Background: Neuromyelitis optica spectrum disorders (NMOSD) encompasses a group of rare inflammatory diseases which primarily target the optic nerves, spinal cord, and brain. Typically, magnetic resonance imaging (MRI) data from single-center studies comprise 20-50 patients, limiting statistical power for outcomes research. Using retrospective data from the PArallel MRI in NmOsd (PAMRINO) study, a novel prospective NMOSD image repository (NMOsDIR) representing multiple international sites was coordinated by Charité-Universitätsmedizin Berlin and the Medical Image Analysis Center (Basel).

Objectives: The PAMRINO study aimed to investigate and analyze retrospective MRIs collected from NMOSD-specialized centers, potentially for the evaluation of disease-related brain and spinal cord changes. NMOsDIR serves as an international imaging research resource (comprising standardized retinal optical coherence tomography and MRI scans) and clinical data hub for prospective studies in NMOSD. Linking imaging and clinical data, as well as enabling analysis pipelines for each modality, will facilitate multi-centered studies using sufficient data and statistical power to advance outcomes research in this rare disease.

Methods: For clinical data collection in PAMRINO, a Research Electronic Data Capture (REDCap) platform was used, where participating centers entered data relevant for NMOSD patient monitoring. An image database (XNAT) was established for image uploads. This large collection of MRI data is currently being analyzed in a joint international effort of NMOSD clinical neuroradiologists and scientists.

Results: Brain, spinal cord, and optic nerve MRI scans with associated clinical data were collected from 514 NMOSD patients and 56 healthy controls from 17 international centers. Roughly 20,000 individual MRI scans from patients and healthy controls were collected. Of these, 78% had T1-weighted cerebral MRIs (55% with 3D scans), 80% had T2-weighted cerebral MRIs (54% with 3D scans), 86% had T2-weighted spinal cord MRIs (55% with 3D scans), and 35% had optic nerve MRIs.

Conclusions: We successfully established PAMRINO, an international collaborative retrospective MRI and clinical data repository. The knowledge gained during this process provided important new insights, where the initial analysis of the dataset has underscored the large degree of heterogeneity in image and clinical data collection in NMOSD-specialized centers. Thus, calling for more standardized methods of data acquisition and imaging analysis, as not to limit research opportunities. The new longitudinal, prospective NMOsDIR will help us to answer many pressing - yet open - questions regarding patients seropositive for aquaporin-4-IgG+, myelin oligodendrocyte glycoprotein-IgG+ and other autoimmune-related diseases. In turn, such a strategy will strengthen future capabilities in research, diagnosis, monitoring and improving NMOSD patient care.