评估靶区与危及器官重叠体积对宫颈癌经验引导调强计划自动优化的影响

Effect of the overlapping volume of target area and organs-at-risk on the automatic optimization of knowledge-based intensity-modulated radiotherapy plan for cervical cancer

目的:通过比较20例宫颈癌术后患者(计划靶区部分与膀胱、直肠重叠)经验引导调强计划(K-IMRT)和传统调强计划(C-IMRT)的剂量学差异,探讨宫颈癌K-IMRT自动优化的可行性。方法:选取接受调强放射治疗的110例宫颈癌患者,100例患者作为模型组进行自动优化模型的建立,模型组外10例靶区与危及器官有重叠体积患者作为试验组E1(开环验证),模型组内10例靶区与危及器官有重叠体积患者作为实验组E2(闭环验证)。E1、E2组利用自动优化模型生成K-IMRT计划,分别比较两组计划与传统优化计划的剂量学差异。结果:两组靶区剂量参数相比,E1组各项靶区剂量参数及E2组最小剂量、HI差异均无统计学意义(P>0.05);E2组K-IMRT计划的最大剂量、CI明显好于C-IMRT计划,差异有统计学意义(P<0.05)。危及器官各参数相比,E1组直肠V20,左右股骨头V_(15)、V_(20)、V_(25);E2组膀胱V_(25)、V_(30)、V_(35)、D_(50),直肠V_(25)、V_(30),左右股骨头V_(15)、V_(20)、V_(25)的K-IMRT计划较C-IMRT计划明显减少,差异有统计学意义(P<0.05)。MU及计划优化时间,K-IMRT计划较C-IMRT计划明显减少。结论:相对于宫颈癌靶区与危及器官相重叠的患者,K-IMRT计划自动优化是可行的。

Objective To evaluate the feasibility of the automatic optimization of knowledge-based intensity-modulated radiotherapy（K-IMRT） plan for cervical cancer by comparing the dosimetric differences between K-IMRT plan and conventional intensitymodulated radiotherapy（C-IMRT） plan for 20 patients receiving radiotherapy for cervical cancer（part of target area overlaps with the bladder and the rectum）. Methods Among the 110 selected cervical cancer patient undergoing radiotherapy, 100 patients served as model group for the establishment of automatic optimization model, and the another 10 cases with overlapping volume of target area and organs-at-risk（OAR） were taken as experimental group E1（open-loop verification）. In model group, 10 patients with overlapping volume of target area and OAR were selected as experimental group E2（closed-loop verification）. In both E1 and E2 groups, the automatic optimization model was used to generate K-IMRT plans which were then compared with C-IMRT plans. Results The target dose parameters in group E1 and the minimum dose and homogeneity index in group E2 didn＇t showed any significant differences between C-IMRT plans and K-IMRT plans（P0.05）. In group E2, the maximum dose and conformity index of K-IMRT plans were significantly better than those of C-IMRT plans, with statistical differences（P〉0.05）. The comparison of OAR parameters showed that compared with C-IMRT plans, K-IMRT plans showed markedly lower V_（20） of the rectal, V_（15）, V_（20）,V_（25） of the left and right femoral head in group E1, and had significant reductions in the V_（25）, V_（30）, V_（35）, D_（50） of the bladder, the V_（25）,V_（30） of the rectum, the V_（15）, V_（20）, V_（25） of the left and right femoral head in group E2（P〈0.05）. Furthermore, K-IMRT plans had a decreased monitor unit and optimization time as comparison with C-IMRT plans. Conclusion The automatic optimization of K-IMRT plan is feasible for cervical cancer patients with overlapping volume of target area and OAR.