3T diffusion MRI inter-site reproducibility dataset for cerebral small vessel disease.

Summary Metrics from diffusion imaging are sensitive and robust markers for microstructural tissue alterations in cerebral small vessel disease (SVD). Reproducibility of diffusion metrics is essential for the use in clinical routine and multi-center trials. We therefore provide this inter-site (inter-scanner) dataset of 10 patients with the genetically-defined small vessel disease CADASIL, who have been scanned on 2 different 3 Tesla Siemens MRI scanners (Magnetom Prisma und Magnetom Skyra) within less than 24 hours. The two scanners were located at the same hospital, but in different buildings with independent infrastructure, which is why we consider this an inter-site dataset. The dataset has already been used to determine the reproducibility of commonly assessed diffusion metrics (Konieczny & Dewenter et al., Neurology 2020, https://doi.org/10.1212/wnl.0000000000011213). Terms of Use The dataset is provided under CC BY-NC 4.0 license and can be freely downloaded and used for non-commercial purposes. The use of the dataset must be acknowledged through proper citation of Konieczny & Dewenter et al., Neurology 2020. Methods Patients (n=10) CADASIL patients (diagosed via molecular genetic testing or skin biopsy) were recruited at the Institute for Stroke and Dementia Research (Munich, Germany). Characteristics n=10 Age [years], median (IQR) 55.5 (13.3) Sex, female, n 5 (50%) Hypertension, n 3 (30%) Hypercholesterolaemia, n 5 (50%) Diabetes, n 0 (0%) Current or past smoking, n 3 (30%) WMH volume [% ICV ], median (IQR) 7.67 (2.62) Lacune count, median (IQR) 3 (7.75) Lacune volume [% ICV ], median (IQR) 0.01 (0.02) Microbleed count, median (IQR) 0.5 (1.75) Brain volume [% ICV], median (IQR) 75.7 (7.03) ICV = intracranial volume, IQR = inter-quartile range, WMH = white matter hyperintensities. MRI protocol Patients were scanned on 2 Siemens 3T Magnetom MRI scanners (Siemens Healthineers, Erlangen, Germany) in separate buildings at the LMU University Hospital. Key acquisition parameters (b-values of shells, diffusion-encoding directions, resolution) were harmonized between the two scanners. However, the weaker gradient system of the Magnetom Skyra resulted in longer repetition and echo times. Magnetom Skyra Magnetom Prisma Coil channels 64 (head-neck) 64 (head-neck) TR [ms] 3800 3200 TE [ms] 104.8 74 Flip angle [˚] 90 90 In-plane resolution [mm] 2 x 2 2 x 2 Slice thickness [mm] 2 2 Base resolution (matrix) 120 120 Number of slices 75 75 b-values [s/mm2] 1000/2000 1000/2000 Directions (per b-value) 30/60 30/60 b=0 [images] 10 10 Receiver bandwidth [Hz/px] 1894 1954 Parallel imaging acceleration factor 2 2 Multi-band acceleration factor 3 3 Data structure The data is stored in Nifti format and organized according to the BIDS standard. The data format (also of bval and bvec files) is according to FSL convention. We provide raw data (converted from DICOM via dcm2niix), and pre-processed data as described in the manuscript, preprocessed with dwidenoise & mrdegibbs (mrtrix), TOPUP & EDDY (FSL). /sub-01 /ses-prisma /dwi_raw /dwi_preprocessed /ses-skyra /dwi_raw /dwi_preprocessed Data privacy For privacy reasons, only an irreversibly anonymized, minimal dataset (only diffusion MRI) and group summary statistics (see patient table above) can be provided.