结构MRI联合基于VR眼动追踪技术的计算机化认知评估在阿尔茨海默病早期诊断中的应用价值  

Application value of structural MRI combined with computerized cognitive assessment based on VR eye-tracking technology in early diagnosis of Alzheimer's disease

作　　者：韦卓男 樊响 余可妍 陈乐乐 尹晨旺 陈慧 戚玉龙[1] 陈旭辉[2] 胡俊[2] 张绪 成官迅[1] WEI Zhuonan;FAN Xiang;YU Keyan;CHEN Lele;YIN Chenwang;CHEN Hui;QI Yulong;CHEN Xuhui;HU Jun;ZHANG Xu;CHENG Guanxun(Department of Medical Imaging,Peking University Shenzhen Hospital,Shenzhen 518036,China;Department of Neurology,Peking University Shenzhen Hospital,Shenzhen 518036,China;National Engineering laboralory for Big Data System Computing Technology,Shenzhen Universitiy,Shenzhen 518061,China)

机构地区：[1]北京大学深圳医院医学影像科,深圳518036 [2]北京大学深圳医院神经内科,深圳518036 [3]深圳大学大数据系统计算技术国家工程实验室,深圳518061

出　　处：《磁共振成像》2024年第6期49-53,共5页Chinese Journal of Magnetic Resonance Imaging

基　　金：深圳市科技计划项目(编号:KCXFZ20211020163408012)。

摘　　要：目的探讨结构MRI联合基于VR眼动追踪技术的计算机化认知评估在阿尔茨海默病(Alzheimer's disease,AD)早期诊断中的临床应用价值。材料与方法77例早期AD患者和58例认知未受损者(cognitive unimpaired,CU)均接受了传统神经心理学量表评估[包括简易精神状态量表(Mini-Mental State Examination,MMSE)、蒙特利尔认知评估量表(Montreal Cognitive Assessment,MoCA)等]、一项时长为5分钟的VR眼动认知评估,以及头颅三维T1加权成像(three dimensions T1-weighted imaging,3D-T1WI)磁共振结构像检查,并利用自动结构分析工具获取海马体积(hippocampal volume,HV)数值。分析VR眼动认知评估的总分与传统认知评估之间的相关性,构建受试者工作特征(receiver operating characteristic,ROC)曲线,分析单个指标(HV)及联合指标(HV联合VR眼动认知评估的总分)对AD的诊断效能。结果(1)HV诊断早期AD的AUC为0.629(95%CI:0.542~0.711),HV联合VR眼动认知评估的总分诊断早期AD的AUC为0.825(95%CI:0.751~0.885),DeLong检验结果显示两者差异有统计学意义(P<0.001);(2)早期AD组的HV(Z=−2.565,P=0.010)较CU组低;(3)VR眼动认知评估的总分与MMSE评分、MoCA评分呈正相关(r=0.531、0.627,P均<0.001)。结论联合基于VR眼动追踪技术的计算机化认知评估能提高HV对早期AD的诊断价值。Objective:To explore the clinical application value of structural MRI in conjunction with computerized cognitive assessment based on VR eye-tracking technology in early diagnosis of Alzheimer's disease(AD).Materials and Methods:Seventy-seven patients with early AD and fifty-eight cognitively unimpaired(CU)participants underwent the traditional neuropsychological scale assessment[including Mini-Mental State Examination(MMSE)and Montreal Cognitive Assessment(MoCA)],a 5-minute VR eye-tracking technology-based cognitive assessment,and 3D T1-weighted MRI.The hippocampal volume(HV)was quantified using an automatic structural analysis tool.Receiver operating characteristic(ROC)curves were constructed to analyze the diagnostic effectiveness of the single index(HV)and the combined index(HV along with the total score from the VR eye-tracking technology-based cognitive assessment)for AD.Results:(1)The area under ROC curve(AUC)of early AD diagnosed by HV was 0.629(95%CI:0.542-0.711).When combining HV with the VR eye-tracking cognitive assessment,the AUC increased to 0.825(95%CI:0.751-0.885).The DeLong test results indicated a significant difference between the two methods(P<0.001).(2)The HV in the early AD group was significantly lower than that in the CU group(Z=−2.565,P=0.010).(3)The total score of VR eye-tracking cognitive assessment was positively correlated with MMSE score and MoCA score(r=0.531,0.627,P<0.001).Conclusions:Combining computerized cognitive assessment using VR eye-tracking technology significantly enhances the diagnostic value of hippocampal volume in the early detection of AD.

关 键 词：阿尔茨海默病 磁共振成像 眼动追踪 计算机化认知评估 海马体积 

分 类 号：R445.2[医药卫生—影像医学与核医学] R749.16[医药卫生—诊断学]