非基于先验知识的胃癌容积弧形调强治疗自动计划

Non-knowledge-based auto-planning of volumetric modulated arc therapy for gastric cancer

目的:评估非基于先验知识的容积弧形调强治疗(VMAT)自动计划在胃癌放疗中的可行性。方法:回顾性地收集胃癌术后放疗患者30例。使用基于Python语言开发的自动VMAT计划程序设计胃癌放疗计划。所有目标参数设置和优化基于处方和目标函数值而非历史计划;然后对比自动计划和人工计划的剂量学参数、计划时间。结果:人工计划与自动计划均能满足靶区剂量覆盖。对比人工计划,自动计划的脊髓、小肠和肾脏的吸收剂量均显著降低(P<0.05)。然而,两种计划中十二指肠、肝脏等邻近靶区的危及器官的吸收剂量差异无统计学意义(P>0.05)。此外,自动计划的中位计划时间较人工计划减少23.9 min(P<0.05)。结论:对于胃癌,非基于先验知识的自动计划能在满足靶区剂量覆盖的同时显著降低远离靶区的危及器官放疗剂量,提高计划效率。

Objective To evaluate the feasibility of non-knowledge-based auto-planning of volumetric modulated arc therapy(VMAT)in radiotherapy for gastric cancer.Methods Thirty cases of gastric cancer receiving postoperative radiotherapy were collected retrospectively.The automatic VMAT planning program developed based on Python language was used to design the radiotherapy plan of gastric cancer.The objective parameters setting and optimization were based on prescription and objective function values instead of historical plans.Then,the dosimetric parameters and planning time of automatic plans and manual plans were compared.Results Both of manual and automatic plans could meet the target dose coverage.Compared with those in manual plans,the absorbed dose of spinal cord,small intestine and kidney in automatic plans were significantly decreased(P<0.05).However,there was trivial difference in absorbed dose between two kinds of plans for the organs-at-risk close to the target area such as duodenum and liver(P>0.05).In addition,the median planning time of autoplanning was 23.9 min less than that of manual planning(P<0.05).Conclusion For gastric cancer,the non-knowledge-based auto-planning can not only meet the target dose coverage,but also significantly reduce the radiation dose to organs-at-risk far away from the target area and improve the planning efficiency.