摘要
目的:探讨阿尔茨海默病(AD)和轻度认知障碍(MCI)患者在嗅觉功能、眼动特征及磁共振成像(MRI)脑结构指标上的差异,分析多模态生物标志物联合应用对AD早期诊断的价值。方法:选取深圳市福田区第二人民医院、深圳市罗湖区人民医院、深圳市第二人民医院2021年7月至2024年8月收治的30例AD患者作为AD组,收治的40例MCI患者作为MCI组,招募深圳市第二人民医院2022年1月至2024年8月的40例体检人员作为正常对照(NC)组,采用中国气味识别测验(CSIT)评估嗅觉功能,基于人工智能的眼动追踪技术记录眼动参数,并通过3.0T MRI测量脑区体积。通过方差分析、Spearman相关性分析及Logistic回归模型评估组间差异、指标相关性及诊断效能。结果:AD组与MCI组患者左右侧海马体积、左右侧杏仁核体积、左右侧前中央回体积、左右侧前额叶体积及左右侧楔叶体积均小于NC组,且AD组左侧海马体积、左右侧杏仁核体积、左右侧前中央回体积、左右侧前额叶体积及左右侧楔叶体积小于MCI组,简易精神状态检查表(MMSE)和蒙特利尔认知评估量表(MoCA)评分低于NC组,且AD组低于MCI组,中国气味识别测验(CSIT)得分、错误眼动纠正频率及新照片停留时间低于或短于NC组,但MCI组高于或长于AD组,差异具有统计学意义(P<0.05);MMSE评分、MoCA评分、左侧海马体体积、右侧海马体体积、左侧杏仁核体积、右侧杏仁核体积、左侧前中央回体积、左侧楔叶体积、错误眼动纠正频率与CSIT得分均存在显著的正相关关系(P<0.05),其中左侧海马体体积与CSIT得分存在较强的正相关性(r=0.622,P<0.05);在MRI脑结构指标方面,左侧海马体体积与错误眼动纠正频率呈显著正相关(r=0.537,P<0.05),与新照片停留时间也有一定相关性(r=0.320,P<0.05);右侧杏仁核体积与新照片停留时间显著正相关(r=0.397,P<0.05);左侧前额叶体积与错误眼动纠正频率显著正相关(r=0.441,P<0.05);受试者工作特征(ROC)曲线分析结果显示,CSIT、MMSE得分、左侧海马体体积、错误眼动纠正频率诊断AD的曲线下面积(AUC)分别为0.799、0.789、0.628、0.687,以CSIT得分、错误眼动纠正频率、左侧海马体体积和MMSE评分为变量构建的逻辑回归预测模型的AUC值达到了0.903,显著高于各项指标单独检测(P<0.05)。结论:嗅觉、眼动及MRI脑结构的多模态生物学标志物均在AD、MCI与NC组间有差异,联合嗅觉–眼动以及MRI脑结构构建的诊断模型可进一步提高早期AD的诊断效能,为AD早期筛查提供了新策略。
Objective To explore the differences in olfactory function,eye movement characteristics,and magnetic resonance imaging(MRI)brain structure indicators between patients with Alzheimer's disease(AD)and mild cognitive impairment(MCI),and to analyze the value of combined application of multimodal biomarkers in the early diagnosis of AD.Methods Thirty patients with AD andforty patients with MCI admitted to the Second People's Hospital of Futian District Shenzhen,Shenzhen Luohu People's Hospital,and Shenzhen Second People's Hospital from July 2021 to August 2024 were selected as the AD group and MCI group.Forty healthy individuals who underwent physical examinations at shenzhen Second People's Hospital from January 2022 to August 2024 were recruited as the normal control(NC)group.Olfactory function was evaluated using the Chinese smell identification test(CSIT),eye movement parameters were recorded using artificial intelligence-based eye tracking technology,and brain region volumes were measured using 3.0 T MRI.Group differences,indicator correlations,and diagnostic efficacy were evaluated using analysis of variance,Spearman correlation analysis,and Logistic regression models.Results The volumes of the left and right hippocampus,left and right amygdala,left and right precentral gyrus,left and right prefrontal and left and right cuneus in the AD and MCI group were smaller than those in the NC group,and the volumes of the left hippocampus,left and right amygdala,left and right precentral gyrus,left and right prefrontal lobe,and left and right cuneus in the AD group were smaller than those in the MCI group.The mini-mental state examination(MMSE)and Montreal cognitive assessment(MoCA)scores in the AD and MCI groups were lower than those in the NC group,and the scores in the AD group were lower than those in the MCI group.The CSIT scores,error eye movement correction frequency,and new photo dwell time in the AD group were lower or shorter than those in the NC group,but the indexes of the MCI group were higher or longer than those of the AD group,with statistically significant differences(P<0.05).There were significant positive correlations between MMSE scores,MoCA scores,left hippocampal volume,right hippocampal volume,left amygdala volume,right amygdala volume,left precentral gyrus volume,left cuneus volume,error eye movement correction frequency,and CSIT scores(P<0.05),among which the left hippocampal volume had a strong positive correlation with the CSIT score(r=0.622,P<0.05).In terms of MRI brain structure indicators,the left hippocampal volume was significantly positively correlated with the error eye movement correction frequency(r=0.537,P<0.05),and also had a certain correlation with the new photo dwell time(r=0.320,P<0.05).The right amygdala volume was significantly positively correlated with the new photo dwell time(r=0.397,P<0.05).The left prefrontal lobe volume was significantly positively correlated with the error eye movement correction frequency(r=0.441,P<0.05).The receiver operating characteristic(ROC)curve analysis showed that the area under the curve(AUC)for diagnosing AD by CSIT,MMSE score,left hippocampal volume,and error eye movement correction frequency were 0.799,0.789,0.628,and 0.687,respectively.The AUC value of the logistic regression prediction model constructed with CSIT score,error eye movement correction frequency,left hippocampal volume,and MMSE score as variables reached 0.903,which was significantly higher than that of each individual indicator(P<0.05).Conclusion Multimodal biological markers of olfaction,eye movement,and MRI brain structure show differences among AD,MCI,and NC groups.The diagnostic model constructed by combining olfaction,eye movement,and MRI brain structure can further improve the diagnostic efficacy of early AD,providing a new strategy for early screening of AD.
作者
黄思博
梁月飞
陈泽楷
何文杰
钟旗
李爱东
夏军
HUANG Sibo;LIANG Yuefei;CHEN Zekai;HE Wenjie;ZHONG Qi;LI Aidong;XIA Jun(Department of Radiology,Shenzhen Second People's Hospital,Guangdong Shenzhen 518035;Department of Neurology,Shenzhen Luohu People's Hospital,Guangdong Shenzhen 518000;Department of Rehabilitation Medicine,The Second People's Hospital of Futian District Shenzhen,Guangdong Shenzhen 518049)
出处
《深圳中西医结合杂志》
2025年第16期1-5,9,I0002,I0003,共8页
Shenzhen Journal of Integrated Traditional Chinese and Western Medicine
基金
国家自然科学基金(82171913)
深圳市科技创新委员会基础研究重点项目(JCYJ20220818101816036)。
关键词
阿尔茨海默病
轻度认知障碍
嗅觉测试
眼动追踪
磁共振成像
早期诊断
Alzheimer's disease
Mild cognitive impairment
Olfactory test
Eye movement tracking
Magnetic resonance imaging
Early diagnosis
