Investigation Of Whole-Brain Iron Accumulation And Cerebral Structural Alteration In Alzheimer’s Disease On MRI

Objective:To investigate the level and distribution of iron accumulation through the whole brain cortex and deep gray nucleus in patients with Alzheimer’s Disease（AD）by using quantitative susceptibility mapping（QSM）,and analyze the relationship between brain structural change and magnetic susceptibility value in AD patients and the diagnostic value of magnetic susceptibility value.Materials and Methods:From November 2018 to March 2020,thirty AD patients from the Memory Clinic of Shandong Provincial Hospital,who were clinically diagnosed probable or possible Alzheimer’s Disease with early stage based on the National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer Disease and Related Disorders Association（NINCDS/ADRDA）criteria,were prospectively recruited as AD group.Twenty-seven age and sex matched healthy volunteers were recruited from local community as the healthy controls（HC）.All the AD patients and healthy volunteers underwent complete MMSE test for the cognitive evaluation.MRI scans were acquired on 3.0 T MRI system（Siemens Healthineers,Erlangen,Germany）with 32 channel head coils.All the subjects underwent complete routine MRI,multi-echo GRE sequence and T1-MPRAGE sequence scan.The magnitude image and phase image were used to generate QSM image through MEDI toolbox.The two-step registration method was conducted on SPM.Firstly,the T1 structural image was co-registered to magnitude image of the first echo in the same native space for each subject.The co-registered T1 structural image was then segmented and standardized to Montreal Neurological Institute（MNI）space,during this process the forward deformation field was saved.Then,the QSM image was spatially normalized and warped to MNI space by using the previous forward deformation parameters.We manually drew the bilateral deep gray matter nucleuses as ROIs（region of interests）on the original QSM with at least three consecutive slices and recorded the mean of magnetic susceptibility value,which included caudate nucleus,putamen,globus pallidus,substantia nigra,red nucleus and dentate nucleus.The automated anatomical labelling atlas 116（AAL 116）template was used to automatically extract the magnetic susceptibility values（MSV）from 116 brain cortical regions on normalized QSM.The voxel-based morphometry（VBM）and surface-based morphometry（SBM）were implemented to analyze the gray matter volume（GMV）and cortical thickness respectively.Using VBM8 and Freesufer to obtain the GMV and cortical thickness of all the subjects by series of segmentation,registration and correction on T1 structural images.In VBM8 software,a general linear model（GLM）with age and sex as covariance was constructed for GMV.Then,two independent samples t-test was used to compare the difference of GMV between two groups and the Family Wise Error（FWE）method was used for multiple comparisons correction.The commands in Freesurfer were used to construct the GLM for cortical thickness and then the different offset same slop（DOSS）model with age and sex as covariance was constructed.The two independent samples t-test was used to compare the two groups difference of GMV on the cluster-wise level.The False Discovery Rate（FDR）method was used for multiple comparisons correction.P<0.05 was considered statistically significant.Following the VBM and SPM analysis we saved the significantly different regions of brain structure（GMV and cortical thickness）as binary masks,and then extracted their corresponding magnetic susceptibility value,GMV and cortical thickness.The rest statistical analysis was conducted on IBM SPSS21 software.The Kolmogorov-Smirnov method was used to test the normal distribution of data.Continuous data are presented as means ± standard deviations.The differences of age and gender between two groups were compared by student’s t test and Chi-squared test respectively.The differences of MSV in deep gray matter nucleus,116 gray matter regions and the structural change regions between two groups was evaluated by Student’s t test.The Pearson correlation analysis was used to assess the correlation between MSV and GMV and cortical thickness.The receiver operating characteristic curve（ROC）and area under curve（AUC）were used to evaluate the diagnostic value of MSV.And,the Yuden index was used for determining the best cut-off value and corresponding specificity and sensitivity.P<0.05 was considered statistically significant.Results:1.Thirty-one AD patients（age=65.1±7.40,male=12,female=18）and twenty-seven healthy controls（62.4±5.47,male=13,female=14）were finally recruited in this study.There were no significant differences of age（p=0.132）and gender（p=0.599）between AD and HC groups.The MMSE score of AD groups was significant lower than that in HC（AD:MMSE=19.43±2.8;HC:MMSE=27.93±0.9,P<0.001）.2.Among the 12 deep gray matter nucleuses,the MSV of AD group in left caudate nucleus（P=0.036）and bilateral putamen（left:P=0.005;right:P=0.013）were significantly greater than that in healthy controls.For the ROIs which were defined from AAL116 atlas,the MSV of total 17 brain regions in AD group were significant higher than HC group,including left olfactory cortex（P=0.034）,right parahippocampal gyrus（P=0.032）,left occipital inferior gyrus（P=0.049）,right parietal inferior gyrus（P=0.043）,right supramarginal gyrus（P=0.028）,right angular gyrus（P=0.025）,left putamen（P=0.048）,right putamen（P=0.021）,left Heschl gyrus（P=0.012）,left temporal superior gyrus（P=0.046）,left temporal inferior gyrus（P=0.035）,Cerebe-lum_Crusl_L（P=0.024）,Cerebelum₇b_L（P=0.040）,Cerebelum₇b_R（P=0.013）,Cerebelum_8L（P=0.029）,Cerebelum_8R（P=0.020）and Cerebelum₁0_L（P=0.019）.3.After the VBM analysis,the GMV of bilateral hippocampus,parahippocampal gyrus,amygdala（left peak MNI coordinate=-28.5,-10.5,-18;right peak MNI coordinate= 28.5,-6,-27）and bilateral frontal inferior gyrus（left peak MNI coordinate=-42,12,30;right peak MNI coordinate=43.5,15,31.5）in AD group was significantly decreased compared to HC group（P<0.05,FWE correction）.4.After the SBM analysis,the significant decrease of cortical thickness were found in bilateral precuneus（left peak MNI coordinate=-15.2,-47.7,31.8;right peak MNI coordinate=13.2,-65.3,35.4;5.8,-56.7,13.0）,left supramarginal gyrus（peak MNI coordinate=-56.1,-39.3,43.3）,bilateral middle temporal gyrus（left peak MNI coordinate=-57.3,-19.4,-15.9;right peak MNI coordinate=59.6,-32.6,-15.9）,left caudal middle frontal gyrus（peak MNI coordinate=-42.7,14.1,43.9）、right superior frontal gyrus（peak MNI coordinate=16.6,27.6,53.5）and bilateral inferior parietal lobe（left peak MNI coordinate=-48.9,-55.3,24.1;right peak MNI coordinate=39.6,-68.6,43.5）in AD group compared to HC group（Cluster level,P<0.05,FDR correction）.5.Among the GMV significantly different regions,the MSV of left para-hippocampal gyrus（P=0.04）and left frontal inferior gyrus（P=0.019）in AD group were significant higher in comparison of HC group.And among the cortical thickness significantly different regions,the MSV of left inferior parietal lobe（P=0.028）,left supramarginal gyrus（P=0.013）and right middle temporal gyrus（P=0.018）were significant higher compared to HC group.Among the structural different regions determined by VBM and SBM analysis,there were no statistical correlation between MSV and GMV or cortical thickness（P>0.05）.6.The ROC analysis of MSV in deep gray matter nucleuses showed that the MSV of left putamen had the biggest AUC which is 0.697（P=0.013）,and the cut-off value was 95.17 ppb,the sensitivity and specificity were 50%and 95.8%respectively.The ROC analysis of MSV based on the AAL116 template presented that the MSV in Cerebelum_Crus1_L had the biggest AUC which is 0.71（P=0.007）,and the cut-off value was-3.30 ppb,the sensitivity and specificity were 63.3%and 81.5%respectively.And in the ROC analysis of MSV in GMV different regions,the MSV of left inferior frontal region owned the biggest AUC with the figure 0.678（P=0.021）,and the best cut-off value was-3.98 ppb,the corresponding sensitivity and specificity were 86.6%and 44.4%respectively.Besides in the ROC analysis of MSV in cortical thickness different brain regions,the biggest AUC of MSV was in right middle temporal gyrus which is 0.669（P=0.029）,when the cut-off value was-0.59 ppb,the sensitivity and specificity were 56.7%and 85.2%respectively.Conclusion:1.QSM can quantitatively measure the iron accumulation of the whole-brain cortex and deep gray nucleus in the patient with Alzheimer’s disease.2.Our research showed increased iron deposited in partial deep gray nucleuses and multiple cortical regions,including parahippocampcal,olfactory cortex and cerebellum,which further can be one of the robust imaging biomarkers of Alzheimer’s disease.3.In the cortical atrophy areas of AD,the iron accumulation can also be found.But there was no correlation between iron deposition and GMV or cortical thickness.4.The magnetic susceptibility value of left putamen presented the best effectiveness for AD diagnosis.