论文部分内容阅读
目的探讨磁共振22b值(0~5000 s/mm2)非高斯弥散加权成像在健康成人脑白质和胶质瘤应用的可重复性。方法 12名健康志愿者和65名胶质瘤患者进行22 b值e DWI(enhanced diffusion weighted imaging)磁共振扫描。前者进行连续2次e DWI扫描,选取基底节层面和半卵圆中心层面深部白质为感兴趣区;后者进行单次e DWI扫描,由2名神经放射医师分别在肿瘤最大层面的肿瘤实质区放置感兴趣区,工作站进行双指数和拉伸指数模型后处理;前者对2次扫描双指数模型拟合的慢弥散系数Dslow(slow diffusion coefficient)、快弥散系数Dfast(fast diffusion coefficient)和快弥散容积分数PF(perfusion fraction of Dfast)以及拉伸指数模型拟合的分布弥散系数DDC(distributed diffusion coefficient)和不均质指数α进行配对t检验、计算变异系数(coefficient of variability,CV)和绘制BlandAltman(B-A)散点图;后者对2次测量的Dslow、Dfast、PF、DDC和α进行观察者间一致性检验,并计算CV。结果健康志愿者各参数2次测量配对比较差异均无统计学意义(P>0.05),CV值均小于5%,B-A散点图绝大多数(≥11/12)测值均位于95%一致性范围内;胶质瘤各参数2次测量一致性系数均大于0.75(P<0.001),Dfast和PF的CV值最大,范围在5.0%~11.5%之间,其次是DDC和Dslow,α最小,CV值小于2%,除α外,其余参数CV值均随胶质瘤级别增高而增大。结论健康脑白质评价中非高斯弥散加权成像各参数均具有较好的可重复性;胶质瘤评价中拉伸指数模型α具有最好的可重复性。
Objective To investigate the reproducibility of magnetic resonance 22b (0 ~ 5000 s / mm2) non-Gaussian diffusion weighted imaging in white matter and glioma of healthy adults. Methods Twelve healthy volunteers and 65 patients with glioma underwent 22 b-value e DWI (enhanced diffusion weighted imaging) magnetic resonance imaging. The former underwent two successive e DWI scans and the deep white matter at the basal and semi-oval centers was selected as the region of interest. The latter was subjected to a single e DWI scan and was performed by two neuroradiologists in the tumor parenchyma The area of interest was placed and the workstation was subjected to dual index and tensile index model post-processing. The former was used to measure the slow diffusion coefficient (Dslow), fast diffusion coefficient (DIFF) and fast dispersion coefficient PF (perfusion fraction of Dfast) and distribution diffusion coefficient (DDC) and nonhomogeneous index (α) of the tensile index model were paired t-test to calculate coefficient of variability (CV) and draw BlandAltman (BA) scatter plot; the latter performed an interobserver agreement test on Dslow, Dfast, PF, DDC, and α for 2 measurements and calculated CV. Results There was no significant difference in paired comparison between the two parameters of healthy volunteers (P> 0.05), the CV values were less than 5%, and the absolute values of BA scattergram (≥11 / 12) were all consistent at 95% (P <0.001). The CV values of Dfast and PF ranged from 5.0% to 11.5%, followed by DDC and Dslow, with the smallest α , CV value is less than 2%, with the exception of α, CV values of other parameters increased with the glioma level increased. Conclusion The parameters of non-Gaussian diffusion weighted imaging in healthy white matter evaluation are both good reproducibility. The tensile index model α in glioma evaluation has the best reproducibility.