肺癌锥形束CT图像不同配准方式的误差分析

目的 研究肺癌锥形束CT(CBCT)图像配准的影响因素.方法 选取2007年间本科采用CBCT作为在线校位的肺癌放疗患者20例.每位患者在疗程中6～19次治疗采集了CBCT图像.根据肺内病变位置与椎体关系分为病变靠近椎体组(A组)和病变远离锥体组(B组),对比两组摆位误差差异.同时对该组病例分别选取从治疗初期、中期和后期CBCT图像,请4位医生分别采用骨配准和灰度配准,比较不同医生间和不同配准方式间差异.结果 A组和B组在头脚、左右、前后方向上的摆位误差分别为-1.31、1.24、-1.88 mm和0.10、1.37、-1.26 mm(t=0.07、0.05、-0.12,P=0.554、0.652、0.321).4位医生骨配准摆位误差头脚方向分别为-0.05、-0.01、0.05、-0.16(F=-0.01,P=0.887),左右方向分别为0.56、0.35、0.51、0.43 mm(F=-0.01,P=0.880),前后方向分别为-1.16、-1.20、-0.88、-1.03 mm(F=0.04,P=0.555);灰度配准摆位误差头脚方向分别为-0.32、-0.34、-0.39、-0.37 mm(F=-0.01,P=0.874),左右方向分别为0.34、0.54、-0.04、0.27 mm(F=-0.03,P=0.622),前后方向分别为-1.12、-1.15、-1.13、-1.04 mm(F=0.00,P=0.812).结论 采用相同配准框和图像质量下,肺癌患者肺内病变位置、不同配准方式和不同医生对图像引导放疗中CBCT图像配准无明显影响.

Abstract：

Objective To analyze the influencing factors of cone-beam CT (CBCT) imagine registration in lung cancer. Methods From Mar. 2007 to Dec. 2007, 20 patients with lung cancer were treated with IGRT. The imagines of CBCT were collected from 6 to 19 fractions during the patients' radiotherapy. To compare the difference of set-up errors between the two groups according to the distance from the lesion in lung to the centrum. At the same time, CBCT imagines from the first, middle and the last fraction of these patients' radiotherapy were registrated in bone and grey methods by four doctors. The difference of set-up errors between different doctors and registrated methods were compared. Results The mean values of set-up errors were ＜2 mm in the two groups without significant difference (x:-1.31mm vs 0. 10 mm (t=0. 07,P=0.554);y:1.24 mm vs 1.37 mm (t=0. 05,P=0. 652);z: - 1.88mm vs -1.26mm (t= -0. 12,P=0.321)). The mean values of set-up errors were ＜ 1.3 mm in four doctors and registrated methods without significant difference, for bone registration,x: -0. 05 mm, -0. 01 mm,0. 05 mm, -0.12 mm and -1.31 mm ( F=-0.01,P=0.887) ;y:0.56 mm,0.35 mm,0.51 mm and 0.43 mm (F= -0.01,P=0.880);z: -1.16 mm, -1.20 mm, -0.88 mm and -1.03 mm (F= -0.04,P=0. 555 ), for grey registration ,x: -0.32 mm, -0.341 mm, -0.395 mm and - 0.37 mm(F=-0.01, P=0.874);y:0.34 mm,0.54 mm, -0.04 mm and 0.27 mm (F= -0.03,P=0.622);x:-1.12 mm,- 1.15 mm, - 1.13 mm and - 1.04 mm (F=0. 00,P=0. 812). Conclusions With the same registrated box and imagine quality, the location of the lesions in lung, registred methods and different doctors are not the influencing factors for CBCT imagine registration.

Cone-beam CT imagine registration of lung cancer

Objective To analyze the influencing factors of cone-beam CT (CBCT) imagine registration in lung cancer. Methods From Mar. 2007 to Dec. 2007, 20 patients with lung cancer were treated with IGRT. The imagines of CBCT were collected from 6 to 19 fractions during the patients' radiotherapy. To compare the difference of set-up errors between the two groups according to the distance from the lesion in lung to the centrum. At the same time, CBCT imagines from the first, middle and the last fraction of these patients' radiotherapy were registrated in bone and grey methods by four doctors. The difference of set-up errors between different doctors and registrated methods were compared. Results The mean values of set-up errors were ＜2 mm in the two groups without significant difference (x:-1.31mm vs 0. 10 mm (t=0. 07,P=0.554);y:1.24 mm vs 1.37 mm (t=0. 05,P=0. 652);z: - 1.88mm vs -1.26mm (t= -0. 12,P=0.321)). The mean values of set-up errors were ＜ 1.3 mm in four doctors and registrated methods without significant difference, for bone registration,x: -0. 05 mm, -0. 01 mm,0. 05 mm, -0.12 mm and -1.31 mm ( F=-0.01,P=0.887) ;y:0.56 mm,0.35 mm,0.51 mm and 0.43 mm (F= -0.01,P=0.880);z: -1.16 mm, -1.20 mm, -0.88 mm and -1.03 mm (F= -0.04,P=0. 555 ), for grey registration ,x: -0.32 mm, -0.341 mm, -0.395 mm and - 0.37 mm(F=-0.01, P=0.874);y:0.34 mm,0.54 mm, -0.04 mm and 0.27 mm (F= -0.03,P=0.622);x:-1.12 mm,- 1.15 mm, - 1.13 mm and - 1.04 mm (F=0. 00,P=0. 812). Conclusions With the same registrated box and imagine quality, the location of the lesions in lung, registred methods and different doctors are not the influencing factors for CBCT imagine registration.