摘要
目的比较体素内不相干运动(IVIM)模型和单指数模型在肾脏DWI后处理分析过程中预测肾小球滤过率的价值。材料与方法共纳入56例因蛋白尿或持续微量蛋白尿就诊的患者。根据估计肾小球滤过率(e GFR)将56例患者分为:明显肾小球滤过功能损伤组(s RI,e GFR≤30 ml/min/1.73 m^2)和非明显肾小球滤过功能损伤组(non-s RI)。56例患者均完成7个b值的DWI序列扫描。多b值DWI图像由1名放射科医师同时行IVIM模型和单指数模型分析,分别得到三个IVIM参数[慢、快两种扩散系数(D_(slow)、D_(fast))及快扩散系数的分数(F_(fast))]和一个单指数模型参数(D_(mon))。选择右肾的最大横断面,将感兴趣区域尽可能地囊括该层面的肾实质区。结果 s RI组的D_(mon)、D_(slow)、D_(fast)和F_(fast)分别依次为(1.961±0.173)×10^(-3) mm^2/s、(1.747±0.153)×10^(-3) mm^2/s、(3.481±0.690)×10^(-3) mm^2/s和(19.000±4.010)%,non-s RI组的D_(mon)、D_(slow)、D_(fast)和F_(fast)分别依次为(2.175±0.165)×10^(-3) mm^2/s、(1.917±0.162)×10^(-3) mm^2/s、(4.210±0.718)×10^(-3) mm^2/s和(23.110±2.809)%,两组间均存在统计学差异(t≥3.793,P<0.05),四个参数也均与e GFR呈线性相关性(t≥0.356,P<0.05)。当区分s RI和non-s RI组时,D_(mon)、D_(slow)、D_(fast)和F_(fast)的受试者工作特征曲线下面积分别依次0.796、0.786、0.773和0.806,相互之间无明显统计学差异(Z≤0.482,P>0.05)。结论基于单指数模型分析的D_(mon)包含组织微灌注和水分子扩散加权信息,更具有评价肾小球滤过率的潜力。
AbstractObjective：To compare the clinical value of intra-voxel incoherent motion （IVIM） with a mono-exponential decay model in the analysis of diffusion weighted imaging for assessing glomerular ifltration rate in vivo.Materials and Methods：In this study, 56 participants with albuminuria or persistent micro-albuminuria were included. A parameter of estimated glomerular ifltration rate （eGFR） was calculated to classify the participants as having severe glomerular ifltration rate injury （sRI, eGFR≤30 ml/min/1.73 m^2） or not （non-sRI）. A DWI sequence with seven b-factors was performed successfully on each participant. Image analysis was performed by a radiologist to generate diffusion coefifcient （Dslow and Dfast） and fraction of fast diffusion （Ffast） in IVIM model, and to generate Dmon in mono-exponential decay model. The regions of interest were curved to cover the renal parenchyma for parameter measurements.Results：The four parameters （Dmon, Dslow, Dfast and Ffast） were （1.961±0.173）×10^-3 mm^2/s, （1.747±0.153）×10^-3 mm^2/s, （3.481±0.690）×10^-3 mm^2/s and （19.000±4.010）%, respectively, in sRI group, and were less than that （2.175±0.165）＆#215; 10-3 mm2/s, （1.917±0.162）×10^-3 mm^2/s, （4.210±0.718）×10^-3 mm^2/s and （23.110±2.809）%in non-sRI group, respectively） in non-sRI （t≥3.793,P〈0.05）, and were positively related with eGFR （t≥0.356,P〈0.05）. For differentiating sRI from non-sRI, receiver operating characteristic curve （ROC） analysis indicated no signiifcant difference between the four parameters （Z≤0.482,P〉0.05） because the areasunder ROC of Dmon, Dslow, Dfast and Ffast were 0.796, 0.786, 0.773 and 0.806, respectively. Conclusions：The Dmon, a mono-exponential modelparameter with combining micro-perfusion and diffusion information, demonstrates the potential for assessing glomerular ifltration rate in vivo.
出处
《磁共振成像》
CAS
CSCD
2016年第9期679-682,共4页
Chinese Journal of Magnetic Resonance Imaging
基金
国家自然科学基金项目(编号:81371513)~~
关键词
肾脏
肾小球滤过率
磁共振成像
弥散磁共振成像
Kidney
Glomerular ifltration rate
Magnetic resonance imaging
Diffusion magnetic resonance imaging
