摘要
目的建立一种基于遗传算法的三维适形放疗治疗机架入射角度最优化算法。方法从Pinnacle v 7．2计划系统上读取患者几何轮廓信息和三维剂量分布文件，把几何轮廓信息和三维剂量分布在同一坐标系下融合后作为最优化算法输入数据。对每组角度组合都计算其最优射野权重，用基于遗传算法角度最优化方法挑选出最优角度组合，最后把得到的角度组合和各角度射野权重重新输入到Pinnacle v 7．2计划系统，和常规三维适形计划相比较。对2例肺癌病例和1例脑瘤病例比较了优化入射角度和常规三维适形计划下剂量分布、DVH图差异和适形度指标。结果和常规三维适形计划相比，优化计划适形指数分别为0．59和0．70，高于常规三维适形计划的0．36和0．58。DVH图上看肺癌病例，使用优化算法后脊髓受到的最高量分别下降了17．8％和22．4％，肺V20分别降低了3．1％和4．4％，肺V30分别降低了4．5％和1．5％；对脑瘤病例，晶体和眼球受照剂量也有显著下降。结论所建立的角度最优化方法可作为计划设计的辅助工具。
Objective To develop a beam orientation optimization algorithm for the gantry orientation in three-dimensional conformal radiotherapy （3DCRT）. Methods Patients＇ data were imported from the Pinnacle v 7.2 treatment planning system,including the DICOMRT and dose distribution flies. These imported files were merged using a uniform coordinate system. The algorithm determined the optimized beam weight for each beam group and optimized the beam orientation with genetic algorithm. The optimized parameters,including the optimized beam orientations and weights, were exported back to the Pinnacle v 7.2 to compare with the conventional 3DCRT plan. The optimized algorithm was implemented with our in-house program. The dose distributions, the DVH diagram and the conformity index of two lung cancer patients were compared. Results For the two lung cancer patients, the conformity index of the optimized plan （0.59 and 0.7） was higher than the conwentional 3DCRT plan （0.36 and 0.58）. The maximum dose in spinal cord was reduced by 17.8% and 22.4% ,the lung V20 reduced by 3.12% and 4.35% ,and V30 reduced by 4.47% and 1.49%. For the brain tumor patient,the dose of lens and eyes was also decreased significantly. Conclusion This beam orientation optimization can be used as an assistant planning tool.
出处
《中华放射肿瘤学杂志》
CSCD
北大核心
2008年第4期304-307,共4页
Chinese Journal of Radiation Oncology
基金
国家自然科学基金部分资助（N010275045）
关键词
三维适形放疗
角度最优化算法
遗传算法
Three-dimensional conformal radiotherapy
Beam orientation optimization algorithm
Genetic algorithm
