method to validate the utility of the camera simulated for in-beam proton range verification during cancer treatment. Materials and Methods: We used the TOPAS Monte Carlo code to model a 3-D PG slit-camera system based on pixelated cadmium zinc telluride (CZT) semiconductor detectors. The hierarchy structure of the system modeling includes a top level parameter file and a set of low level parameter files needed to specify particle source, physics setting, geometry, scoring, time-dependent motion, data and graphical output of the system. A high-performance computer cluster to perform the timeconsuming computing was used. The effectiveness of the Monte Carlo computer simulation for the system modeling was assessed through statistical simulation experiments. The energy spectra detected by the CZT camera were simulated by irradiating a polymethyl methacrylate (PMMA) phantom with a proton beam of 160 MeV. The axial beam shifting within the PMMA phantom and ranges of the distal beam spots of C-Shape dose distribution in the TG-119 phantom were measured by the modeled camera, respectively. Through simulating a double-head slit camera system, the transverse positions of the proton beam spots were determined based on the measurement of time of flights of photons from positron annihilation events induced by protons within patient. Results: The results from the Monte Carlo computer simulation showed that the energy spectra detected by the CZT camera consisted of main characteristic peaks, including the 4.44 MeV peak from de-excitation of C-12 and the 511 KeV peak from positron annihilation interaction. The gamma emission between 4.0 to 5.0 MeV contributed to the majority of the detected profiles. In the PMMA phantom experiments with 2.2 x 10(8) protons at 160 MeV simulated, a 1 mm standard deviation on range estimation was achieved. With the collimator at 25 cm away from the beam axis, a less than 3 mm standard deviation on range estimation in TG-119 phantom for the individual distal beam spots was obtained. Using the time-of-flight method, the transverse position accuracy of the beam spot measured was less than 1 cm using the simulated double-head camera. Conclusions: The study demonstrated the effectiveness of Monte Carlo computer simulation framework to model the proton beam range monitoring process. Moreover, the computationally modeled proton range verification capability using the 3-D PG camera system, with millimeter accuracy, can be readily translated from computer modeling to medical practice. (C) 2019 Elsevier B.V. All rights reserved.