利用ADC全域直方图鉴别胶质母细胞瘤和原发中枢神经系统淋巴瘤

ADC Histogram Analysis for Differentiating Glioblastomaa from Primary Central Nervous System Lymphoma

目的探讨胶质母细胞瘤(GBM)和原发中枢神经系统淋巴瘤(PCNSL)的表观扩散系数(ADC)直方图特征以及其对鉴别诊断的价值。方法回顾性分析经病理证实的30例GBM和38例PCNSL患者的资料,使用MaZda软件于ADC图像上在肿瘤的每一层面提取感兴趣区(ROI),对两组肿瘤全域的直方图参数进行统计学分析,比较各项参数之间的差异。结果利用全域ADC直方图分析提取出的9个参数特征在GBM和PCNSL间均有统计学差异(P≤0.001),第50百分位数具有最高的鉴别诊断价值。GBM的第50百分位数93.38±21.58小于PCNSL 143.35±34.34。以第50百分位数111.0为临界值,绘制受试者工作特征曲线(ROC),曲线下面积(AUC)为0.904,鉴别的灵敏度为84.21%,特异度为86.67%。结论 ADC全域直方图分析可对GBM和PCNSL的鉴别提供更精准的数据支持,为两种肿瘤的鉴别诊断提供了新的思路和方法。

Objective To investigate the diagnostic value of the histogram analysis derived from ADC MR imaging in identification of glioblastomaa from primary central nervous system lymphoma.Methods Retrospective analysis of pathologically confirmed 30 cases of glioblastoma,and 38 cases of primary central nervous system lymphoma.Drawing the region of interest(ROI)on every slice of the tumor in the ADC MR image using the software of MaZda.Histogram analysis for 9 texture parameters was evaluated to find out the significant differences between the two kinds of tumors.Results All of the 9 texture parameters have statistical significance between two groups.Among them,the perc50%has the highest differential diagnosis value.The sensitivity,specificity and AUC(Area Under Curve)were 84.21%,86.67%and 0.904.Conclusion Histogram analysis can provide more quantitative information characteristics,which provides a new method for the differential diagnosis of GBM and PCNSL.