基于蒙特卡罗方法与人体模型的硼中子俘获治疗硼浓度在线监测方法

针对现有的技术难以实现硼浓度在线分布实时监测，提出一种基于康普顿相机的双能瞬发γ硼浓度在线监测方法，并利用蒙特卡罗模拟对其可行性进行验证，实现了利用生成对抗网络算法快速提升康普顿相机重建图像质量的方法，训练后的网络可在0.83 s内优化原始图像，与反投影重建图像相比更接近真实的瞬发伽马分布图像；搭建双能瞬发γ硼浓度监测系统模型实现了基于新方法的硼浓度分布计算，肿瘤区域平均硼浓度计算结果为86.91 ppm，正常组织区域硼浓度为18.37 ppm，肿瘤区域硼浓度与正常组织区域的硼浓度比值平均为4.75，硼浓度比值的计算误差为5.00%，该结果验证了双能瞬发γ硼浓度监测新方法的可行性，为硼中子俘获治疗硼浓度在线监测方法提供了新的思路和理论指导。

Online BNCT boron concentration monitoring method based on Monte Carlo simulation and human body model

Abstract: The existing technology is difficult to realize real-time and online monitoring of boron concentration distribution. Therefore, a novel method for the online monitoring of boron concentration based on Compton camera for dual-energy prompt gamma imaging is proposed, and its feasibility is verified with Monte Carlo simulation. The proposed method uses generative adversarial network to quickly improve the quality of the reconstructed image of Compton camera. The trained network can optimize the original image within 0.83 s, and compared with the reconstructed image with back projection, the optimized image is closer to the real prompt gamma distribution image. The calculation of boron concentration distribution using the proposed method is realized on the constructed model of dual-energy prompt gamma detection based boron concentration monitoring system. The calculated average boron concentration is 86.91 ppm in the tumor area, and 18.37 ppm in the normal tissue area. The calculated average ratio of the boron concentration in the tumor area to that in the normal tissue area is 4.75, with a calculation error of 5.00%, which verifies the feasibility of the novel method for boron concentration monitoring based on dual-energy prompt gamma detection. The study provides new ideas and theoretical guidance for the online monitoring of boron concentration in boron neutron capture therapy.