摘要
对243例脑转移瘤和88例星形胶质细胞瘤的CT影像特点和人口统计特征进行了逐步回归分析,筛选了具有统计学显著意义的因素。选入因素对因变量的复相关指数(R^2)为0.6287。建立了两类脑肿瘤的判别函数模型,并对研究病例进行回代试验。回代总符合率为93.35%,56例单发转移瘤的回代符合率为85.71%,187例多发转移瘤回代符合率为96.26%,88例胶质细胞瘤的回代符合率为92.05%。
For the purpose of differentiating intracranial metastasis from astrocytoma,the CT scan results and some demographic factors of 243 in-tracranial metastasis cases and 88 astrocytoma cases have been analyzed with
discriminant analysis.In order to select statistically significant factors into discriminant model,the stepwise regression was adopted as a screen method.The R^2 of the established multivariate model is 0.6287 and the significant factors are size of the tumor,age,density of the contrast enhancement,diameter of edema surrounding a tumor,calcification,numbers of tumors,sex,hemorrhage and density before contrast enhancement.The total accuracy of the discriminant model is 93.35%.The accuracy for the 88 astrocytomas is
92.05%,for the 56 solitary cerebral metastasis and 187 multiple metastasis
are 85.71% and 96.26% respectively.
出处
《山东医科大学学报》
1992年第1期61-63,共3页
Acta Academiae Medicinae Shandong
关键词
脑转移瘤
星形细胞瘤
CT
Intracranial metastasis
Astrocytoma
CT image
Discriminant analysis
