【摘要】 目的探讨基于4D-CT不同时相结合肺转移瘤内靶区勾画的差异。方法选择20例行4D-CT模拟定位扫描的肺转移瘤患者,按不同时相(0%～90%)、最大密度投影(MIP)和平均方式进行图像重建,分别基于10个时相叠加形成(A方式,参考靶区)、基于最大密度投影图像勾画(B方式)、基于CT0(吸气末)和CT50(呼气末)时相叠加形成(C方式)、基于CT0、CT20和CT50相叠加形成(D方式)、基于CT0、CT20、CT50和CT70相叠加形成(E方式)和基于平均时相勾画的靶区按前后左右各3 mm和头脚方向5 mm方式外扩形成(F方式)内靶区,用重合度指标、形状相似性系数和Jaccard系数评估B、C、D、E、F方式与A方式的内靶区差异。结果 B、C、D、E和F方式5种内靶区创建方式与参考靶区重合度指标分别为0.81±0.08,0.80±0.10,0.85±0.08,0.88±0.08,0.93±0.07;形状相似性系数分别为0.88±0.05,0.88±0.06,0.92±0.05,0.93±0.05,0.73±0.08;Jaccard系数0.79±0.10,0.80±0.10,0.85±0.08,0.86±0.08,0.57±0.10;其中B方式与C方式靶区重合度指数、形状相似性系数和J系数没有统计学差异(P=0.486,0.863,0.803)。结论基于两极限呼吸和CT20时相创建内靶区可实现勾画质量和勾画效果的平衡。更多还原

【Abstract】 Objective To investigate the differences in internal target volume(ITV)for lung metastases based on 4 D-CT.Methods 20 patients with lung metastases were simulated in 4 D-CT type in this study, the images were reconstructed according to the respiratory phases, maximum density projection(MIP)and average images.The ITV were delineated on different ways, the fusion of 10 phase images and regarded as the reference ITV(mode A),MIP images(Mode B),the fusion of CT0(end inhale phase)and CT50(end exhale phase)(Modes C),the fusion of CT0,CT20,and CT50(Mode D),the fusion of CT0,CT20,CT50 and CT70(Modes E),expanded on average images.The differences between the ITV of the B,C,D,E and F methods and reference ITV were evaluated by the overlap ratio(OR),Dice-similarity coefficients(DSC),and Jaccard index.Results The OR of the ITV created by B,C,D,E and F methods and the reference target were 0.81±0.08,0.80±0.10,0.85±0.08,0.88±0.08,and 0.93±0.07,respectively.DSC and Jaccard index were 0.88±0.05,0.88±0.06,0.92±0.05,0.93±0.05,0.73±0.08,and 0.79±0.10,0.80±0.10,0.85±0.08,0.86±0.08,0.57±0.10,respectively.The differences between mode B and mode C on the OR,DSC and Jaccard index were no statistically significant(P=0.486,0.863,0.803).Conclusion The ITV created based on the two limiting respiration time phases and CT20 time phase can achieve a balance between outline quality and outline effect.更多还原