【作者】 王晓东；

【导师】 刘亚强； 徐志勇；

【作者基本信息】 清华大学 ， 生物医学工程（专业学位）， 2014， 硕士

【摘要】 光子射线穿过空腔时，由于存在空腔边缘电子不平衡现象，将可能导致空腔边缘出现大于5%的剂量误差，直接影响到靶区剂量的均匀性，明显降低肿瘤病人的治疗效果。因此，明确放射治疗计划系统（Treatment Planning System,TPS）计算空腔边缘剂量的准确性在临床中有着十分重要的意义。本研究拟比较Elekta的两种剂量算法Monaco的XVMC算法和Xio的MGS算法在小体积空腔边缘侧向和纵向方向剂量计算的准确性，为涉及小体积空腔的放疗计划评估提供参考。采用EGS4程序用对医科达直线加速器进行蒙特卡罗(Monte Carlo)模拟，并使用IBA蓝水箱三维扫描系统验证该Monte Carlo算法的准确性。以Monte Carlo算法为纵向剂量计算的参考标准，以EBT2胶片测量结果为侧向剂量分布的参考标准。分别用XVMC和MGS算法计算带有空腔的等效水体模的中心轴纵向剂量分布和侧向剂量分布，深度方向计算结果与Monte Carlo算法进行比较，侧向剂量分布计算结果与EBT2胶片测量结果进行比较。进一步分析射野大小、能量、野数等因素对XVMC和MGS算法剂量计算的误差的影响。在小体积空腔边缘，XVMC和MGS都能预测到建成效应的发生，但是都低估了建成效应；即XVMC和MGS算法都高估了剂量。剂量计算误差与射野面积相关，射野面积趋小则剂量计算误差趋大。不同光子能量下，XVMC和MGS算法的计算结果都高估了剂量，随着能量的增加计算的误差增大。空腔越大二次建成效应就越大，空腔越大时XVMC和MGS算法误差越大。7射野空腔下表面的误差相比单野、三野和五野射野明显减少。计算网格大小对XVMC和MGS算法在处理空腔边缘剂量时造成的影响不大。XVMC和MGS算法在空腔边缘计算上均存在一定误差；因些，在放疗计划评估中，应根据所采用的算法而进行适当的调整。更多还原

【Abstract】 The effect of dose distribution influenced by tissue inhomogeneity is particularlyimportant in clinical practics. Human body contains several cavities. Studies haveshown that when photon beams pass through air cavity, the charged particledisequilibrium will appear near the air tissue interface. Larger than5%dose errors willoccur when this effect was ignored. The dose calculation accuracy directly affects thetreatment result of largest part of tumor patients. Monaco algorithm and Xio are twokinds of the most commonly used commercial TPS system.The accuracy of XVMC of Monaco and MGS of Xio near the interface in lateraland longitudinal direction was compared in our study. A Monte Carlo simulation on theElekta linear accelerator is performed with EGS4code. And measurements were madeusing IBA Dosimetry “blue phantom”3D scanning systems to verify Monte Carlosimulation. Results confirm the reliability of this Monte Carlo algorithm.Central axisdepth dose distribution and profiles in a water equivalent phantom with variously-sizedair cavities were calculated by XVMC and MGS, and compared with Monte Carlosimulation and EBT2film, respectively. The influence of cavity shape, cavity size, fieldsize, energy, the number of field on the XVMC and MGS dose calculation errors wasinvestigated. Our study has shown that XVMC and MGS both overestimate the dosenear the air tissue interface, XVMC performs better than MGS. Prediction errors arebase on filed size, energy photon beam the number of filed cavity density and grid does.The results of this study can be used as a guide in the clinical. The XVMC and MGSalgorithm should be enhanced to fully account for electron disequilibrium near aircavity interface.更多还原