目的 探讨使用不同模态MR成像进行脑胶质瘤放疗靶区勾画的一致性。方法 对33例胶质瘤术后放疗患者进行回顾性 分析，分别在T2WI、FLAIR以及DWI模态上手动勾画肿瘤靶区，通过面积求和法计算临床靶体积（Clinical Target Volume， CTV）。以随机区组方差分析法比较3组MR模态下CTV的差异；以Bland-Atlman法评估3组MR模态下勾画胶质瘤CTV的95%一 致性界限（95% Limits of Agreement, 95%LoA）。结果 T2WI、FLAIR以及DWI不同模态下的CTV无统计学差异（F＝0.088，P＝ 0.916）。Bland-Atlman法结果表明：T2WI组与FLAIR组、DWI组比较的95%LoA分别为-78.3~98.4 mm3、-91.2~114.1 mm3；FLAIR 组与DWI组的95%LoA为-101.3~98.4 mm3；均存在6.06%（2/33）的点位于95%LoA以外，超出临床可接受范围。结论 不同模态MR 图像测量胶质瘤术后放疗CTV一致性欠佳，尚不能相互取代，临床应用时应注意不同模态的联合分析。

Objective To explore the consistency of different modality MR imaging in target delineation of brain glioma radiotherapy. Methods A total of 33 patients with postoperative radiotherapy for glioma were retrospectively analyzed. Tumor target areas were manually delineated on T2WI, FLAIR and DWI modes, respectively. Clinical Target Volume (CTV) was calculated by area summation method. The differences of CTV among the 3 groups using different MR mode were compared by randomized blocks analysis of variance. The 95% limits of agreement (95%LoA) of CTV between the 3 groups were assessed by Bland-Atlman method. Results There was no significant difference in CTV among different modes of T2WI, FLAIR and DWI（F＝0.088，P＝0.916）.The 95% LoA of T2WI and FLAIR, T2WI and DWI, FLAIR and DWI were （-78.3-98.4）mm3,（-91.2-114.1）mm3 and（-101.3-98.4）mm3, respectively. There were 6.06% (2/33) points outside 95%LoA, beyond the clinically acceptable range. Conclusion The consistency of CTV measurment using different modality MR images for gliomas postoperative radiotherapy is not good and can not be replaced by each other. Clinical application should pay more attention to the joint analysis of different modes.