The need for such guidance is rooted in biology. AQP4+ NMOSD and MOGAD are antibody-mediated inflammatory conditions of the central nervous system, but they differ fundamentally in their immunopathology. AQP4 antibodies target a water channel concentrated on astrocytic endfeet, producing lesions that cluster around the ventricles, in the area postrema, and along regions of high aquaporin 4 expression. MOG antibodies, by contrast, attack a myelin surface protein, generating a distinct pattern of demyelination that often involves the optic nerves, spinal cord and, in children, the cerebral cortex. These differences translate into different imaging signatures, different disease courses and different treatment strategies, making accurate radiological interpretation a matter of clinical urgency.

Yet the panel acknowledges that the diagnostic landscape remains treacherous. Imaging features of the two diseases overlap with each other and with multiple sclerosis, lesions evolve in unpredictable ways over time, and acquisition protocols vary widely between centers. A patient presenting with optic neuritis or transverse myelitis may harbor any of these conditions, and the wrong diagnosis can lead to the wrong therapy, with potentially devastating consequences. The new recommendations therefore serve a dual purpose: standardizing practice and sharpening the differential diagnosis at the bedside.

At the heart of the document are standardized protocols. The authors define a core set of MRI sequences that should be performed in every patient, alongside optional sequences for specific clinical questions. Brain MRI, spinal cord imaging and orbital imaging each receive detailed attention, with guidance on the timing of scans at presentation, during acute attacks, and in the follow-up period. The recommendations extend the MAGNIMS framework, long established for multiple sclerosis, into these antibody-mediated diseases, creating a parallel infrastructure for image acquisition and reporting that can be implemented in routine clinical practice.

The characteristic conventional MRI findings of each disease receive careful treatment. In AQP4+ NMOSD, lesions typically affect the spinal cord over long segments, the area postrema causing intractable hiccups and nausea, periependymal surfaces of the ventricles, and the optic chiasm. In MOGAD, longitudinally extensive myelitis is also common but tends to involve the conus medullaris, while optic nerve involvement is frequently anterior and bilateral, and cortical lesions with seizures, including the FLAMES pattern of unilateral cortical FLAIR hyperintensity, point toward MOG antibodies. Recognizing these patterns helps clinicians distinguish the two diseases from each other and from multiple sclerosis, where short-segment cord lesions, central vein signs and paramagnetic rim lesions dominate.

One of the most clinically valuable sections addresses the acute attack. MRI plays a decisive role in determining whether a patient is experiencing a true relapse or a pseudo-relapse, a distinction that changes management entirely. The panel also highlights the phenomenon of radiologic lag in MOGAD, in which brain lesions may appear or evolve days to weeks after the clinical onset of an attack, meaning an early normal scan does not exclude active disease. Serial imaging during the attack window can therefore reveal lesion dynamics that a single time point misses, and the timing of follow-up scans matters as much as their content.

Post-attack recovery is another area where the recommendations break new ground. MOGAD lesions, unlike those of many other demyelinating diseases, have a striking tendency to resolve on T2-weighted images, and the panel discusses the timing and predictors of this lesion resolution. The degree of tissue recovery, visible as the fading of lesions and the relative preservation of tissue volume, carries prognostic weight and informs decisions about how aggressively to treat and how long to continue immunotherapy. Spinal cord atrophy, which correlates with disability in NMOSD, is identified as a key longitudinal measure.

Beyond conventional imaging, the document surveys the added value of non-conventional MRI techniques. Diffusion-based measures, myelin-sensitive sequences, magnetization transfer imaging and volumetric analysis can probe microstructural damage, inflammatory activity and tissue repair that standard T2 and T1 sequences cannot capture. Advanced approaches such as neurite orientation dispersion and density imaging, 7 Tesla imaging of myelin signal in normal-appearing white matter, and even indices of glymphatic system dysfunction are reviewed as research tools that may mature into clinical biomarkers. The panel also points to the emerging role of machine learning and radiomics, including deep learning models trained to distinguish MOGAD from multiple sclerosis, as a frontier for automated diagnostic support.

The recommendations confront one of the field’s most stubborn problems: seronegative NMOSD. Patients who present with the clinical and radiological phenotype of NMOSD but lack detectable AQP4 antibodies form a heterogeneous group whose classification remains uncertain, and whose response to therapy may differ from that of seropositive patients. The panel outlines the diagnostic complexity of this population and argues that novel imaging biomarkers are needed to improve classification and treatment precision, potentially allowing MRI itself to stratify patients when antibody testing is inconclusive.

Ultimately, the consensus document represents a maturing of the field. With targeted therapies now available for both AQP4+ NMOSD and MOGAD, and with growing interest in when immunotherapy can be safely discontinued, standardized imaging is no longer a luxury but a clinical necessity. By codifying what to scan, when to scan and how to interpret the results, the MAGNIMS panel has given neurologists, neuroradiologists and researchers a common language for these diseases, one that should accelerate both better individual care and the multicenter trials needed to push the field forward.

Subject of Research: Consensus recommendations for the use of MRI in diagnosing and monitoring neuromyelitis optica spectrum disorder and MOG antibody-associated disease

Article Title: Use of MRI in neuromyelitis optica spectrum disorder and myelin oligodendrocyte glycoprotein antibody-associated disease — MAGNIMS consensus recommendations

Article References: Cortese, R., Hacohen, Y., Arrambide, G., Bianchi, A., DeLuca, G., Filippi, M., Gasperini, C., Geraldes, R., Granziera, C., Haider, L., Messina, S., Sastre-Garriga, J., Preziosa, P., Rocca, M. A., Rovira, A., Tintoré, M., Toosy, A., Tortorella, C., Vaneckova, M., … Schoonheim, M. (2026). Use of MRI in neuromyelitis optica spectrum disorder and myelin oligodendrocyte glycoprotein antibody-associated disease — MAGNIMS consensus recommendations. Nature Reviews Neurology. https://doi.org/10.1038/s41582-026-01271-2

Image Credits: AI Generated

DOI: 10.1038/s41582-026-01271-2

Keywords: MRI, NMOSD, MOGAD, aquaporin-4 antibodies, myelin oligodendrocyte glycoprotein, demyelinating disease, neuroimmunology, MAGNIMS, optic neuritis, myelitis, diagnostic criteria, biomarkers

NMOSD is a rare, autoimmune inflammatory disorder of the central nervous system that predominantly targets the optic nerves and spinal cord. In most patients, the disease is driven by pathogenic immunoglobulin G antibodies directed against aquaporin-4 (AQP4), a water channel protein abundantly expressed on astrocytes. Each relapse can leave permanent neurological damage, ranging from irreversible vision loss to paralysis, which is why suppressing the underlying B-cell–driven immune attack is the central goal of long-term management. Before targeted therapies earned regulatory approval, clinicians commonly prescribed oral immunosuppressants (ISTs) such as azathioprine, mycophenolate mofetil, and methotrexate off-label to keep the disease at bay.

Those older drugs have well-recognized shortcomings. Comparative studies have associated oral ISTs with a greater frequency of relapses and a shorter time to relapse after treatment initiation compared with approved monoclonal antibodies. They are also linked to gastrointestinal and hematologic side effects and a high frequency of infections, problems that accumulate over years of continuous use. Inebilizumab, by contrast, is a humanized, affinity-optimized, glycoengineered monoclonal antibody that targets CD19, a marker expressed broadly across the B-cell lineage, depleting the cells responsible for producing the pathogenic AQP4 antibodies. The randomized, placebo-controlled N-MOmentum trial demonstrated its efficacy in AQP4-seropositive NMOSD and led to regulatory approval.

A key question remained, however. Many participants entering N-MOmentum had years of prior IST exposure, and it was unclear whether that treatment history might blunt or otherwise alter the drug’s long-term efficacy and safety. To investigate, researchers analyzed 202 of the 213 AQP4-seropositive participants from the randomized controlled period (RCP) and 197 of the 201 from the open-label period (OLP), splitting them roughly evenly between those with prior IST treatment and those who were IST-naïve. Only ISTs taken before day 1 of the trial were considered; ISTs were prohibited once the trial began.

The randomized portion of the trial delivered an unambiguous message. Participants on inebilizumab experienced far fewer adjudicated NMOSD attacks than those on placebo, and the magnitude of benefit was essentially identical in both subgroups. Among participants with prior IST use, the hazard ratio for attack was 0.21 (95% confidence interval 0.09–0.48), while among IST-naïve participants it was 0.23 (0.09–0.59). NMOSD-related inpatient hospitalizations were also less frequent with inebilizumab in the prior-IST group. Worsening on the Expanded Disability Status Scale (EDSS) was significantly less common with inebilizumab versus placebo in both groups: 19.2% versus 43.5% among those with prior IST exposure, and 13.6% versus 28.0% among those who were IST-naïve, both comparisons reaching nominal statistical significance.

Because the randomized controlled period lasted only about 28 weeks, the research team turned to modeling to assess long-term outcomes over years rather than months. In the any-INEB population, the adjusted annualized attack rate was 0.11 (0.07–0.17) for participants with prior IST use and 0.08 (0.05–0.14) for those without, a difference so small it is unlikely to matter clinically. A high probability of remaining attack-free persisted through week 286 of treatment in both groups, and EDSS scores actually improved during the open-label extension regardless of treatment history. NMOSD-related hospitalizations were equal in number across the two subgroups during long-term follow-up.

The most striking results emerged when inebilizumab was compared against synthetic historical comparator groups built from published Kaplan–Meier survival curves of patients treated with azathioprine or other broad-spectrum ISTs, or with placebo. Using a time-varying spline model with two internal knots, selected through formal information-criterion testing and visual fit assessment, the researchers found the time to NMOSD attack was significantly longer with inebilizumab than with azathioprine or other ISTs (hazard ratio 0.29; p < 0.001) and than with placebo (hazard ratio 0.15; p < 0.001). The modeled four-year attack-free probability was 77% for inebilizumab, 36% for azathioprine or other ISTs, and just 12% for placebo. Notably, the widening gap over time suggests inebilizumab’s relative advantage may grow the longer patients stay on treatment.

Safety outcomes were reassuringly similar across subgroups. Through the open-label period, roughly 92% of participants in both groups experienced at least one treatment-emergent adverse event, but drug-related events were somewhat more common among IST-naïve participants, 46.6% versus 30.9%. Infections, the adverse event category of greatest concern for a B-cell–depleting therapy, occurred at nearly identical rates in both groups, affecting 72.3% of those with prior IST use and 76.7% of those without, and were not increased by prior immunosuppressant exposure. Opportunistic infections were rare, one case in each group, no anaphylactic reactions were reported, and deaths were similarly uncommon.

The authors caution that this was a post hoc, exploratory analysis with limitations. Subgroup sample sizes were modest, randomization was not stratified by duration of IST history, and p values were not adjusted for multiple comparisons, so all statistical results are nominal. The synthetic historical comparators, reconstructed by digitizing published survival curves, introduce approximation error and cannot fully control for differences in study populations and designs across the source studies. These constraints mean the long-term comparative estimates should be interpreted as suggestive rather than definitive evidence from head-to-head trials, which have never been conducted.

Even so, the practical implications for patients and clinicians are substantial. Switching from off-label immunosuppressants to immunotherapies generally requires an overlapping treatment period to avoid triggering an attack upon discontinuation of the old drug, and many providers overlap ISTs for up to six months, although concurrent inebilizumab and IST use is not recommended as a routine regimen. The new analysis supports inebilizumab treatment for patients with AQP4-seropositive NMOSD regardless of whether they previously received first-line immunosuppressants, with sustained reductions in attack risk, stabilized or improved disability scores, and a safety profile consistent with the overall trial population. For the many patients still managed with older oral agents, the data offer a quantitative basis for a confident transition to targeted B-cell depletion.

Subject of Research: Long-term efficacy and safety of inebilizumab in AQP4-seropositive NMOSD patients with or without prior immunosuppressant use

Article Title: Efficacy of Inebilizumab in N‐MOmentum Trial Participants With or Without Prior Immunosuppressants

Article References: Cree, B. A. C., Suero, B., Walsh, S., Marignier, R., Lindsey, J. W., Kim, H. J., She, D., Cimbora, D., Cavida, D., & Paul, F. (2026). Efficacy of Inebilizumab in N‐ MOmentum Trial Participants With or Without Prior Immunosuppressants. Annals of Clinical and Translational Neurology, 13(9), 1930-1936. https://doi.org/10.1002/acn3.70426

Image Credits: AI Generated

DOI: 10.1002/acn3.70426

Keywords: inebilizumab, NMOSD, N-MOmentum trial, aquaporin-4, CD19 B-cell depletion, azathioprine, mycophenolate mofetil, immunosuppressants, clinical trial, neuroimmunology, EDSS, autoimmune disease