Multimode MRI Research On Brain Of Systemic Lupus Erythematosus Without Neuropsychiatric Symptoms

Part one:Non-Gaussian diffusion MRI assessment of brain microstructure in systemic lupus erythematosus individuals without neuropsychiatric symptomsObjective:We investigated the capability of DKI parameters for detecting microstructural changes in both gray matter and white matter in systemic lupus erythematosus patients without neuropsychiatric symptoms and sought to determine whether these DKI parameters could serve as imaging biomarkers for early diagnosis and to further understand the brain damage mechanism.We also assessed the correlations between white matter DKI parameters and course of the disease, SLEDAI and the dose of corticosteroid.Materials and Methods:1. SubjectsA total of 31 patients with non-NPSLE were recruited from Nanfang Hospital, Guangzhou, China in this study from March 2014 to January 2015. Inclusion criteria as follows:female patients with ages between 18 and 50 years old; diagnosed SLE patients according to 1997 ACR SLE classification standards; diagnosed non-NPSLE patients according to 1999 NPSLE classification standards; right-handed; ethnic Han; patients could cooperate with the examine; the images of the brain were normal on routine MRI sequences; patients without anxiety and depression symptoms. Exclusion criteria included:psychiatric disorders or major neurologic disorders (e.g., severe head injury, stroke, epilepsy or visible lesions); ischemic diseases including acute ischemic cerebrovascular disease, acute peripheral arterial occlusion, advanced liver or heart failure; a history of diabetes; and substance abuse including drugs, alcohol and cigarettes, a history of CNS infections. We recruited 31 age, sex and education levels matched health controls. Written informed consent was obtained from each participant. We also recorded the course of the disease and the dose of corticosteroid, calculated the SLEDAI scores.2. DKI and 3DT1WI data acquisitionMRI data were obtained using a 3.0T MR scanner (Philips Achiva 3.0T) with a 8 channel head coil to receive the signal. The conventional imaging sequences including T2-weighted images and T2-FLAIR images were obtained for every subject to detect clinically lesions. DKI data were acquired using SE-EPI sequence. The DKI acquisition parameters were as follows:repetition time (TR)= 2000ms, echo time (TE)=69ms, In-plane resolution= 88x87, FOV=224mm×224mm×132mm, NSA=2. A total of 44 axial slices of 3.0 mm thickness were collected with no intersection gap. Thirty-two diffusion gradient directions with b-value=0,1000 s/mm2 and 2000 s/mm2. Each DKI scan lasted about 19 minutes.3DT1WI data were acquired using gradient echo FFE sequence. The 3DT1WI acquisition parameters were as follows: repetition time (TR)=9.0ms, echo time (TE)=4.0ms, In-plane resolution=256x256, FOV=256mmx256mmx 176mm, NSA=1. A total of 176 saggital slices of 1.0 mm thickness were collected with no intersection gap. Each 3DT1WI scan lasted about 8 minutes. 3. DKI data post-processing①All the DICOM images were converted to NIFTI format using MRIcron. ② Diffusion-weighted images were first corrected for eddy-current distortion and head motion using FSL in reference to the b0 images. ③ Non-brain tissues of DKI images were removed by BET. ④The apparent diffusion and kurtosis coefficients were calculated to obtain the diffusional tensor and kurtosis tensor, which were used to derive fractional anisotropy, mean diffusivity and mean kurtosis using DKE. ⑤ Non-brain tissues of 3DT1WI images were removed by FSL. ⑥ For each subject, T1-weighted images and b0 images were first oriented in the same direction. The T1-weighted images were then co-registered to the b0 images. Next, the co-registered T1-weighted images were segmented into GM, WM and cerebrospinal fluid images. Last the three DKI parameter images were normalized by the matrix of standard space. ⑦Regions of interest for major white matter tracts were defined by an intersection of the individual spatially normalized WM probability map thresholded at 50% with the JHU ICBM-DTI-81 white matter labels atlas supplied with FSL. Cortical gray matter regions were similarly defined by the intersections of the individual spatially normalized GM and WM probability maps (respectively, each thresholded at 50%) with the Harvard-Oxford cortical atlas. Subcortical structures were defined using FIRST. ⑧Quantification analyses of WM and GM in the regions of interest were extracted by software based on MATLAB 2013a programs.4. Statistic analysisTwo-sample t-test was performed with age, years of education between non-NPSLE patients and controls. Two-sample t-test was performed with selected brain white matter regions in mean FA value, MD value and MK value between non-NPSLE patients and controls. Two-sample t-test was performed with selected brain gray matter regions in mean MD value and MK value between non-NPSLE patients and controls. We used a statistical significance level of P<0.05. Then, we used Spearman correlation analysis to investigate the relationship between mean FA value, MD value and MK value of non-NPSLE patients and disease duration, SLEDAI and the dose of corticosteroid.Results1. Compared to the control subjects, non-NPSLE patients revealed significantly lower MK values in the body of the CC and fornix; significantly higher MD values in the body of the CC, the splenium of the CC, fornix, bilateral posterior thalamic radiation, right inferior longitidinal fasciculus and inferior fronto-occipital fasciculus; significantly lower FA values in bilateral corticospinal tract and right posterior corona radiate.2. Compared to the control subjects, non-NPSLE patients revealed significantly higher MD values in left thalamus and bilateral caudate. Compared to the control subjects, non-NPSLE patients revealed significantly higher MD values and MK values in many brain regions in frontal lobe, temporal lobe, occipital lobe and parietal lobe. There were only significant differences in MD values of precentral gyrus, temporal pole, postcentral gyrus, intracalcarine cortex, subcallosal cortex and parahippocampal gyrus. There were only significant differences in MK values of inferior temporal gyrus, anterior division.3. There were significant positive correlation between MD values in the body of the CC, the splenium of the CC and fornix and the duration of the disease. There were significant negative correlation between MK values in the body of the CC and the duration of the disease. There were significant negative correlation between FA values in right posterior corona radiate and the duration of the disease. There were no significant associations between DKI measurements and SLEDAI and the dose of corticosteroid.Conclusion1. Abnormal microstructure changes of both gray matter and white matter already happened in SLE even when the routine MRI findings are negative or nonspecific. So it is necessary for early assessment and intervention in patients with systemic lupus erythematosus.2. DKI could detect microstructure changes of white matter in SLE more earlier and sensitivily than DTI and white matter damage degree increase with the course of the disease. DKI also can be applied to the research of gray matter microstructure damage.3. We may suppose that in SLE gray and white matter alterations relate to different underlying mechanisms:white matter lesions mainly related to demyelination and axonal injury and loss; gray matter injury mainly related to the loss of neuronal cell and glial cell proliferation.Part two:Research on brain functional network of systemic lupus erythematosus individuals without neuropsychiatric symptoms via Resting-State fMRIObjective:This study aimed to investigate alterations of functional connectivity and small-world topological properties related to systemic lupus erythematosus individuals without neuropsychiatric symptoms through functional connection technology and graph-theory based on resting-state functional MRI.Materials and Methods:1. SubjectsA total of 42 patients with non-NPSLE were recruited from Nanfang Hospital, Guangzhou, China in this study from March 2014 to January 2015. Inclusion criteria as follows:female patients with ages between 18 and 50 years old; diagnosed SLE patients according to 1997 ACR SLE classification standards; diagnosed non-NPSLE patients according to 1999 NPSLE classification standards; right-handed; ethnic Han; patients could cooperate with the examine; the images of the brain were normal on routine MRI sequences; patients without anxiety and depression symptoms. Exclusion criteria included:psychiatric disorders or major neurologic disorders (e.g., severe head injury, stroke, epilepsy or visible lesions); ischemic diseases including acute ischemic cerebrovascular disease, acute peripheral arterial occlusion, advanced liver or heart failure; a history of diabetes; and substance abuse including drugs, alcohol and cigarettes, a history of CNS infections. We recruited 33 age, sex and education levels matched health controls. Written informed consent was obtained from each participant. We also recorded the course of the disease and the dose of corticosteroid, calculated the SLEDAI scores.2. Data acquisitionMRI data were obtained using a 3.0T MR scanner (Philips Achiva 3.0T) with a 8 channel head coil to receive the signal. The conventional imaging sequences including T2-weighted images and T2-FLAIR images were obtained for every subject to detect clinically lesions. Resting-state fMRI images covering the whole brain were acquired axially using an gradient echo-echo planar imaging sequence (GRE-EPI), TR= 2000ms, TE=35ms, flip angle=90°, matrix=64x64, FOV=230mm×230mm× 141mm, slice thickness=3.6mm, slice gap= 0.7mm, NSA=1. For each subjects, the resting state fMRI scanning lasted eight minutes, thus collecting 240 volumes. 3. Data post-processingData preprocessing was carried out using SPM8 software.The reprocessing as follows: ①Data with DICOM format were converted to NIFTI format; ②The first 10 images were discarded to ensure the magnetization equilibrium; ③Slice timing, correct differences in image acquisition time between slices; ㏕hen the remaining images were realigned, the subjects would be excluded for head translation or rotation exceeded ±1mm or ±1°; ⑤Normalization, normalize images into a standard space by MNI template images which supplied with SPM8; ⑥The images were proceed with linear detrend and low-frequency filtering; ⑦Regression of several nuisance signals of white matter signal, cerebrospinal fluid signal and head-motion profiles. The images were segmented into 90 anatomical regions of interests (ROIs) (45 ROIs for each hemisphere) using anatomically labeled-90 (AAL-90) template. These anatomical ROIs were extracted by the MarsBar toolbox. The resting state BOLD time series were correlated region by region for each subject across the full length of the resting time series. Then a square 90X90 correlation matrix was obtained for each subject,4005 (C290=90×89/2=4005) inter-regional correlations were subjected to statistic analysis.The topological properties of the brain functional networks were defined on the basis of a 90 × 90 binary graph, which was consisted of nodes and edges, the edges between nodes could be constructed by applying a correlation matrice, the regional centroid of each node was positioned by using Brainnet viewer Version 1.42 software. 4. Statistic analysisPermutation test based on matlab 2013a was performed with functional connectivity between non-NPSLE patients and controls. To account for multiple comparisons, the false discovery rate method was applied, P<0.05 was supposed to be a significantly functional connectivity. Permutation test based on matlab 2013a was performed with network topological measures between non-NPSLE patients and controls. P<0.05 was supposed to be a significantly difference.Results1. Compared to the control subjects, non-NPSLE patients revealed significantly lower functional connectivity between specific ROIs, e.g. right hippocampus and right posterior cingulate gyrus; right superior parietal gyrus and right middle frontal gyrus, orbital part; right inferior parietal gyrus and right middle frontal gyrus, orbital part; left precuneus and left superior frontal gyrus, dorsolateral; left precuneus and left middle frontal gyrus; right precuneus and right superior frontal gyrus, dorsolateral; right parahippocampal gyrus and right lenticular nucleus, putamen; left pallidum and left superior frontal gyrus, dorsolateral; right pallidum and right parahippocampal gyrus; left thalamus and right olfactory cortex; left thalamus and left hippocampus; right olfactory cortex and right heschl gyrus; left inferior temporal gyrus and right caudate nucleus; right inferior temporal gyrus and median cingulate and paracingulate gyri.2. A graph-based network efficiency analysis revealed that all the networks obeyed small-world organizations, statistical comparisons revealed significant differences in the quantitative network measures between the two groups. Compared to the HC group, the non-NPSLE patients showed significantly decreased global efficiency and clustering coefficient (P<0.05)3. Compared to the HC group, the non-NPSLE patients showed significantly decreased node parameters. The significantly decreased node degree regions located at frontal lobe, subcortical tissue and parietal lobe. The significantly decreased node efficiency regions located at occipital lobe, frontal lobe and parietal lobe.ConclusionSignificant weakened functional connectivity between multiple brain regions were found in non-NPSLE, mainly related to DMN and cortico-basal ganglia-thalamic-cortical circuit. The topology structure of functional network of non-NPSLE was damaged and the information transmission ability was reduced both in local and whole brain. Our findings provide evidence for the disconnection nature of non-NPSLE’s brain and therefore have important implications for understanding the pathogenesis of the disease.