利用四维锥形束CT指导肺癌内靶区的勾画

目的 研究四维锥形束CT(4D-CBCT)指导肺癌内靶区勾画的可行性。方法 简单随机法选取本院24例肺癌患者。平静呼吸下CT模拟定位获得CT图像,首次治疗前4D-CBCT扫描获得4D-CBCT中位图像,将4D图像重建算法更改为3D,获得3D-CBCT图像。图像融合算法取骨配准,分别在定位CT、4D-CBCT中位图像、3D-CBCT图像上勾画大体肿瘤靶区(GTV),定义GTV到临床靶区(CTV)的外扩边界为7 mm,获得CTV_CT、ITV_4D和ITV_3D,基于CTV_CT在三维方向上外扩5 mm得到ITV_CT。比较靶区间中心点位置、体积、相似度和相互包含关系的差异。结果 ITV_CT与ITV_3D在中下叶组的头脚方向上,靶区中心点位置差异有统计学意义(Z=-2.027,P<0.05)。在靶区体积方面,ITV_CT最大,与ITV_3D相比差异有统计学意义(Z=-2.941,P<0.05),ITV_4D最小,但与ITV_3D相比差异无统计学意义(P>0.05)。ITV_CT与ITV_3D相似度均数<75%,ITV_4D与ITV_3D相似度均数>90%(Z=-2.940、-2.975,P<0.05)。ITV_CT、ITV_4D未被ITV_3D包含的比例均数为40%和5%(Z=-2.952、-3.185,P<0.05)。结论 4D-CBCT的中位图像可以缩小内靶区的勾画范围,为肺癌的自适应放疗提供选择。

Feasibility of delineation internal target volumes of lung cancer using four-dimensional cone-beam CT

Objective To study the feasibility of contouring internal target volumes of lung cancer using four-dimensional cone-beam CT(4D-CBCT). Methods Twenty-four cases of lung cancer coming from Nanjing Medical University First Affiliated Hospital were selected with simple random sampling. Locating CT images were obtained with simulating CT under quiet breathing,and the mid-position image of 4D-CBCT were obtained in the first fraction, three-dimensional cone-beam CT(3D-CBCT) were obtained by changing image reconstruction algorithm from four-dimensional to three-dimensional. Register the planning images using the bone registration. GTVs were contoured respectively based on CT, mid-position image of 4D-CBCT and 3D-CBCT. CTV_CT, ITV_4D and ITV_3D were defined with a margin of 7 mm accounting for microscopic disease. ITV_CT were defined based on CTV_CT with an isotropic internal margin of 5 mm. The differences in the position, volume, Dice's similarity coefficient (DSC) and inclusion relationship of different volumes were compared. Results In the direction of CC,the differences between ITV_CT and ITV_3D for the middle-lower lobe had statistical significance in the center of target location (Z=-2.027, P<0.05). ITV_CT was the biggest and ITV_4D was the smallest in terms of volume, and the former had statistical significance compared with ITV_3D (Z=-2.941, P<0.05), while the latter had no statistical significance compared with ITV_3D (P>0.05). The mean of DSC between ITV_CT and ITV_3D was less than 75%, while the one between ITV_4D and ITV_3D was more than 90% (Z=-2.940, -2.975, P<0.05). The mean percentages of ITV_CT and ITV_4D excluding ITV_3D were approximately 40% and 5% respectively (Z=-2.952, -3.185, P<0.05). Conclusions The volume of internal target volumes of lung cancer can be reduced using the mid-position image of 4D-CBCT, which might provide an alternative of individualized adaptive radiotherapy for lung cancer.