质子治疗中瞬发 γ 射线分布的模拟与实验研究

[Background] Proton therapy achieves high precision irradiation of tumors by precisely controlling the position of the Bragg peak of the proton beam. To further enhance treatment accuracy, real-time monitoring of the Bragg peak position is necessary, and the use of prompt gamma rays for real-time monitoring has become a research focus with related detector systems under development. [Purpose] This study aims to investigate the variation trends of characteristic gamma ray yields under different proton energies and to explore the spatial flux distribution of prompt gamma rays in proton therapy. [Methods] Firstly, a LaBr₃(Ce) detector was used to perform comprehensive measurements of prompt gamma ray spectra at various angles (45°, 90°, and 135°) during solid water irradiation by proton beams under different energies (59.8 MeV, 121.1 MeV, 172.8 MeV, and 221.1 MeV). Then, the detector efficiency was corrected using FLUKA Monte Carlo software, and the yields of 511 keV, 718 keV, 2.22 MeV, and 4.4 MeV characteristic gamma rays were obtained. Finally, the Monte Carlo simulation toolkit Geant4 was utilized to model the Bragg peak position within solid water phantoms, the intensity and angular distribution of prompt gamma ray emissions, and the energy spectra as a function of proton penetration depth. [Results] The experimental and simulation results both indicate that the yields of the 511 keV, 718 keV, 2.22 MeV, and 4.4 MeV characteristic gamma rays increase as the proton beam energy increases. However, the Geant4 simulation results significantly overestimate the yields of the 2.22 MeV and 4.4 MeV characteristic gamma rays. The yields of the four characteristic gamma rays show a clear backward trend at proton energies of 59.8 MeV and 121.1 MeV, and this trend gradually weakens as the proton energy increases to 172.8 MeV and 221.1 MeV. The Geant4 simulation results well reproduce the angular distribution trends at proton energies of 59.8 MeV, 121.1 MeV, and 172.8 MeV, but for 221.1 MeV proton energy, the simulation results show a forward trend in angular distribution, contrary to experimental observations. [Conclusions] The Geant4 toolkit demonstrates reasonable predictive capability for characteristic gamma ray yields across varying proton energies, showing generally good agreement with experimental measurements despite systematic deviations in absolute yield calculations. The simulation accurately reproduces yield trends but overestimates 2.22 MeV and 4.4 MeV gamma ray yields by significant factors. Notable discrepancies between simulated and experimental angular distributions are observed, particularly at 221.1 MeV proton energy, which stem from the Geant4 physical cascade model's inadequate consideration of angular momentum coupling effects during the gamma ray emission process.