Department of Radiation Oncology, UC San Francisco Comprehensive Cancer Center, San Francisco, California 94143-1708, USA.
Med Phys. 2009 Oct;36(10):4577-91. doi: 10.1118/1.3218764.
The purposes of this study are to improve the accuracy of source and geometry parameters used in the simulation of large electron fields from a clinical linear accelerator and to evaluate improvement in the accuracy of the calculated dose distributions.
The monitor chamber and scattering foils of a clinical machine not in clinical service were removed for direct measurement of component geometry. Dose distributions were measured at various stages of reassembly, reducing the number of geometry variables in the simulation. The measured spot position and beam angle were found to vary with the beam energy. A magnetic field from the bending magnet was found between the exit window and the secondary collimators of sufficient strength to deflect electrons 1 cm off the beam axis at 100 cm from the exit window. The exit window was 0.05 cm thicker than manufacturer's specification, with over half of the increased thickness due to water pressure in the channel used to cool the window. Dose distributions were calculated with Monte Carlo simulation of the treatment head and water phantom using EGSnrc, a code benchmarked at radiotherapy energies for electron scatter and bremsstrahlung production, both critical to the simulation. The secondary scattering foil and monitor chamber offset from the collimator rotation axis were allowed to vary with the beam energy in the simulation to accommodate the deflection of the beam by the magnetic field, which was not simulated.
The energy varied linearly with bending magnet current to within 1.4% from 6.7 to 19.6 MeV, the bending magnet beginning to saturate at the highest beam energy. The range in secondary foil offset used to account for the magnetic field was 0.09 cm crossplane and 0.15 cm inplane, the range in monitor chamber offset was 0.14 cm crossplane and 0.07 cm inplane. A 1.5%/0.09 cm match or better was obtained to measured depth dose curves. Profiles measured at the depth of maximum dose matched the simulated profiles to 2.6% or better at doses of 80% or more of the dose on the central axis. The profiles along the direction of MLC motion agreed to within 0.16 cm at the edge of the field. There remained a mismatch for the lower beam energies at the edge of the profile that ran parallel to the direction of jaw motion of up to 1.4 cm for the 6 MeV beam, attributed to the MLC support block at the periphery of the field left out of the simulation and to beam deflection by the magnetic field. The possibility of using these results to perform accurate simulation without disassembly is discussed. Phase-space files were made available for benchmarking beam models and other purposes.
The match to measured large field dose distributions from clinical electron beams with Monte Carlo simulation was improved with more accurate source details and geometry details closer to manufacturer's specification than previously achieved.
本研究旨在提高临床直线加速器大电子场模拟中源和几何参数的准确性，并评估计算剂量分布的准确性的提高。
为了直接测量组件几何形状，将一台未投入临床使用的临床机器的监测腔和散射箔拆除。在重新组装的各个阶段测量剂量分布，减少了模拟中的几何变量数量。发现测量的光斑位置和射束角度随射束能量而变化。在出口窗和次级准直器之间发现了来自弯曲磁体的磁场，其强度足以在离出口窗 100 厘米处将电子沿束轴偏 1 厘米。出口窗比制造商的规格厚 0.05 厘米，其中超过一半的增加厚度是由于用于冷却窗的通道中的水压力。使用 EGSnrc 对治疗头和水模体进行蒙特卡罗模拟计算剂量分布，该代码在放射治疗能量下对电子散射和韧致辐射产生进行了基准测试，这对模拟都至关重要。允许模拟中二次散射箔和监测腔从准直器旋转轴的偏移随射束能量而变化，以适应磁场引起的射束偏转，而磁场未进行模拟。
从 6.7 到 19.6 MeV，束流的能量与弯曲磁体电流呈线性变化，在最高束能量下弯曲磁体开始饱和。用于补偿磁场的次级箔片偏移范围为 0.09 cm 横切和 0.15 cm 矢状，监测腔偏移范围为 0.14 cm 横切和 0.07 cm 矢状。测量深度剂量曲线的 1.5%/0.09 cm 匹配或更好。在最大剂量深度处测量的轮廓与模拟轮廓匹配，在中央轴上 80%或更多剂量处的匹配度为 2.6%或更好。在叶片运动方向上测量的轮廓在叶片边缘处的吻合度为 0.16 厘米或更好。对于与颌运动方向平行的轮廓的较低束能边缘仍存在不匹配，对于 6 MeV 束，不匹配高达 1.4 厘米，这归因于场边缘未包含在模拟中的叶片支撑块以及磁场引起的射束偏转。讨论了使用这些结果进行无需拆卸的精确模拟的可能性。提供了相空间文件以用于基准测试束模型和其他目的。
通过使用更准确的源细节和更接近制造商规格的几何细节，与临床电子束的大野剂量分布的蒙特卡罗模拟匹配得到了改善，超过了以前的水平。
