目的 评估MR扩散峰度成像(diffusion kurtosis imaging,DKI)及扩散张量成像(diffusion-tensor imaging,DTI)在鉴别低级别胶质瘤与脑炎中的价值.材料与方法 回顾性分析58例低级别胶质瘤或脑炎患者的影像学资料,所有患者在术前或保守治疗前均行MRI常规序列及DKI序列扫描.用NeuDiLab软件处理DKI图像,获得基于DKI模型的平均峰度(mean kurtosis,MK)、轴向扩散峰度(axial kurtosis,AK)和径向扩散峰度(radial kurtosis,RK)以及基于DTI模型的平均扩散系数(mean diffusivity,MD)、轴向扩散系数(axial diffusivity,AD)、径向扩散系数(radial diffusivity,RD)和各向异性分数(fractional anisotropy,FA)参数图及b= 0 s/mm2的扩散图像(B0).用ITK-SNAP软件在B0图像上手动勾画肿瘤的感兴趣容积(volume of interest,VOI),将其配准到其他参数图上.用FAE软件提取各参数图的平均值.根据病理检查或脑脊液检查结果将患者分为胶质瘤组和脑炎组.采用卡方检验、独立样本t检验及Mann-Whitney U检验比较两组患者的一般资料、常规MRI表现及扩散参数的差异.计算Cohen's d值评估各扩散参数的效应量.绘制受试者工作特征(receiver operating characteristic,ROC)曲线,计算曲线下面积(area under the curve,AUC)、敏感度、特异度和准确度,并用DeLong检验比较各参数的鉴别诊断效能.结果 共51例患者被纳入研究,包括29例低级别胶质瘤患者、22例脑炎病变患者.RK鉴别诊断脑炎组低级别胶质瘤的表现最好,AUC为0.878,当阈值取0.662时,其敏感度为72.7%,特异度为89.7%.DeLong检验显示DKI模型的诊断表现显著优于DTI模型.结论 DKI序列有助于鉴别诊断低级别胶质瘤与脑炎.
The value of DKI and DTI in the differential diagnosis of low-grade gliomas and encephalitis
Objective:To evaluate the value of MR diffusion kurtosis imaging(DKI)and diffusion-tensor imaging(DTI)in differentiating low-grade gliomas from encephalitis.Materials and Methods:The imaging data of 58 patients with either low-grade glioma or encephalitis were retrospectively collected.All patients underwent routine MRI and DKI sequence scans before surgery or conservative treatment.The DKI images were processed with NeuDiLab software to obtain the parameter maps of DKI-based mean kurtosis(MK),axial kurtosis(AK)and radial kurtosis(RK),as well as DTI-based mean diffusivity(MD),axial diffusivity(AD),radial diffusivity(RD)and fractional anisotropy(FA)and diffusion images of b=0 s/mm2(B0).The volumes of interest(VOIs)of the tumor were manually delineated on the B0 image with ITK-SNAP software and were registered to other parametric maps.The mean values of each parameter were extracted with FAE software.Patients were divided into glioma group and encephalitis group according to pathological examination or cerebrospinal fluid examination results.The chi-square test,independent samples t test and Mann-Whitney U test were used to compare the general data,routine MRI findings and diffusion parameters between the two groups.Cohen's d values were calculated to evaluate the effect sizes of diffusion parameters.The receiver operating characteristic(ROC)curve was drawn to calculate the area under the curve(AUC),sensitivity,specificity and accuracy.The DeLong test was used to compare the differential diagnostic performance of diffusion parameters.Results:A total of 51 patients were included in the study,including 29 patients with low-grade gliomas and 22 patients with encephalitis.RK demonstrated the best performance in distinguishing low-grade gliomas from the encephalitis group,with an AUC of 0.878.When the threshold was set at 0.662,the sensitivity was 72.7%,and the specificity was 89.7%.The DeLong test indicated that the diagnostic performance of the DKI model was significantly superior to DTI.Conclusions:DKI is helpful in the differential diagnosis of low-grade gliomas and encephalitis.
万方数据
维普
