Poon Emily, Verhaegen Frank
Medical Physics Unit, McGill University, 1650 Cedar Avenue, Montreal H3G 1A4, Canada.
Med Phys. 2005 Jun;32(6):1696-711. doi: 10.1118/1.1895796.
This work involves a validation of the photon and electron transport of the GEANT4 particle simulation toolkit for radiotherapy physics applications. We examine the cross sections and sampling algorithms of the three electromagnetic physics models in version 4.6.1 of the toolkit: Standard, Low-energy, and Penelope. The depth dose distributions in water for incident monoenergetic and clinical beams are compared to the EGSNRC results. In photon beam simulations, all three models agree with EGSNRC to within 2%, except for the buildup region. Larger deviations are found for incident electron beams, and the differences are affected by user-imposed electron step limitations. Particle distributions through thin layers of clinical target materials, and perturbation effects near high-Z and low-Z interfaces are also investigated. The electron step size artifacts observed in our studies indicate potential problems with the condensed history algorithm. A careful selection of physics processes and transport parameters is needed for optimum efficiency and accuracy.
这项工作涉及对GEANT4粒子模拟工具包在放射治疗物理应用中的光子和电子输运进行验证。我们研究了该工具包4.6.1版本中三种电磁物理模型的截面和采样算法：标准模型、低能模型和佩内洛普模型。将单能入射束和临床束在水中的深度剂量分布与EGSNRC的结果进行比较。在光子束模拟中，除了剂量建成区外，所有三种模型与EGSNRC的结果一致，偏差在2%以内。对于入射电子束，发现偏差较大，且这些差异受用户设定的电子步长限制的影响。还研究了通过临床靶材料薄层的粒子分布以及高Z和低Z界面附近的微扰效应。我们研究中观察到的电子步长大小伪影表明凝聚历史算法存在潜在问题。为了实现最佳效率和准确性，需要仔细选择物理过程和输运参数。
