扩散峰度成像在胶质瘤分级和预测细胞增殖中的诊断效能

目的 探讨MR扩散峰度成像（DKI）在胶质瘤分级的价值，及其与Ki-67标记指数的相关性。方法 收集经手术病理证实的脑胶质瘤患者32例，根据病理结果分为高级别胶质瘤组（HGG组，n=18）和低级别胶质瘤组（LGG组，n=14），术前行常规MR、DWI及DKI扫描，测定肿瘤实质区的平均扩散峰度（MK）、轴向扩散峰度（Ka）、径向扩散峰度（Kr）、平均扩散系数（MD）、部分各向异性（FA）和ADC值，比较高低级别胶质瘤各参数值的差异和诊断效能、肿瘤区各参数与Ki-67标记指数的相关性。结果 高低级别胶质瘤各参数值差异均有统计学意义（P均<0.01）。HGG组的MK、Ka、Kr、FA值高于LGG组，而MD、ADC值低于LGG组。MK在鉴别高低级别胶质瘤的ROC曲线下面积最大（0.82，P<0.01），特异度最高（90.40%）；ADC值敏感度最高（80.20%）。MK、Ka、Kr、MD和ADC值与Ki-67标记指数均有相关性，MK值与Ki-67标记指数相关性最大（r_s=0.61，P<0.01）。结论 DKI参数对鉴别高低级别胶质瘤有一定的价值，其诊断效能总体高于传统扩散参数。DKI参数值对预测肿瘤细胞增殖有潜在价值。

Efficacy of diffusion kurtosis imaging in grading gliomas and predicting cell proliferation

Objective To quantitatively evaluate the diagnostic efficacy of diffusion kurtosis imaging (DKI) in grading gliomas and to assess the correlation between diffusion parameters and cell proliferation biomarker (Ki-67).Methods Thirty-two patients with pathologically proved glioma (18 high-grade glioma, 14 low-grade glioma) were included. All patients underwent routine MR, DWI and DKI scan. The values of mean kurtosis (MK), axial kurtosis (Ka), radial kurtosis (Kr), mean diffusivity (MD), fractional anisotropy (FA), apparent diffusion coefficient (ADC) were calculated. The difference of each parameter between LGG and HGG group were compared. The diagnostic efficiency of each parameter were evaluated using ROC analysis, and their correlation with Ki-67 was performed.Results All diffusion parameters had significantly differentiate between HGG and LGG group (P<0.01). MK, Ka, Kr, FA were significantly higher, whereas MD and ADC was significantly lower in HGG group than those in LGG group. Area under the ROC (AUC) was highest for MK (0.82, P<0.01). The highest specificity diffusion parameter was MK (90.40%), whereas the highest sensitivity diffusion parameter was ADC (80.20%). MK, Ka, Kr, MD and ADC values were positively correlated with Ki-67 values; the correlation coefficient was highest for MK (r_s=0.61,P<0.01).Conclusion DKI parameters have a certain value in differentiating high and low grade gliomas, and their diagnostic efficiency is higher than traditional diffusion parameters. DKI parameter values have potential value in predicting tumor cell proliferation.