Multimodle MRI Study Of The Brain In Patients With Subcortical Ischemic Vascular Dementia

Background and purposeDementia is a syndrome with numerous symptoms, including loss of memory, judgment, reasoning, and changes in mood, behavior, and communication. Vascular dementia(Va D) is the second most common cause of dementia following Alzheimer’s disease(AD). Va D comprises different entities based on the vascular pathologies and clinical manifestations. Subcortical ischemic vascular dementia(SIVD) is a relatively homogeneous subtype of Va D subtype, which results from small vessel disease. It is characterized by subcortical lesions such as lacunes and incomplete white matter types of infarction which affect the inner parts of the brain.Currently there are many difficulties in the diagnosis of SIVD. The clinical diagnosis is mainly based on the cognitive impairment of neurocognitive test, evidence of subcortical vascular disease from medical imaging or clnical history, and the confirmation of their causal relationship. However, because SIVD is often caused by chronic ischemia and begins with no obvious symptoms, It is very difficult to identify the causal relationship between cerebrovascular disease and cognitive impairment by the time[1]. Objective examination methods are conventional medical imaging and neurophysiological examination. They both showed limited value for accurately diagnosis. Therefore it is important to investigate a more objective and specific biomarkers for the clinical diagnostics of SIVD.. The recent development of MRI has allowed researchers to detect the sturctral, metabolic and functional pattern of the brain. Multimode MRI technology can give us more objective and comprehensive understanding of the pathological changes of SIVD. It may help the clinical diagnosis SIVD, and provide valuableinsights into the pathomechanism of dementia.Material and methodThe present study was divided into three parts:1. Thirty five patients with SIVD and 35 control subjects were enrolled in this part. Clinical diagnosis of SIVD was made according to the criteria set by Erkinjuntti and the DSM-IV(Diagnostic and Statistical Manual of Mental Disorders, IV). All of them underwent Three-dimensional magnetization-prepared rapid gradient-echo imaging(3D MPRAGE) and rs-f MRI using an echo-planar imaging sequence(EPI)scan. The structural data were processed using the voxel-based morphometry 8 toolbox(VBM8). The rs-f MRI data were processed using Statistical Parametric Mapping(SPM8), Data Processing Assistant for Resting-State f MRI(DPARSF) software in three different frequency bands(global bands: 0.01-0.08 Hz; slow-5: 0.01-0.027 Hz; slow-4: 0.027-0.073 Hz). Five SIVD patients were excluded according to the criteria that individuals must not have an estimated maximum displacement in any direction greater than 2 mm or a head rotation greater than 2°. Within-group analysis was performed with a one-sample Student’s t-test to identify brain regions with ALFF value larger than the mean. Intergroup analysis was performed with a two-sample Student’s t-test to identify ALFF differences of whole brain between SIVD and control subjects. Partial correlations between ALFF values and Montreal Cognitive Assessment(Mo CA) and Mini-Mental State Examination(MMSE) scores were analyzed in the SIVD group across the parameters of age, gender, years of education, and GM volume.2. 34 SIVD patients and 35 healthy controls wereincluded in this part. Structural data using 3D MPRAGE scan was initially analyzed with FSL-VBM to measure gray matter volume. The cortical thickness was measured with the CIVET MRI analysis pipeline. Automated segmentation and labeling of the deep gray matter was performed using fsl_anat. Between-group comparisons were conducted to test for the differences in the VBM data using permutation-based nonparametric testing and in cortical thickness/surface area data using Surfstat toolbox based on random field theory. Correlations between neuropsychological results and the cortical thickness, deep gray matter measures in SIVD patients were performed.Furthermore, we selected 14 patients with Binswanger’s disease(BD) from SIVD patients and recruited 14 age, sex and education mathed patients with AD. Patients with probable AD fulfilled the NINCDS-ADRDA criteria. BD patients were diagnosed according to the criteria proposed by Erkinjuntti et al. The volumes of the hippocampus and amygdala, and morphologic parameters(volume, surface area, cortical thickness and mean curvature) of the entorhinal cortex(ERC) and perirhinal cortex(PRC) were calculated using an automated approach. Group differences in the morphometric results while controlling for age, gender and TIV. Partial correlation analyses were used for associations between brain morphometric results and cognitive scores(MMSE and Mo CA) while controlling for age and gender. All statistical analyses were conducted using SPSS software.3. Thirty five patients with SIVD and 35 healthy controls were included in this part. They all underwent high-resolution susceptibility-weighted imaging. The phase shift value of the bilateral hippocampus(HP), caudate nucleus(CN), globus pallidus(GP), putamen(PU), thalamus(TH), red nucleus(RN), substantia nigra(SN), anterior cingulate cortex, posterior cingulate(ACC), parietal cortex(PC) and frontal white matter(FWM) were calculate by SPIN software and correlated with neuropsychological scores for SIVD patients. Age, gender, education level and white matter hyperintensities(WMH)(grading scales presented by Fazekas) of each subject were imported as covariates in the statistical analysis. SPSS software package was used for the statistics.Result:1. In the global band of resting state(0.01-0.08 Hz) study, within-group analysis showed that the bilateral anterior cingulate cortex(ACC), posterior cingulate cortex, medial prefrontal cortex(MPFC), inferior parietal lobe(IPL), occipital lobe, and adjacent precuneus had significantly higher standardized ALFF values than the global mean ALFF value in both groups. Compared to the controls, patients with SIVD presented lower ALFF values in the bilateral precuneus(t=-3.68, p < 0.01) and higher ALFF values in the bilateral ACC(t=4.57, p < 0.01), left insula(t=4.25, p < 0.01) and hippocampus(t=4.02, p < 0.01) with Alphasim corrected. Including GM volume as an extra covariate, the ALFF inter-group difference exhibited highly similar spatial patterns to those without GM volume correcting. There were negative correlations between the Mo CA scores and the average ALFF values of the left insula(r =-0.388, p = 0.046) and the left hippocampus(r =-0.564, p = 0.002). Negative correlation between the MMSE scores and the average ALFF values of the left insula(r =-0.542, p = 0.004) and the left hippocampus(r =-0.421, p = 0.029) were also found.In slow-5 band, compared with the controls, the SIVD patients exhibited significant higher ALFF in the bilateral anterior cingulate cortex(t=4.65), right putamen(t=3.59) and right supplementarymotor area(t=4.67), while lower ALFF in the right precuneus(t=-3.83) and right angular gyrus(t=-3.95)(p < 0.05, FWE corrected). In the slow-4 band, SIVD patients relative to controls exhibited increased ALFF in the bilateral anterior cingulate cortex(t=4.51), right putamen(t=4.53) and left fusiform gyrus(t=3.68)(p < 0.05, FWE corrected). A close correlation was found between the ALFF value of the right angular gyrus at slow-5 and activities of daily living(ADL) scores(r = 0.378, p = 0.039).2. SIVD patients had significant gray matter volume reduction compared to controls in bilateral temporal lobes, including fusiform gyrus, lateral temporal lobes and medial temporal lobe(hippocampus and parahippocampal gyrus); bilateral orbitofrontal cortices, left supramarginal gyrus, thalamus and right cerebellar region. SIVD patients showed significant cortical thinning in bilateral perisylvian areas, concretely in the lateral temporal lobe, supramarginal gyrus, insula, precentral-postcentral association region and inferior frontal gyrus that extended to the orbitofrontal cortex. There were significant group differences(p ≤ 0.005) of volume in most subcortical gray matter. Mo CA scores showed a positive correlation trend with the cortical thickness in the right hippocampus and anterior cingulate(p < 0.01, uncorrected). Mo CA scores were positively(r = 0.62, p = 0.01) correlated with the volumes of the right thalamus(β = 0.45, p = 0.02) and caudate nucleus(β = 0.46, p = 0.01). CDR scores were also negatively(r = 0.65, p = 0.007) related to the volumes of the left amygdala(β =-0.58, p = 0.004) and right caudate nucleus(β =-0.39, p = 0.03).Volume reduction of the hippocampus, amygdala, ERC, and disturbance of the PRC curvature was found in both AD and BD patients compared with the controls(p < 0.05, uncorrected). There was no significant differences among all the structural measures between the AD and BD patients. Compared with the controls, the AD patients exhibited significantly decreased volume and cortical thickness in the ERC, and increased mean curvature in the PRC; while the BD patients only showed significant increase in mean curvature(p < 0.005) and a reduced trend in surface area(p = 0.09) and volume(p = 0.07) in the PRC. The partial correlation analyses revealed significant associations between ERC thickness and Mo CA scores(r = 0.70, p = 0.01) in the AD group. In the BD group, only PRC volume was significantly associated with Mo CA(r = 0.62, p = 0.03) and MMSE(r = 0. 70, p = 0.01) scores. There is a close relationship between the hippocampal volume and the volume of each other structure(r = 0.91, p < 0.001 for amygdala; r = 0.93, p < 0.001 for ERC; r = 0.95, p < 0.001 for PRC) in the AD group. In the BD group, the hippocampal volume was also correlated with the volume of the amygdala(r = 0.60, p = 0.04) and ERC(r = 0.64, p = 0.03).3. There is a close positive correlation(r=0.902, P=0.002) between SWI phase shift values and the published regional real iron concentrations. Patients with SIVD exhibited significant increased phase shift values in the bilateral HP, CN, PU, right GP and left SN(p < 0.05). No significant phase shift differences was found in the bilateral PC, TH, FWM, RN, ACC, PCC, left GP and right SN between SIVD and control groups(p < 0.05). Close correlations were found between the phase shift value of the left HP and HIS scores(r=-0.043, P=0.016), right CN and ADL scores(r=-0.395, P=0.031).Conclusion1. There is a spontaneous aberrant activity characteristic for SIVD and that measurement of the ALFFs in the above regions may aid in its detection. In different frequency bands, the abnormalities of brain function in SIVD patients exhibited different spatial patterns. ALFF of right angular gyrus at slow-5 band is more specific for SIVD and may be a useful tool to help SIVD diagnosis. The results highlight abnormalities of ALFF in SIVD and provide novel insights into the understanding of the pathophysiological mechanism of it.2. We demonstrate volume reductions and cortical thinning of the hippocampus and their associated outflow areas in SIVD. Besides, our study confirms the widespread cortical thinning in the perisylvian area. Changes in anterior cingulate and caudate nucleus might subserve the characteristic impairment of executive function in SIVD.AD and BD exhibit similar atrophy patterns in the medial temporal cortices and deep gray matte. Although atrophy of the MTL structures is a sensitive biomarker for AD, it is not superior for discrimination between AD and BD. Our results lend new insight into the impaired patterns of MTL structures in AD and BD patients, and provide an update on biomarkers for the prediction of cognition in demented patients.3. Our data demonstrate that SIVD patients have widespread abnormal brain iron deposition, which showed close correlation with the severity of cognitive impairment. Brain iron deposition may be a new biomarker of SIVD and play an important role in the pathophysiological mechanism.