基于定积分的放疗权重因子自动优化方法研究  

Research on Automatic Optimization of Weighting Factors in Radiotherapy Based on Definite Integral

作　　者：郭彩萍[1] 张俊生[1,2] 张晓娟[1,2] Guo Caiping;Zhang Junsheng;Zhang Xiaojuan(Electronic Engineering Department,Taiyuan Institute of Technology,Taiyuan 030008,China;Institute of Information and Communication Engineering,North University of China,Taiyuan 030051,China)

机构地区：[1]太原工业学院电子工程系,太原030008 [2]中北大学信息与通信工程学院,太原030005

出　　处：《中国生物医学工程学报》2022年第3期320-327,共8页Chinese Journal of Biomedical Engineering

基　　金：山西省重点研发计划项目(201803D121069);山西省高校科技创新项目(2019L0924);太原工业学院青年学科带头人资助项目(Y201701)。

摘　　要：针对调强放疗逆向计划过程中手动调整目标函数权重因子所需时间长、效率低的问题,利用定积分理论构造了各物理准则子目标函数权重因子调整策略,提出一种自动迭代调整权重因子的放疗优化方法。首先,根据基于定积分提出的权重因子调整策略自动迭代调整权重因子。迭代完成后,根据评价准则评估所得计划是否临床接受,如果可以接受,则结束优化过程,否则通过引入额外的补偿系数,赋予未满足剂量目标相关的子目标函数更高惩罚,然后重复上述步骤直到生成满意计划或者达到最大迭代次数。通过与同病例的人工计划相比,从剂量学和生物学两方面比较和评价,在10例前列腺癌病例上验证所提出方法的有效性。实验表明,就DVH曲线和剂量统计特性而言,自动计划较人工计划能够在保证靶区剂量覆盖特性相似的前提下,膀胱的平均剂量、V_(65)、V_(70)、正常组织并发症概率和广义等效均匀剂量分别降低了0.53 Gy、4.6%、3.33%、0.37%和0.22 Gy;直肠的最大剂量、V_(50)和V_(65)分别降低了0.54Gy、0.66%、1.64%。除膀胱V_(75)有显著性差异外(P<0.05),其它统计指标均无显著性差异(P>0.05)。采用自动优化方法1~3 min即可生成满足临床剂量要求的计划,而采用人工试误的方法则需1~3 h。总之,基于定积分的放疗权重因子自动优化方法能够提高放疗效率,生成满意的放疗计划。To solve the problem of time-consuming and low efficiency that exist in the process of manual inverse planning,a strategy adjusting weighting factor for sub-objective function of physical criteria was constructed based on the definite integral theory,and an automatic iterative adjustment method of weighting factors was proposed in this paper.In the automatic method,the weighting factors are automatically and iteratively adjusted first based on proposed penalty strategies.Then,plan evaluation was performed to determine whether the obtained plan was acceptable.If not,a higher penalty was assigned to the unsatisfied objective by multiplying it by a compensation coefficient.The optimization processes were performed alternately until an acceptable plan was obtained or the maximum number of iteration was reached.The effectiveness of the method was verified on 10 prostate cancer cases and compared with the manual planning from the perspective of dosimetry and biology.Experimental results showed that,in terms of DVH curves and dose statistics,the mean dose,V_(65),V_(70),normal tissue complications probability and generalized equivalent uniform dose of bladder were reduced by 0.53 Gy,4.6%,3.33%,0.37%and 0.22 Gy,respectively,under the premise that the dose coverage characteristics of target area were similar.And the maximum dose of rectum,V_(50) and V_(65) decreased by 0.54 Gy,0.66%and 1.64%,respectively.There was no significant difference in other indexes(P>0.05)except V_(75) of bladder(P<0.05);it took 1~3 minutes to produce an acceptable plan applying our proposed automatic optimization method of weighting factors,while for the manual trial-and-error method,it needed professional physicians to take 1~3 hours to obtained an acceptable plan.In conclusion,the automatic optimization method of radiotherapy weighting factors based on definite integral improved the efficiency of radiotherapy and generated satisfactory radiotherapy plans.

关 键 词：权重因子 逆向计划 定积分 自动计划 物理优化 

分 类 号：R318[医药卫生—生物医学工程]