自动uARC计划在食管癌同步推量放疗中的可行性研究

Feasibility study of automatic uARC planning for esophageal cancer using simultaneous integrated boost radiotherapy

目的比较联影计划系统中食管癌同步推量自动uARC计划与人工计划的剂量学和放射生物学参数差异,为临床应用提供参考。方法选取100例同步推量食管癌患者的优质uARC计划,统计靶区和危及器官的剂量学参数均值,建立联影计划系统uRT-TPOIS智能计划的优化目标表。另选取21例食管癌病例作为测试病例,使用uRT-TPOIS计划系统分别设计人工计划和自动uARC计划。比较两种计划靶区的平均剂量D_(mean)、PTV的近似最小剂量D_(98%)和近似最大剂量D_(2%)、均匀性指数(HI)、适形指数(CI)、危及器官剂量、平均计划时间、机器跳数(MU)、肿瘤控制概率(TCP)和正常组织并发症概率(NTCP)的差异。符合正态分布的数据采用配对t检验比较两组数据的差异,不符合的采用非参数Wilcoxon检验。结果自动uARC计划的靶区D_(98%)(PTV_(60 Gy):P<0.001,PTV_(54 Gy):P=0.001)、CI(PTV_(60 Gy):P<0.001,PTV_(54 Gy):P=0.002)和处方剂量覆盖靶区体积(V_(54 Gy):P<0.001)均优于人工计划,而D_(mean)和HI的差异无统计学意义[PTV_(54 Gy)(59.32±1.87)∶(59.13±1.64)Gy,(0.19±0.02)∶(0.18±0.02),均为P>0.05];脊髓D_(mean)和D_(max)均优于人工计划[(13.22±4.27)∶(13.75±4.44)Gy,P=0.020;(36.99±1.67)∶(38.14±1.31)Gy,P=0.011];肺V_(20 Gy)两者差异无统计学意义(P>0.05),但肺V_(5 Gy)和V_(10 Gy)均值人工计划均低于自动计划(均为P<0.001)。自动计划和人工计划的平均计划时间分别为(11.79±1.71)min和(53.36±8.23)min,两者差异有统计学意义(P<0.001);和人工计划相比,自动计划没有增加跳数[(762.84±74.83)∶(767.41±80.63)MU,P>0.05]。自动计划提高了靶区的TCP值(PTV_(60 Gy) 89.15%±0.49%∶86.75%±6.46%,P=0.004和PTV_(54 Gy) 79.79%±3.48%∶77.51%±5.04%,P=0.006),增加了肺的NTCP值(0.46%±0.40%∶0.35%±0.32%,P<0.001)。但心脏和脊髓的NTCP值,两种计划之间的差异无统计学意义(均为P>0.05)。结论使用联影智能计划模块实现食管癌自动uARC计划具有临床可行性,并且能够提高靶区适形性,缩短计划设计时间。

Objective To compare dosimetric and radiobiological parameters between automatic and manual uARC plans in the treatment of esophageal cancer patients,aiming to provide reference for clinical application.Methods High-quality uARC plans of 100 patients with esophageal cancer were selected,and the mean values of the dosimetric parameters in the target area and organs at risk(OAR)were counted,and the goal table of uRT-TPOIS intelligent plan was established.Automatic and manual uARC plans were generated with UIH(United Imaging)treatment planning system(TPS)for 21 esophageal cancer patients.The differences in mean dose(D_(mean)),approximate minimum(D_(98%))and maximum(D_(2%))dose of planning target volume(PTV),homogeneity index(HI)and conformity index(CI),dose of OAR,mean planning time,monitor unit(MU),tumor control probability(TCP)and normal tissue complication probability(NTCP)were compared between automatic and manual uARC plans.Normally distributed data between two groups were compared by paired t-test,and non-normally distributed data were assessed by nonparametric Wilcoxon test.Results The D_(98%)(PTV_(60 Gy):P<0.001,PTV_(54 Gy):P=0.001),CI(PTV_(60 Gy):P<0.001,PTV_(54 Gy):P=0.002)and target volume of area covered by prescription dose(V_(54 Gy):P<0.001)of the automatic uARC plans were better than those of manual uARC plans(all P<0.05).There was no significant difference in Dmean or HI between the two plans[PTV_(54 Gy)(59.32±1.87)Gy vs.(59.13±1.64)Gy,(0.19±0.02)vs.(0.18±0.02),all P>0.05].The D_(mean) and D_(max) of spinal cord of the automatic plan were better than those of the manual plan[(13.22±4.27)Gy vs.(13.75±4.44)Gy,P=0.020 and(36.99±1.67)Gy vs.(38.14±1.31)Gy,P=0.011].There was no significant difference in the mean dose of V_(20 Gy) of the lung between two plans(P>0.05),whereas the mean doses of V_(5 Gy) and V_(10 Gy) of the lung of the manual plan were less than those of the automatic plan(both P<0.001).Automatic uARC plan had a significantly shorter mean planning time than manual uARC plan[(11.79±1.71)min vs.(53.36±8.23)min,P<0.001].MU did not significantly differ between two plans[(762.84±74.83)MU vs.(767.41±80.63)MU,P>0.05].The TCP of the automatic plan was higher than that of the manual plan(PTV_(60 Gy) 89.15%±0.49%vs.86.75%±6.46%,P=0.004 and PTV_(54 Gy) 79.79%±3.48%vs.77.51%±5.04%,P=0.006).However,manual plan had a lower NTCP of the lung than automatic uARC plan(0.46%±0.40%vs.0.35%±0.32%,P<0.001).There was no significant difference in NTCP of heart and spinal cord between two plans(all P>0.05).Conclusion It is feasible to generate automatic uARC plan with uRT-TPOIS TPS for esophageal cancer patients,which can increase the target CI and shorten the plan design time.