DCE-MRI直方图分析在脑胶质瘤分级中的应用

发布时间：2018-02-28 08:40

本文关键词： 直方图动态增强磁共振成像胶质瘤　出处：《中国医学计算机成像杂志》2017年01期 　论文类型：期刊论文

【摘要】：目的:探索动态增强磁共振成像(DCE-MRI)结合以整个肿瘤体积为基础的直方图(Histogram)分析方法在脑胶质瘤分级中的应用价值。方法:回顾性分析31例经过DCE-MRI检查的胶质瘤患者,其中包括8例低级别胶质瘤(LGG)和23例高级别胶质瘤(HGG)。应用血流动力学模型计算出各个病例的渗透参数,包括容积转运常数(K~(trans))、回流速率常数(Kep)、血管外细胞外容积比(Ve)和血浆容积(Vp);选取肿瘤实质区作为感兴趣区(ROI),并计算出各渗透参数的直方图参数,包括平均值、标准差、峰度、偏度及第90百分位数。不同级别胶质瘤的K~(trans)、Kep、Ve和Vp直方图参数值的比较采用Mann-Whitney U检验,采用ROC分析直方图各参数值鉴别HGG和LGG的诊断效能,应用组内相关系数(ICC)分析数据测量的可重复性。结果:K~(trans)值的平均值、标准差、峰度、偏度及第90百分位数在鉴别LGG与HGG中的差异均具有统计学意义(P0.05)。ROC分析显示以上各值在鉴别高、低级别胶质瘤中均有较好的诊断效能,其中K~(trans)值的偏度诊断准确性最高。各组ICC值显示出较好的可重复性。结论:DCE-MRI成像结合全肿瘤直方图分析方法在胶质瘤的分级中具有重要的应用价值。
[Abstract]:Objective: to explore the value of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) combined with histogram based on the whole tumor volume in the grading of glioma. Methods: 31 cases of glioma with DCE-MRI were retrospectively analyzed. These include 8 cases of low grade gliomas (LGG) and 23 cases of high grade gliomas (HGGGG). The osmotic parameters of each case were calculated using hemodynamic models. These include the volume transport constant (K), the reflux rate constant (K), the extracellular volume ratio (E) and the plasma volume (V); the tumor parenchyma region is selected as the region of interest, and the histogram parameters of the osmotic parameters, including mean value, standard deviation, kurtosis, are calculated. Deviation and 90 percentile. The comparison of KeptransVe and VP histogram parameters of different grade gliomas was performed by Mann-Whitney U test, and the diagnostic efficacy of HGG and LGG was identified by ROC analysis. The repeatability of data measurement was analyzed by ICC.Results the average value, standard deviation, kurtosis of the value of:. The difference of bias and 90 percentile in differentiating LGG from HGG was statistically significant (P 0.05). The ICC value of each group showed good reproducibility. Conclusion the weight DCE-MRI imaging combined with the whole tumor histogram analysis method has important application value in the classification of glioma.
【作者单位】：南京医科大学第一附属医院影像科;通用电气医疗集团;
【分类号】：R445.2;R739.41

【相似文献】