RayStation/GATE Monte Carlo simulation framework for verification of proton therapy based on the12N imaging.

Abstract

Objective.¹²N, having a half-life of 11 ms, is a highly effective positron emitter that can potentially provide near real-time feedback in proton therapy. There is currently no framework for comparing and validating positron emission imaging of¹²N. This work describes the development and validation of a Monte Carlo (MC) framework to calculate the images of¹²N, as well as long-lived isotopes, originating from activation by protons.Approach. The available dual-panel Biograph mCT PET scanner was modeled in GATE and validated by comparing the simulated sensitivity map with the measured one. The distributions of¹²N and long-lived isotopes were calculated by RayStation and used as the input of GATE simulations. The RayStation/GATE combination was verified using proton beam irradiations of homogeneous phantoms. A 120 MeV pulsed pencil beam with 10⁸protons per pulse was used. Two-dimensional images were created from the GATE output and compared with the images based on the measurements and the 1D longitudinal projection of the full 2D image was used to calculate the¹²N activity range.Main results. The simulated sensitivity in the center of the FoV (5.44%) agrees well with the measured one (5.41%). The simulated and measured 2D sensitivity maps agree in good detail. The relative difference between the measured and simulated positron activity range for both¹²N and long-lived isotopes is less than 1%. The broadening of the¹²N images relative to those of the longer-lived isotopes can be understood in terms of the large positron range of¹²N.Significance. We developed and validated a MC framework based on RayStation/GATE to support the in-beam PET method for quality assurance of proton therapy. The inclusion of the very short-lived isotope¹²N makes the framework useful for developing near real-time verification. This represents a significant step towards translating¹²N real-time in vivo verification to the clinic.