Objective To discriminate Alzheimer disease (AD) and mild cognitive impairment (MCI) from normal controls with 3D texture features, in order to explore the new approach for the early diagnosis of AD. Methods 3D texture analysis was performed on MR images of 12 early AD patients (AD group), 12 MCI patients (MCI group) and 12 normal controls (NC group). Texture features of the hippocampus and corpus callosum were extracted from gray level co-occurrence matrix and run length matrix. The texture features that existed significant differences among groups were used as features in a classification procedure based on support vector machines (SVM). The accuracy was evaluated with leave-one-out cross-validation. Results The classification accuracy for NC and MCI group, MCI and AD group, NC and AD group was 79.17％, 83.33％ and 91.67％, respectively. Conclusion 3D texture characteristics can be used to discriminate patients with early AD and patients with MCI from normal controls, and would be helpful to early diagnosis of AD.%目的 利用三维纹理特征对阿尔茨海默病(AD)患者和轻度认知障碍(MCI)患者进行分类识别,以探索AD早期诊断新途径.方法 对12例早期AD患者(AD组)、12例MCI患者(MCI组)及12名健康对照者(NC组)的MR图像进行三维纹理分析,采用灰度共生矩阵和游程长矩阵提取每位受试者左、右侧海马结构及胼胝体的三维纹理特征,选取三组间存在显著性差异的纹理参数作为特征变量,采用支持向量机(SVM)方法对各组进行分类,利用留一法估算分类准确率.结果 对NC组与MCI组、MCI组与AD组、NC组与AD组进行分类识别的最高准确率分别为79.17%、83.33%、91.67%.结论 利用三维纹理分析可分类识别早期AD患者及MCI患者,有助于AD的早期诊断.
展开▼
