中国癌症杂志 ›› 2019, Vol. 29 ›› Issue (9): 700-708.doi: 10.19401/j.cnki.1007-3639.2019.09.004

• 论著 • 上一篇    下一篇

乳腺癌患者MRI体素内不相干运动成像中感兴趣区选择方法对参数一致性的影响

孙瑞红 1 ,王 翔 1 ,沈丽娟 2 ,孟凡华 3 ,尹化斌 3   

  1. 1. 华中科技大学同济医学院附属武汉中心医院放射科,湖北 武汉 430014 ;
    2. 复旦大学附属肿瘤医院放射诊断科,复旦大学上海医学院肿瘤学系,上海 200032 ;
    3. 复旦大学附属上海市第五人民医院放射科,上海 200240
  • 出版日期:2019-09-30 发布日期:2019-11-20
  • 通信作者: 沈丽娟 E-mail: 13817959783@163.com

Intravoxel incoherent motion parameters of breast cancer at MRI: effect of region of interest on parameter measurement consistency

 SUN Ruihong 1 , WANG Xiang 1 , SHEN Lijuan 2 , MENG Fanhua 3 , YIN Huabin 3   

  1. 1. Department of Radiology, Central Hospital of Wuhan, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430014, Hubei Province, China; 2. Department of Radiology, Fudan University Shanghai Cancer Center, Department of Oncology, Shanghai Medical College, Fudan University, Shanghai 200032, China; 3. Department of Radiology, the Fifth People’s Hospital of Shanghai, Fudan University, Shanghai 200240, China
  • Published:2019-09-30 Online:2019-11-20
  • Contact: SHEN Lijuan E-mail: 13817959783@163.com

摘要: 背景与目的:近年来,基于双指数计算模型的体素内不相干运动(intravoxel incoherent motion,IVIM)成像技术在肿瘤领域的研究越来越多,其定量参数的准确性成为临床关注的重点。探讨乳腺癌患者在IVIM成像中采用不同感兴趣区(region of interest,ROI)选择方法对参数一致性的影响,并分析乳腺腺体类型、微钙化及肿瘤组织学分级对IVIM参数重复性的影响。方法:对华中科技大学同济医学院附属武汉中心医院收治的60例乳腺癌患者行双侧乳腺12个b值(0~1200 s/mm 2 )IVIM检查。2名医师分别在肿瘤最大层面采用肿瘤轮廓法和固定半径法勾画ROI,测量IVIM参数值,包括真实扩散系数(D)、灌注分数(f)、假性扩散系数(D * )。其中医师1间隔1个月后重复上述测量。采用单因素方差分析、Bland-Altman法比较各参数,计算组内相关系数(intraclass correlation coefficients,ICC),并比较乳腺腺体类型、微钙化及肿瘤组织学分级对IVIM参数可重复性的影响。结果:医师1第1次采用肿瘤轮廓法、固定半径法测得的D值、f值、D * 值差异均无统计学意义(P>0.05)。同一医师两次以及不同医师之间采用不同ROI勾画方法所测得的D值、f值可重复性均较好(ICC≥0.773),D * 值可重复性较差(ICC<0.75)。Bland-Altman散点图显示肿瘤轮廓法测定的D值分布更集中,提示肿瘤轮廓法在同一医师及不同医师间的重复测量中一致性区间小,准确性好。微钙化、乳腺腺体类型及肿瘤组织学分级分组间IVIM各参数重复性均较好(ICC≥0.811)。结论:乳腺癌患者采用肿瘤轮廓法与肿瘤固定半径法测得的IVIM参数差异无统计学意义,采用肿瘤轮廓法重复测量一致性评估变异度小,可重复性较好。乳腺腺体类型、微钙化及肿瘤组织学分级对IVIM参数的重复性影响不明显。

Abstract: Background and purpose: Intravoxel incoherent motion (IVIM) based on the theory of biexponential model is increasingly researched in the field of oncology. There is growing interest in the reproducibility of biexponential diffusion parameters. This study aimed to investigate the reproducibility of IVIM parameters of breast cancer based on different regions of interest (ROIs) and the influence of breast density, microcalcifications and histological grade on this reproducibility. Methods: In this retrospective study, 60 patients with breast cancer who treated in Central Hospital of Wuhan, Tongji Medical College, Huazhong University of Science and Technology underwent 3.0 T magnetic resonance imaging (MRI), including an IVIM diffusion-weighted imaging (DWI) examination with 12 different b-values (0-1 200 s/mm 2 ). Two radiologists independently delineated different ROIs that were used to measure IVIM parameters [true diffusion coefficient (D), pseudo-diffusion coefficient (D * ) and perfusion fraction (f)]. Intraclass correlation coefficients (ICC) were calculated to assess measurement reproducibility of the parameters. Variability between readers was assessed by using Bland-Altman 95% limits of agreement and analysis of variance (ANOVA). The effects of breast density, microcalcifications, and histological grade on measurement reproducibility were also assessed. Results: There was no significant difference in the IVIM parameters (D, f and D * ) among the three tumor ROIs. The intra- and interobserver reproducibility were good for D and f (ICC≥0.773). However, D* had poor measurement reproducibility (ICC<0.75). Bland-Altman analysis revealed the narrowest 95% limits of agreement using the outlined-tumor ROI for D between readers. All breast cancer parameters were highly reproducible irrespective of breast density, histological grade or coexistence of microcalcifications (ICC≥0.811). Conclusion: Outlined-tumor ROI was more reproducible with respect to IVIM parameter measurements than the tumor fixed-radius ROI method. Breast density, microcalcifications and histological grade did not significantly affect the repeatability of the IVIM parameters.