MRI增强全域灰度直方图在鉴别儿童后颅窝肿瘤的应用

Application of whole-tumor histogram analysis of enhanced MRI for posterior fossa tumors in children

目的:研究MRI增强全域灰度直方图分析对三种常见的儿童后颅窝肿瘤(室管膜瘤、星形细胞瘤、髓母细胞瘤)的鉴别诊断价值。方法:回顾性分析我院进行脑部MRI检查经并病理证实的76例儿童后颅窝肿瘤患者,其中室管膜细胞瘤25例,星形细胞瘤26例,髓母细胞瘤25例,分别在三组MR增强矢状位图像上每一层肿瘤层面用Mazda软件勾画感兴趣区并进行灰度全域直方图分析,对获得的三组直方图参数特征分别进行统计学分析,比较各参数的统计学意义。结果:通过灰度全域直方图分析得到的9个参数中,均值、变异度、偏度、第10百分位数、第50百分位数及第99百分位数这6个参数差异有统计学意义(P均<0.05),其余3个参数峰度、第1百分位数、第90百分位数在三组间无显著性差异(P均>0.05)。结论:MRI全域灰度直方图分析可作为鉴别儿童常见的三种后颅窝肿瘤的重要手段,第50百分位数与变异度具有较高诊断效能。

Objective: To study the value of whole tumors enhanced MRI gray histogram analysis of differential diagnosis in three common pediatric posterior fossa tumors(ependymoma, astrocytoma, medulloblastoma). Methods: A retrospective analysis was conducted by brain MRI examination and pathology diagnosis of 76 cases of posterior fossa tumors in children in our hospital. Among them, there were 25 cases of ependymoma, 26 cases of astrocytoma, 25 cases of medulloblastoma. Respec-tively, we drew the region of interest(ROI) in the enhanced MR sagittal images of three groups on each layer of tumor level by using Mazda software and analyzed the whole tumors gray histogram, then performed statistical analysis on the three sets of parameters obtained from histograms to find out statistical difference of each parameter. Results: Through histogram analysis of9 parameters, the difference of these 6 parameters were statistically significant(all P<0.05), including mean, variance, skewness,Perc.10%, Perc.50% and Perc.99%, the remaining 3 parameters, including kurtosis, Perc.01%, Perc.90% had no significant dif-ference(all P>0.05). Conclusion: The MRI gray histogram analysis based on whole tumors is helpful for the identification of three kinds of pediatric posterior fossa tumors,the Perc.50% and variance had a high diagnostic efficiency.