宫颈癌术后放疗中自动计划与人工计划的剂量学比较

Dosimetric comparison between automated and manual volumetric modulated arc therapy planning for postoperative cervical cancer

目的 比较容积弧形调强（VMAT）人工计划与自动计划的剂量学参数，评估自动计划在临床宫颈癌术后放疗应用的可行性以及剂量学特点。方法 选取23例宫颈癌术后病例，其中ⅡA期8例，ⅡB期15例，使用Pinnacle3放射治疗计划系统分别设计人工VMAT和自动VMAT计划。比较两种计划靶区的Dmean、D95、适形指数（CI）、均匀性指数（HI）、危及器官（OAR）的剂量体积直方图（DVH）、平均优化时间、调试次数以及机器跳数（MU）。结果 自动计划得到的靶区Dmean、CI以及HI均优于人工计划（t=4.65～14.92，P〈0.05），而D95无明显差异（P〉0.05）；自动计划得到的OAR各参数均优于人工计划（t=3.30～14.42，P〈0.05）；自动计划明显减少了计划的平均优化时间（72 min，t=3.85，P〈0.05）和调试次数（2次，t=5.41，P〈0.05）；自动计划和人工计划的平均机器跳数分别为（819±53）和（638±41） MUs，自动计划跳数增加了181 MUs。结论 使用Pinnacle3自动设计的宫颈癌术后VMAT计划具有临床可行性，并且自动VMAT计划与人工VMAT计划相比提高了计划质量，减少了计划设计时间、人为因素对于计划质量的影响。

Objective To compare dosimetric parameters between automated and manual volumetric modulated arc therapy（VMAT） plans in the treatment of postoperative cervical cancer patients, and to investigatethe feasibility and dosimetric advantage of the automated VMAT planning. Methods Automated and manual VMAT plans were generated with Pinnacle3 treatment planning system（TPS）for twenty-three postoperative cervical cancer patients, including eight patients in stage ⅡA and fifteen in stage ⅡB, respectively. The differences in Dmean, D95,conformity index （CI） and homogeneity index （HI） of target, as well as dose volume histogram （DVH） of organs at risk （OAR）,planning time, average optimization time and monitor unit （MU） were compared between automated and manual VMAT plans. Results The average Dmean, CI and HI of automated VMAT plans were better than those of manual VMAT plans （t=4.65-14.92, P〈0.05）. There was no significant difference in D95 （P〉0.05）. The automated VMAT plans achieved better average dosimetric parameters on OARs compared with the manual VMAT plans （t=3.30-14.42, P〈0.05）. Automated VMAT plans had a significantly shorter planning time （72 min, t=3.85, P〈0.05） and interruption frequency （twice, t=5.41, P〈0.05） than manual VMAT plans. However, automated VMAT plans had a higher average MU than manual VMAT plans with an average MU of 819±53 and 638±41 for automated and manual VMAT plans, respectively. Conclusions It is feasible to generate automated VMAT plans with Pinnacle3 TPS for postoperative cervical cancer patients. The automated VMAT plans increase the plan quality and reduce the optimization time compare with manual VMAT plans. Automated technique also eliminates the influence of human factors on the plan quality.