Monaco的蒙特卡罗算法网格间距和不确定度对γ通过率的影响

Effects of different grid spacing and statistical uncertainty in MC algorithm of Monaco TPS on gamma pass rate

目的比较Monaco系统的蒙特卡罗(MC)算法网格间距(GS)和不确定度(SU)对临床计划验证γ通过率的影响,为日常放疗患者的质量保证(PSQA)提供参考。方法回顾性选择2023年7—11月中国医学科学院肿瘤医院深圳医院已治疗的20例患者计划,其中头部鼻咽癌5例,胸部肺癌2例、食管癌3例、乳腺癌5例,腹部宫颈癌4例、直肠癌1例。所有患者的计划均在同一天由同一物理师在同一台机器重新测量得到测量的剂量分布文件。Monaco系统的GS选取2、3、4 mm共计3种(分别为GS2、GS3、GS4),SU选取SU_(CP1)、SU_(CP2)、SU_(CP3)、SU_(CP4)、SU_(CL1)共计5种百分比不确定度,分别重新计算验证计划,每例临床计划总计15个验证计划。采用3%/2 mm的评价标准分析γ通过率,统计每个计划不同GS和SU的γ通过率。采用双侧配对样本t检验分析同一病例计划不同GS和SU的γ通过率。结果所有患者计划,以GS2的γ通过率为基准,相同SU不同GS的γ通过率差异均具有统计学意义(P<0.05)。对于SU_(CP1),GS3、GS4的γ通过率比GS2分别降低0.4%、1.4%;对于SU_(CP2),GS3、GS4的γ通过率比GS2分别降低0.5%、1.5%;对于SU_(CP3),GS3、GS4的γ通过率比GS2分别降低0.5%、1.5%;对于SU_(CP4),GS3、GS4的γ通过率比GS2分别降低0.5%、1.5%;对于SU_(CL1),GS3、GS4的γ通过率比GS2分别降低0.7%、2.0%。以SU_(CP1)的γ通过率为基准,相同GS不同SU,在GS2的时候,本研究所选取的SU其γ通过率差异均无统计学意义,在GS3和GS4时,SU_(CP4)和SU_(CL1)相比SU_(CP1)差异具有统计学意义(GS3时,P=0.049和0.012;GS4时,P=0.045和<0.001),γ通过率分别降低0.1%、0.4%,0.2%、0.6%。结论GS和SU值均在一定程度上影响Monaco系统的MC算法的临床计划验证γ通过率,推荐Monaco系统的MC算法日常PSQA,选取GS2,SU选取SU_(CL1)计算验证计划。

Objective To compare the impact of Monte Carlo(MC)algorithm grid spacing(GS)and statistical uncertainty(SU)of Monaco on clinical plan validation gamma pass rate,and to provide reference for daily patient-specific quality assurance(PSQA).Methods Twenty patients treated in Chinese Academy of Medical Sciences and Peking Union Medical College Cancer Hospital&Shenzhen Hospital from July to November 2023 were retrospectively selected,including 5 cases of nasopharyngeal carcinoma in the head,2 cases of lung cancer,3 cases of esophageal cancer,5 cases of breast cancer in the chest,4 cases of cervical cancer,1 case of rectal cancer in the abdomen,respectively.All selected patient plans were re-measured on the same day by the same physicist on the same machine to obtain dose distribution files.Three types of GS of Monaco,including 2 mm,3 mm,and 4 mm(GS2、GS3、GS4),and 5 types percentage of SU,including SU_(CP1),SU_(CP2),SU_(CP3),SU_(CP4),and SU_(CL1) were selected.The validation plans were recalculated,with a total of 15 validation plans for each clinical plan.Using a 3%/2 mm evaluation standard,the gamma pass rates of each plan at different GS and SU were analyzed.The gamma pass rates of different GS and SU in the same case plan were analyzed by paired sample t-test.Results Based on the gamma pass rate of GS2,the differences in gamma pass rates between different GS for the same SU were statistically significant(all P<0.05).For SU_(CP1),the gamma pass rates of GS3 and GS4 were decreased by 0.4% and 1.4% compared to GS2,respectively.For SU_(CP2),the gamma pass rates of GS3 and GS4 were decreased by 0.5% and 1.5% compared to GS2,respectively.For SU_(CP3),the gamma pass rates of GS3 and GS4 were decreased by 0.5% and 1.5% compared to GS2,respectively.For SU_(CP4),the gamma pass rates of GS3 and GS4 were decreased by 0.5% and 1.5% compared to GS2,respectively.For SU_(CL1),the gamma pass rates of GS3 and GS4 were decreased by 0.7% and 2.0% compared to GS2,respectively.Based on the gamma pass rate of SU_(CP1),for the same GS but different SU,there was no statistically significant difference in the gamma pass rate of the SU selected in this study at GS2.However,at GS3 and GS4,the difference between SU_(CP4) and SU_(CL1) compared to SU_(CP1) was statistically significant(at GS3,P=0.049 and 0.012;at GS4,P=0.045 and<0.001),with gamma pass rates reduced by 0.1%,0.4%,0.2% and 0.6%,respectively.Conclusions Both GS and SU values affect the gamma pass rate of PSQA to a certain extent.It is recommended to use Monaco MC algorithm for daily PSQA,selecting GS2 and SU with SU_(CL1) to calculate the validation plan.