Development of techniques for increasing the accuracy of dose calculation in a convolution method used for body stereotactic radiotherapy

以下的研究已经完成了：·从六个线性加速器中重建了一套高能量X射线光谱仪，使用十个能量桶。每套套装都可以准确地应用于从低Z型材料到高Z型材料的防水材料。这是如何分析的，每个人都设置了不同的轴外距离。水中的一对主剂量和散射剂量内核被构造成用于每一个能量单元使用Monte Carlo模拟的基于不同方法的单色辐射。主剂量内核指数相对较小的值围绕光子交互点，当与一个主剂量内核在一个大水幻影中心与一个光子交互点进行比较时。这一事实可能导致在爪和MLC野外边界上增加剂量计算的准确性时，使用基于不同方法的主剂量内核。在水中施用，计算使用卷积方法，一组主要和分散剂量的内核为响应能量瓶，通常在± 1%时与测量相比较。命名为“这可能是有启发性的，因为当卷积被用于每个能量箱中的单色辐射，光束的硬化效应不需要被纳入帐户。”由于考虑到在一点上对空气中X射线强度的计算，对X射线目标和平坦过滤器平面使用错误函数的方法已经开发出来。对于一个打开的爪子字段，错误函数是在区域视觉上使用的，通过爪子拼贴器。对于MLC字段，一个叶场输出删减方法是通过点的区域视觉引入错误函数，并通过每个MLC叶子从点的视觉引入错误函数。

The following studies have been performed:・High-energy x-ray spectra sets from six linear accelerators were reconstructed using about ten energy bins. Each set can be applied accurately to attenuators from low Z materials like water to high Z materials like lead. It was analyzed how each set varies with the off-axis distance.・A pair of primary and scatter dose kernels in water was constructed for monochromatic radiation at each energy bin using Monte Carlo simulations on the basis of the differential method. The primary dose kernel indicates relatively smaller values around the photon interaction point, when compared with a primary dose kernel yielded with a photon interaction point at the center of a large water phantom. This fact may leads to the increase of dose calculation accuracy around the jaw and MLC field borders when using primary dose kernels based on the differential method.・The in-water doses, calculated using a convolution method with sets of primary and scatter dose kernels yielded for respective energy bins, are generally within ±1% when compared with measurements. Namely, it can be understood that when the convolution is performed for monochromatic radiation at each energy bin, the beam hardening effect does not need to be taken into account.・With respect to the calculation of in-air x-ray intensity at a point, a method of using error functions on the x-ray target and flattening filter planes has been developed. For an open jaw field, the error functions are used in the area visual via the jaw collimators from the point. For an MLC field, a leaf-field output subtraction method is introduced using the error functions in the area visual via the jaw collimators from the point and using the error functions in the are a visual via each MLC leaf from the point.