UCSF Helen Diller Family Comprehensive Cancer Center, San Francisco, CA 94143-1708, USA.
Phys Med Biol. 2012 Jun 7;57(11):3295-308. doi: 10.1088/0031-9155/57/11/3295. Epub 2012 May 9.
A general method for handling time-dependent quantities in Monte Carlo simulations was developed to make such simulations more accessible to the medical community for a wide range of applications in radiotherapy, including fluence and dose calculation. To describe time-dependent changes in the most general way, we developed a grammar of functions that we call 'Time Features'. When a simulation quantity, such as the position of a geometrical object, an angle, a magnetic field, a current, etc, takes its value from a Time Feature, that quantity varies over time. The operation of time-dependent simulation was separated into distinct parts: the Sequence samples time values either sequentially at equal increments or randomly from a uniform distribution (allowing quantities to vary continuously in time), and then each time-dependent quantity is calculated according to its Time Feature. Due to this modular structure, time-dependent simulations, even in the presence of multiple time-dependent quantities, can be efficiently performed in a single simulation with any given time resolution. This approach has been implemented in TOPAS (TOol for PArticle Simulation), designed to make Monte Carlo simulations with Geant4 more accessible to both clinical and research physicists. To demonstrate the method, three clinical situations were simulated: a variable water column used to verify constancy of the Bragg peak of the Crocker Lab eye treatment facility of the University of California, the double-scattering treatment mode of the passive beam scattering system at Massachusetts General Hospital (MGH), where a spinning range modulator wheel accompanied by beam current modulation produces a spread-out Bragg peak, and the scanning mode at MGH, where time-dependent pulse shape, energy distribution and magnetic fields control Bragg peak positions. Results confirm the clinical applicability of the method.
我们开发了一种处理蒙特卡罗模拟中时变量的通用方法，以使医学领域能够更广泛地应用于放射治疗中的各种应用，包括剂量和剂量率计算。为了以最通用的方式描述时变变化，我们开发了一种我们称之为“时间特征”的函数语法。当模拟量（例如几何对象的位置、角度、磁场、电流等）来自时间特征时，该量随时间变化。时变模拟的操作分为几个明显的部分：序列要么以相等的增量顺序采样时间值，要么从均匀分布中随机采样（允许数量随时间连续变化），然后根据其时间特征计算每个时变数量。由于这种模块化结构，即使存在多个时变数量，也可以在单个模拟中以任意给定的时间分辨率高效地执行时变模拟。这种方法已经在 TOPAS（Particle 模拟工具）中实现，旨在使临床和研究物理学家更容易使用 Geant4 进行蒙特卡罗模拟。为了演示该方法，我们模拟了三种临床情况：一个可变的水柱状物体，用于验证加利福尼亚大学 Crocker 实验室眼部治疗设备的布拉格峰的稳定性；麻省总医院（MGH）被动束散射系统的双散射治疗模式，其中旋转调强器轮伴随着束流调制产生扩展布拉格峰；MGH 的扫描模式，其中时变脉冲形状、能量分布和磁场控制布拉格峰位置。结果证实了该方法的临床适用性。
