遗忘型轻度认知损害的神经影像与神经心理特点

Characteristics of neuroimaging and neuropsychological in amnestic mild cognitive impairment patients

目的 了解aMCI脑结构改变和相关认知损害特点，探讨准确诊断aMCI的神经影像和神经心理生物指标。方法 3-DMRI采集35例aMCI患者和35名健康对照者脑结构信息，LEAP软件计算左右侧和均侧海马、杏仁核和颞角体积，MoCA测评认知功能。结果两组被试年龄、性别差异无统计学意义（P〉0．05），受教育年限差异有统计学意义（P〈0．01）。aMCI的左、右侧及均侧海马萎缩和右侧颞角扩大（P〈0．05），MoCA总评分和注意力、重复句子、抽象能力、延迟回忆、定向力等亚项得分减低（P〈0．05），MoCA与右、均侧海马及右侧杏仁核呈正相关（P〈0．05）。ROC分析显示Mo—cA总评分、左、右及均侧海马、左、右及均侧杏仁核等指标诊断aMCI准确性高（P〈0．01）。单独认知指标MoCA总评分和重复句子（wilks’Lambda=0．299，X^2=80．905，df=2，P〈0．01）区分aMCI和NC的准确性为88．6％，单独脑结构指标均侧海马和杏仁核（wilks’Lambda=0．515，X^244．509，df=2，P〈0．01）的准确性为81．4％，认知和脑结构综合指标MoCA总评分、均侧海马和左侧杏仁核（Wilks’Lambda=0．261，X^2=89．228，df=3，P〈0．01）的准确性为95．7％。结论 aMCI期已出现特异的海马、杏仁核萎缩和弥漫性认知损害，在AD非痴呆期联合应用认知测试和神经影像指标更有助于AD早期准确诊断。

Objective To understand the brain structure changes and associated cognitive impairment characteristics in patients with amnestic mild cognitive impairment （aMCI）, and to discuss the diagnostic value of neuroimaging and neuropsychologica[ biomarkers for aMCI. Methods Totals of 35 cases of aMCI patients and 35 healthy subjects were chosen in Shanghai community-dwelling older adults from China Longitudinal Aging Study. 3-D MRI data acquisition for brain structure information was used to calculate left and right side and mean volume of hippocampus, amygdala and temporal angle by using Learning embeddings atlas propagation （LEAP）. MoCA was used to assess cognitive function. Results No significant differences in age and gender were observed among the two groups （P〉 0.05）, however, education was statistically significant difference （P 〈0.01）. Multivariate analysis of variance indicated that the NC group performed better than the patients with aMCI in most of the cognitive domains, including MoCA, attention, repeating sentences, abstraction, delayed recall, and orientation （P 〈 0.05）. The aMCI showed volume loss in left, right and mean hippocampus, and right temporal angle （P 〈0.05）. MoCA scores were correlated with right and mean hippocampus （P 〈 0.05）, and right amygdala （P〈0.05）. ROC analysis showed that MoCA, left, right and mean hippocampus, and amygdale had high diagnostic accuracy for aMCI （P〈 0.01）. For cognitive measures, Wilks＇ Lambda stepwise discriminant analysis showed that MoCA and repeating sentence achieved a classification accuracy of 88.6% （Wilks＇ Lambda= 0. 299, ）X^2 = 80. 905 ,df = 2, P〈0. 01）. For brain structure, mean hip- pocampus and amygdala was 81.4M （Wilks＇ Lambda=0. 515, X^2 = 44. 509,df = 2,P〈0. 01）. Furthermore, mean hippocampus and left amygdala, combined to MoCA, achieved a classification accuracy of 95.7%（Wilks＇ Lambda=0. 261, X^2= 89. 228,df = 3,P〈 0.01）. Conclusions Hippocampus and amygdala atrophy, and cognitive impairment have specific value for identifying aMCI stage. Combining two classes of measures for the early detection of MCI provides more accurate biomarkers than those based on neuroimaging or cognitive measures alone.