基于T2WI的灰度直方图分析在儿童后颅窝肿瘤诊断中的价值

The value of whole tumors volume based T2 histogram analysis for differential diagnosis in pediatric posterior fossa tumors

目的研究T2WI全域灰度直方图分析在儿童后颅窝肿瘤诊断中的价值。方法回顾性分析在我院行脑部MRI检查经并病理证实的133例儿童后颅窝肿瘤患者，其中室管膜细胞瘤49例（男25例、女24例），毛细胞星形细胞瘤（WHOⅠ级）36例（男19例、女17例），髓母细胞瘤48例（男36例、女12例），年龄1~14岁，平均（7±2）岁。分别在3组T2WI轴面图像上每一层肿瘤层面用Mazda软件勾画ROI，并进行灰度全域直方图分析，获得均值（mean）、变异度（variance）、峰度（kurtosis）、偏度（skewness）、第1百分位数（perc.01%）、第10百分位数（perc.10%）、第50百分位数（perc.50%）、第90百分位数（perc.90%）及第99百分位数（perc.99%）这9个参数。因3组直方图参数特征均不符合正态分布，3组间比较采用Kruskal-Wallis H检验，组间两两比较进行Bonferroni法检验，比较各参数的统计学意义。结果通过T2WI全域灰度直方图分析得到的9个参数在3组间差异均有统计学意义（P均〈0.05）。其中在室管膜瘤与星形细胞瘤两组的鉴别之间偏度的敏感度为69.4%，特异度为77.8%，曲线下面积为0.744，最佳临界值为-0.08；星形细胞瘤与髓母细胞瘤两组的鉴别之间峰度的敏感度为95.8%，特异度为97.2%，曲线下面积为0.972，最佳临界值1.68；室管膜瘤与髓母细胞瘤两组的鉴别之间峰度的敏感度为91.7%，特异度为89.8%，曲线下面积为0.932，最佳临界值1.90，具有较高的鉴别效能。结论T2WI全域灰度直方图分析可作为鉴别儿童常见的3种后颅窝肿瘤的重要手段，在室管膜瘤与星形细胞瘤之间偏度、星形细胞瘤与髓母细胞瘤之间变异度和峰度、室管膜瘤与髓母细胞瘤之间峰度具有较高诊断效能。

ObjectiveTo study the value of whole tumors T2WI gray histogram analysis for differential diagnosis of three commonly presented pediatric posterior fossa tumors.MethodsA retrospective analysis was conducted to compare brain MRI examination and pathology diagnosis of 133 cases of posterior fossa tumors in pediatric patients in our hospital. Among them, there were 49 cases of ependymoma（there were 25 males and 24 females）, 36 cases of astrocytoma（WHO grade Ⅰ）（there were 19 males and 17 females）, 48 cases of medulloblastoma（there were 36 males and 12 females）, aged 1 to 14 years old, with an average （7±2） years old. Respectively draw the region of interest （ROI） in the T2 MR transaxial images of three groups on each layer of tumor level by using Mazda software and analyze the whole tumors gray histogram, including mean,variance, kurtosis,skewness, perc.01%, perc.10%,perc.50%,perc.90% and perc.99%.The histogram parameters of the three groups were not normally distributed. The Kruskal-Wallis test was used to compare the three groups. Bonferroni test was used to compare the two groups to find out statistical significance of each parameter.ResultsThrough histogram analysis of 9 parameters, these 9 parameters were statistically significant （all P 〈0.05）,including mean,variance, kurtosis,skewness, perc.01%, perc.10%,perc.50%,perc.90% and perc.99%. The sensitivity of skewness to differentiate ependymoma and the astrocytoma was 69.4%, the specificity was 77.8%, the area under the curve was 0.744, and the best cut-off value was -0.08. The sensitivity of kurtosis to differentiate astrocytoma and the medulloblastoma was 95.8%, the specificity was 97.2% and the area under the curve was 0.972, the cut-off was 1.68. The sensitivity of kurtosis to differentiate ependymoma and the medulloblastoma was 91.7%, the specificity was 89.8% and the area under the curve was 0.932,the cut-off was 1.90.They had a high identification efficiency.ConclusionsThe T2WI gray histogram analysis based whole tumors is helpful for the dignosis of three types of pediatric posterior fossa tumors. Skewness has a high diagnostic efficiency to differentiate ependymoma and the astrocytoma. Kurtosis has a high diagnostic efficiency to differentiate ependymoma and the medulloblastoma. Kurtosis has a high diagnostic efficiency to differentiate ependymoma and medulloblastoma.