2026-10-13T03:54:32Z https://www.scidb.cn/oai

10.57760/sciencedb.0109y 2026-08-31T10:25:51Z

2026-08-31 DTI-ALPS in NMOSD doi:10.57760/sciencedb.0109y en Materials and MethodsParticipantsThis prospective cross-sectional observational study was conducted at the Neurology Department of the First Affiliated Hospital of Zhejiang University, School of Medicine between August 2021 and June 2024. We enrolled 44 patients with a confirmed diagnosis of NMOSD according to the 2015 International Consensus Diagnostic Criteria [1], comprising 21 patients in acute relapse and 23 in stable remission. Eighteen age- and sex-matched healthy volunteers were recruited as controls. All NMOSD patients had been receiving immunosuppressive therapy (glucocorticoid, mycophenolate mofetil, azathioprine, rituximab, or inebilizumab therapy) prior to enrollment.Acute status was defined as clinical relapse within 90 days of MRI scanning. Remission was defined as clinical stability for at least 90 days without new symptoms or MRI lesions. Exclusion criteria included: MRI contraindications, comorbid CNS disorders, severe systemic diseases affecting diffusion metrics, prior intracranial surgery, and a documented history of NMOSD brain syndrome.Clinical assessments included the Expanded Disability Status Scale (EDSS), relapse history (number of prior attacks), disease duration, lesion location, serum AQP4-IgG titer, and the interval from the last clinical onset to MRI scanning (days). This interval was computed as the difference between the date of the most recent clinical relapse and the MRI acquisition date, providing a quantitative temporal metric for assessing phase-dependent marker dynamics. All patients provided written informed consent. The study protocol was approved by the Ethics Committee of the First Affiliated Hospital of Zhejiang University (approval No. 779; 2018).MRI AcquisitionAll participants underwent MRI scanning at the Center for Brain Imaging Science and Technology, Zhejiang University, on a 3.0 T Siemens MAGNETOM Prisma scanner (Siemens, Erlangen, Germany) using a 32-channel head coil. Participants were positioned supine with foam padding around the head to reduce motion and were instructed to keep still with eyes closed during imaging.The imaging protocol included: (1) 3D T1-weighted magnetization-prepared rapid gradient echo (MPRAGE) sequence for anatomical reference (160 sagittal slices, slice thickness/gap = 1.2/0 mm, in-plane matrix = 512 × 512, repetition time [TR] = 5000 ms, echo time [TE] = 2.9 ms, inversion time [TI] = 700 ms, flip angle = 4°, field of view [FOV] = 256 × 256 mm²); and (2) diffusion tensor imaging (DTI) with 32 non-collinear diffusion gradients (b = 1000 s/mm²) and one b = 0 reference image.DTI PreprocessingDTI data were preprocessed using FSL (FMRIB Software Library, version 6.0.4). Eddy current-induced distortions and subject motion were corrected using the eddy tool with outlier replacement and concurrent b-matrix reorientation to compensate for rotational head motion. Brain extraction was performed using the Brain Extraction Tool (BET) with a fractional intensity threshold of 0.3. Diffusion tensors were estimated at each voxel using a weighted linear least-squares algorithm to generate fractional anisotropy (FA), mean diffusivity (MD), and individual tensor element (Dxx, Dyy, Dzz, Dxy, Dxz, Dyz) maps. Preprocessed data underwent independent visual quality control; subjects exhibiting >2 mm maximum absolute displacement or >2° rotation in the motion parameter estimates, or visible signal dropout affecting the perivascular target regions, were excluded from further analysis.DTI-ALPS CalculationThe DTI-ALPS index was calculated according to the method originally described by Taoka et al. [12], adapted for standardized multi-subject analysis. Four spherical regions of interest (ROIs, radius = 4 mm) were defined in standard MNI152 1-mm space, positioned at the level of the lateral ventricle body to intersect the deep medullary veins where perivascular spaces are predominantly oriented along the left-right (x) axis: (1) left projection fiber ROI (ROI_proj_left) centered at (-26, -36, 32) mm; (2) right projection fiber ROI (ROI_proj_right) centered at (26, -36, 32) mm; (3) left association fiber ROI (ROI_asso_left) centered at (-32, -43, 32) mm; and (4) right association fiber ROI (ROI_asso_right) centered at (32, -43, 32) mm.These ROIs were mapped to each subject's native anterior commissure–posterior commissure (ACPC)-aligned DTI space using the inverse deformation field from MNI152 to ACPC transformation. Nonlinear registration was performed using Advanced Normalization Tools (ANTs; version 2.4.0) with symmetric normalization (SyN) and cross-correlation as the similarity metric. Each subject's non-diffusion-weighted b0 image was registered to the MNI152 template, and the resulting inverse warp field was applied to transform the ROIs into native DTI space. Nearest-neighbor interpolation was used for binary ROI mask resampling to preserve discrete voxel boundaries, followed by linear interpolation for tensor element extraction within the resampled ROIs. Subjects with >20% of ROI voxels exhibiting CSF-like diffusion characteristics (MD > 2.5 × 10⁻³ mm²/s) or excessive spatial deformation were excluded after independent anatomical verification.For each ROI, the diffusion tensor was extracted and the ALPS index was computed as the ratio of diffusivity along perivascular space directions relative to orthogonal fiber directions. For each hemisphere, the ALPS index was mathematically defined as:ALPS_left = Dx,proj_left / ((Dy,asso_left + Dz,proj_left) / 2)ALPS_right = Dx,proj_right / ((Dy,asso_right + Dz,proj_right) / 2)where Dx,proj represents the mean diffusivity along the x-axis within the projection fiber ROI (aligned with perivascular spaces along deep medullary veins), Dy,asso represents the mean diffusivity along the y-axis within the association fiber ROI, and Dz,proj represents the mean diffusivity along the z-axis within the projection fiber ROI. The mean ALPS index (ALPS-mean) was calculated as the arithmetic average of the bilateral values: ALPS-mean = (ALPS_left + ALPS_right) / 2.To statistically isolate the perivascular-specific diffusivity signature from non-specific white matter microstructural alterations (e.g., inflammatory demyelination, axonal loss, or vasogenic edema frequently observed in AQP4-IgG-positive NMOSD), mean FA and MD values within the same ROIs were included as covariates in group-level analyses. This adjustment accounts for potential confounding effects of local tissue integrity on glymphatic diffusivity metrics. Additionally, ALPS indices deviating more than 3 SD from the cohort mean were winsorized to mitigate the influence of extreme outliers. Age and sex were included as nuisance covariates in the final regression models unless otherwise specified. Neuromyelitis optica spectrum disorder; Aquaporin-4; DTI-ALPS; Neurofilament light chain; Perivascular spaces 魏瑞理 RESTRICTED dataset Science Data Bank