基于VBM-DARTEL的AD脑萎缩特征检测方法

摘要

Objective To investigate the brain grey matter atrophy in Alzheimer disease (AD)and mild cognitive impairment (MCI ), and provide a detective method for exploring the evolution mechanism of AD.Methods By combining voxel based morphometry (VBM) and diffeomorphic anatomical registration through exponentiated lie algebra (DARTEL),we firstly register and segment three T1 structural MRI datasets of 58 normal control (NC),40 AD patients,72 MCI patients including 26 progressive MCI (PMCI)patients and 46 stable MCI (SMCI ) patients.Then a specific template is built by using DARTEL method.Through deformation fields,the grey matter images are registered to MNI space with preserving the total amount of voxels by applying modulation method.Finally,statistical analysis is made on the processed datasets with two sample t test (P≤0.005,uncorrected).Results Compared to NC,the n atrophy regions in AD mainly locate in the bilateral temporal lobe, the bilateral hippocampus and parahippocampal gyrus,the bilateral amygdala,the bilateral insula,the left middle occipital gyrus,the left precuneus,the left posterior cingulate gyrus.The atrophy regions in MCI locate in the bilateral putamen,the left amygdale and the left hippocampus.The atrophy regions in PMCI locate in the left putamen,the left amygdala and the left hippocampus.There is no atrophy region found in SMCI. Compared to MCI,the atrophy regions in AD are the bilateral temporal lobe,the bilateral hippocampus,the bilateral precuneus,the bilateral middle frontal gyrus,the left cingulate gyrus,the left insula,the right amygdala,the right parahippocampal gyrus,the right superior parietal gyrus.There is no atrophy region in MCI compared to AD. Compared to SMCI, the atrophy region in PMCI is in left inferior temporal gyrus,yet there is no atrophy region found in SMCI compared to PMCI. Conclusions VBM-DARTEL based method can achieve a more accurate registration of MRI images and detect subtle volume changes of cerebral grey matter,which is helpful in providing reliable radiology evidence for the early AD diagnosis.%探究阿尔茨海默病（Alzheimer disease，AD ）与轻度认知障碍（mild cognitive impairment，MCI）大脑灰质萎缩的特征，为AD的演化机制探索提供一种有意义的检测方法。方法基于体素形态学（voxel based morphometry，VBM）结合通过取幂李代数的微分同胚解剖配准（diffeomorphic anatomical registration through exponentiated lie algebra，DARTEL）方法，首先预处理了58例正常人（normal control，NC）、26例进展型 MCI（progressive MCI，PMCI）患者、46例稳定型 MCI（stable MCI， SMCI）患者和40例 AD 患者的 T1结构磁共振图像（magnetic resonance imaging，MRI ）数据，利用DARTEL的方法创建特异性模板，再利用变形场将三组灰质图像配准到 MNI 空间并通过调制以保证体素的总数量不变。最后采用双样本 t检验对 MNI 空间平滑后的图像数据进行统计学分析（P≤0.005，未校正）。结果相对于 NC 组，AD患者主要在双侧颞叶、双侧海马及海马旁回、双侧杏仁核、双侧脑岛、左侧枕中回、左侧楔前叶、左侧后扣带回存在严重萎缩。MCI患者大脑灰质的萎缩区域主要集中在双侧豆状壳核、左侧杏仁核和左侧海马区域。PMCI患者大脑萎缩区域主要分布在左侧杏仁核、左侧海马和左侧豆状壳核，而 SMCI患者大脑灰质未见显著萎缩。相对于 MCI，AD 患者在双侧颞叶、双侧海马、双侧楔前叶、双侧额中回、左侧后扣带回、左侧脑岛、右侧杏仁核、右侧海马旁回、右侧顶上回存在严重萎缩。相对于 AD，未发现 MCI患者大脑灰质出现明显萎缩区域。相对于 SMCI，PMCI 主要表现在左侧颞下回出现显著萎缩，而相对 PMCI，SMCI未发现显著萎缩。结论利用 VBM-DARTEL方法可实现 MRI 图像的精确配准以检测大脑灰质体积的微小改变，为 AD早期诊断提供可靠的影像学特征依据。