卢宝兰,肖晓娟,杨心悦,陈琰,文自强,余深平.直肠癌体素内不相干运动MRI参数的一致性和可重复性[J].中国医学影像技术,2018,34(4):568~572
直肠癌体素内不相干运动MRI参数的一致性和可重复性
Repeatability and consistency of intravoxel incoherent motion parameters of rectal cancers
投稿时间:2017-10-19  修订日期:2018-01-27
DOI:10.13929/j.1003-3289.201710075
中文关键词:  直肠肿瘤  磁共振成像  体素内不相干运动  一致性  可重复性,结果
英文关键词:Rectal neoplasms  Magnetic resonance imaging  Intravoxel incoherent motion  Consistency  Repeatability of results
基金项目:广东省科技计划项目(2014A020212126)。
作者单位E-mail
卢宝兰 中山大学附属第一医院放射科, 广东 广州 510080  
肖晓娟 北京大学深圳医院医学影像科, 广东 深圳 518036  
杨心悦 中山大学附属第一医院放射科, 广东 广州 510080  
陈琰 中山大学附属第一医院放射科, 广东 广州 510080  
文自强 中山大学附属第一医院放射科, 广东 广州 510080  
余深平 中山大学附属第一医院放射科, 广东 广州 510080 yuquntogether@163.com 
摘要点击次数: 1534
全文下载次数: 779
中文摘要:
      目的 探讨直肠癌体素内不相干运动(IVIM)MRI参数测量的一致性和可重复性。方法 对128例经病理证实的直肠腺癌患者于治疗前行直肠常规高分辨MR和IVIM扫描。选择两种b值组合(A组:0、5、10、20、30、40、60、80、100、150、200、400、600和1 000 s/mm2;B组:0、5、10、20、30、40、60、80、100、150、200、400、600、1 000、1 500和2 000 s/mm2),拟合计算出两组IVIM参数图,分别测量肿瘤区域单纯扩散系数(D)、假性扩散系数(D*)和灌注分数(f)值。由同一观察者间隔3个月重复测量肿瘤区域D、D*和f值;采用组内相关系数(ICC)和Bland-Altman图评价测量的一致性和可重复性。结果 两种b值组合图像所测量D、D*、f值差异均有统计学意义(P均< 0.001);D、D*及f值的ICC及95%置信区间分别为0.968(0.955,0.977)、0.780(0.688,0.845)和0.957(0.934,0.970);Bland-Altman图显示差异百分比的±1.96标准差分别为(10.8%,22.4%)、(14.8%,61.9%)、(-45.3%,-10.2%)。同一观察者2次测量D*值差异有统计学意义(P=0.001);D、D*及f值的ICC及95%置信区间分别为0.826(0.670,0.908)、0.678(0.392,0.830)和0.910(0.830,0.952);Bland-Altman图显示差异百分比的±1.96标准差分别为(-15.3%,12.4%)、(-39.6%,61.2%)、(-22.6%,22.9%)。结论 两种b值组合图像测量直肠癌各IVIM参数的一致性良好,D值和f值测量的可重复性良好。
英文摘要:
      Objective To explore the repeatability and consistency of MR intravoxel incoherent motion (IVIM) parameters of rectal cancers. Methods Routine high resolution rectal MR and IVIM sequence were performed on 128 patients with pathologically proved rectal cancers before treatment. IVIM maps were generated from two sets of b-values (group A:0, 5, 10, 20, 30, 40, 60, 80, 100, 150, 200, 400, 600 and 1 000 s/mm2; group B:0, 5, 10, 20, 30, 40, 60, 80, 100, 150, 200, 400, 600, 1 000, 1 500 and 2 000 s/mm2), and IVIM parameters (D, D* and f) were measured. Then IVIM parameters were measured again by the same observer three months later. The intra-class correlation coefficient (ICC) and Bland-Altman graph analysis were performed to explore the repeatability and consistency of IVIM parameters. Results The differences of IVIM parameters (D, D* and f) derived from the two sets of b-values were statistically different (all P<0.001). ICC and 95% confidence interval (CI) of D, D* and f was 0.968 (0.955, 0.977), 0.780 (0.688, 0.845) and 0.957 (0.934, 0.970), respectively. Bland-Altman analysis showed that the 95% limits of agreement of D, D* and f was (10.8%, 22.4%), (14.8%, 61.9%) and (-45.3%, -10.2%), respectively. There was significant difference of D* between twice measurement by the same observer (P=0.001); ICC and 95%CI of D, D* and f was 0.826 (0.670, 0.908), 0.678 (0.392, 0.830) and 0.910 (0.830, 0.952), respectively. Furthermore, Bland-Altman analysis showed that the 95% limits of agreement of D, D* and f were (-15.3%, 12.4%), (-39.6%, 61.2%) and (-22.6%, 22.9%), respectively. Conclusion Under the two different sets of b-values, all IVIM parameters of rectal cancers demonstrate good consistency. In addition, D and f value show good repeatability.
查看全文  查看/发表评论  下载PDF阅读器