内照射小器官剂量计算中减方差技巧的比较和应用  

Comparison and application of variance-reduction techniques used in internal radiation dose calculations for small organs

作　　者：张崭[1] 李君利[1] 武祯[1] 范佳锦[1] 曾志[1] 

机构地区：[1]清华大学工程物理系,北京100084

出　　处：《清华大学学报（自然科学版）》2007年第z1期1051-1056,共6页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金资助项目(10575065)

摘　　要：在用人体模型计算内照射比吸收分数的Monte Carlo模拟中,对小尺寸、远距离器官的计算结果通常精度不够,可信度差,这就要求采用减方差技巧。该文采用一个简单模型比较了直接模拟和区域粒子分裂-轮盘赌、强迫碰撞与DXTRAN球、指向概率强迫碰撞3种减方差技巧的效果。结果表明指向概率强迫碰撞方法能更有效解决此类问题。利用中国人体数学模型,选择一组典型的源靶器官对,计算了一个实例,得到了可信度高的吸收分数值并体现了指向概率强迫碰撞方法的优越性。In Monte Carlo calculations for specific absorbed fractions of internal radiation dose,the results of some target organs,which are small and far from the source,are usually unreliable with low precision,which calls for variance-reduction techniques.Three variance-reduction techniques,including geometry splitting with Russian roulette,DXTRAN with force collision,and pointing probability with force collision,are compared with direct Monte Carlo simulation in a simple model representing human body.The results show that the pointing probability with force collision method is most efficient.The Chinese Mathematical Phantom was then used with a typical source-target group selected to calculate the absorbed fractions.The results give a set of reliable absorbed fraction data and illustrate the advantage of the pointing probability with force collision method.

关 键 词：MONTE CARLO方法 减方差技巧 内照射 吸收分数 

分 类 号：R144[医药卫生—公共卫生与预防医学]