中文摘要：

目的 通过联合基于MRI的联合海马体积和内嗅皮层体积,提高阿尔茨海默病（Alzheimer’s disease,AD）和轻度认知障碍（mild cognitive impairment,MCI）的诊断准确率。方法 对133例AD患者、305例MCI患者及190例健康对照（normal control,NC）的MRI数据进行脑区分割,测量获得双侧海马和内嗅皮层体积,采用Logistic回归模型合并双侧海马和内嗅皮层体积作为联合指标,利用受试者工作特性曲线（receiver operating characteristic curve,ROC）评价联合指标和单一指标对三组被试的分类效果。结果 NC-AD、MCI-AD组间双侧海马和内嗅皮层体积均有显著差异（P〈0.05）;NC-MCI组间双侧海马和左侧内嗅皮层体积均有显著差异（P〈0.05）。联合指标对于NC-MCI、NCAD、MCI-AD的ROC曲线下面积（area under the curve,AUC）分别为0.902、0.987、0.974,且最优阈值点的特异度和灵敏度均超过了80%,分类准确率分别为84.3%、95.1%、92.3%,均高于四种单一指标。结论 利用双侧海马和内嗅皮层联合指标可提高MCI和AD患者的诊断准确率。

英文摘要：

Objective To improve the diagnostic accuracy of Alzheimer＇s disease （AD） and mild cognitive impairment （MCI） by combining the magnetic resonance imaging （MRI）-based volume of hippocampus and the entorhinal cortex. Methods Brain regions were segmented based on MRI images for 133 AD patients, 305 MCI patients and 190 normal con- trols （NC）, and then, the volumes of the bilateral hippocampus and entorhinal cortex were attained for all participants. The Logistic regression model was then used to combine them into a new joint index. Receiver operating characteristic curve （ROC） was utilized to evaluate the classification ability of the joint index and four single indexes （i. e. , the volume of the bilateral hippocampus and entorhinal cortex）. Results The volume of the bilateral hippocampus and entorhinal cor- tex were significantly different between AD and NC（ P 〈0.05）, MCI（ P 〈0.05）. The volume of the bilateral hippocam- pus and left entorhinat cortex were significantly different between NC and MCI （ P 〈0.05）. The area under ROC（AUC） of the new joint index for the classification of NC-MCI, NC-AD, and MCI-AD were 0. 902, 0. 987, and 0. 974. Corre- spondingly, the classification accuracies of the new joint index were 84.3%, 95.1%, and 92.3% for the point with the optimal threshold, with sensitivities and specificities over 80 %, which were higher than those of four single indexes. Con- clusion The combination of the volume of hippocampus and the entorhinal cortex could significantly improve the diagnos- tic accuracy of MCI and AD patients.