Severe disease exacerbation in a patient with neuromyelitis optica spectrum disorder during treatment with dimethyl fumarate

Transformer-Based Deep-Learning Algorithm for Discriminating Demyelinating Diseases of the Central Nervous System With Neuroimaging

Background

Differential diagnosis of demyelinating diseases of the central nervous system is a challenging task that is prone to errors and inconsistent reading, requiring expertise and additional examination approaches. Advancements in deep-learning-based image interpretations allow for prompt and automated analyses of conventional magnetic resonance imaging (MRI), which can be utilized in classifying multi-sequence MRI, and thus may help in subsequent treatment referral.

Methods

Imaging and clinical data from 290 patients diagnosed with demyelinating diseases from August 2013 to October 2021 were included for analysis, including 67 patients with multiple sclerosis (MS), 162 patients with aquaporin 4 antibody-positive (AQP4+) neuromyelitis optica spectrum disorder (NMOSD), and 61 patients with myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD). Considering the heterogeneous nature of lesion size and distribution in demyelinating diseases, multi-modal MRI of brain and/or spinal cord were utilized to build the deep-learning model. This novel transformer-based deep-learning model architecture was designed to be versatile in handling with multiple image sequences (coronal T2-weighted and sagittal T2-fluid attenuation inversion recovery) and scanning locations (brain and spinal cord) for differentiating among MS, NMOSD, and MOGAD. Model performances were evaluated using the area under the receiver operating curve (AUC) and the confusion matrices measurements. The classification accuracy between the fusion model and the neuroradiological raters was also compared.

Results

The fusion model that was trained with combined brain and spinal cord MRI achieved an overall improved performance, with the AUC of 0.933 (95%CI: 0.848, 0.991), 0.942 (95%CI: 0.879, 0.987) and 0.803 (95%CI: 0.629, 0.949) for MS, AQP4+ NMOSD, and MOGAD, respectively. This exceeded the performance using the brain or spinal cord MRI alone for the identification of the AQP4+ NMOSD (AUC of 0.940, brain only and 0.689, spinal cord only) and MOGAD (0.782, brain only and 0.714, spinal cord only). In the multi-category classification, the fusion model had an accuracy of 81.4%, which was significantly higher compared to rater 1 (64.4%, p=0.04<0.05) and comparable to rater 2 (74.6%, p=0.388).

Conclusion

The proposed novel transformer-based model showed desirable performance in the differentiation of MS, AQP4+ NMOSD, and MOGAD on brain and spinal cord MRI, which is comparable to that of neuroradiologists. Our model is thus applicable for interpretating conventional MRI in the differential diagnosis of demyelinating diseases with overlapping lesions.

Development and validation of a simple and practical method for differentiating MS from other neuroinflammatory disorders based on lesion distribution on brain MRI

There is an unmet need to develop practical methods for differentiating multiple sclerosis (MS) from other neuroinflammatory disorders using standard brain MRI. To develop a practical approach for differentiating MS from neuromyelitis optica spectrum disorder (NMOSD) and MOG antibody-associated disorder (MOGAD) with brain MRI, we first identified lesion locations in the brain that are suggestive of MS-associated demyelination ("MS Lesion Checklist") and compared frequencies of brain lesions in the "MS Lesion Checklist" locations in a development sample of patients (n = 82) with clinically definite MS, NMOSD, and MOGAD. Patients with MS were more likely than patients with non-MS to have lesions in 3 locations only: anterior temporal horn (p < 0.0001), periventricular ("Dawson's finger") (p < 0.0001), and cerebellar hemisphere (p = 0.02). These three lesion locations were used as predictor variables in a multivariable regression model for discriminating MS from non-MS. The model had area under the curve (AUC) of 0.853 (95% confidence interval: 0.76-0.945), sensitivity of 87.1%, and specificity of 72.5%. We then used an independent validation sample with equal representation of MS and NMOSD/MOGAD cases (n = 97) to validate our prediction model. In the validation sample, the model was 76.3% accurate in discriminating MS from non-MS. Our simple method for predicting MS versus NMOSD/MOGAD only requires a neuroradiologist or clinician to ascertain the presence of lesions in three locations on conventional MRI sequences. It can therefore be readily applied in the real-world setting for training and clinical practice.

[A patient with multiple sclerosis who developed bilateral optic neuritis and central trigeminal myelin lesion]

A 70-year-old woman was admitted to our hospital due to bilateral optic neuritis and left facial sensory disturbance that became exacerbated over 10 days. Both serum and cerebrospinal fluid (CSF) were negative for aquaporin 4 antibody and myelin oligodendrocyte glycoprotein antibody. A high level of myelin basic protein (MBP) in her CSF was observed. Brain MRI showed a high T₂ signal and contrast enhancement of the bilateral optic nerve, intramedullary tract and central myelin lesion in the trigeminal nerve. After intravenous methylprednisolone pulse therapy, her visual impairment and facial sensory disturbance gradually improved. She was diagnosed with clinically isolated syndrome, based on 2017 McDonald criteria. A diagnosis of multiple sclerosis (MS) was suspected due to the trigeminal myelin lesion confined to the central myelin portion and high level of MBP in the CSF. Treatment with dimethyl fumarate has been effective for preventing recurrence in 13 months in this patient. The central-peripheral myelin transitional zone at the trigeminal nerve is located 1-6 mm from the pons, where central myelin changes to the peripheral myelin. This patient showed a high T₂ signal at the trigeminal nerve 3 mm from the pons on MRI, suggesting the involvement of a central trigeminal myelin lesion. Findings of a central trigeminal myelin lesion on MRI may aid in differentiating between MS and seronegative neuromyelitis optica spectrum disorder.

FTY720 Exacerbates Blood-Brain Barrier Dysfunction Induced by IgG Derived from Patients with NMO and MOG Disease

Neuromyelitis optica (NMO) and myelin oligodendrocyte glycoprotein (MOG) antibody-related disease (MOG disease) are inflammatory demyelinating diseases of the central nervous system (CNS). The disruption of the blood-brain barrier (BBB) is considered a key step in the pathogenesis of NMO and MOG disease. Although a previous report indicated that circulating immunoglobulin G (IgG) from NMO patients disrupts the BBB, the effect of IgG from patients with MOG disease has not been elucidated. In addition, it has been reported that some disease-modifying drugs for multiple sclerosis are harmful to NMO by an unknown mechanism. This study aimed to examine the effects of IgG from patients with NMO or MOG disease on BBB integrity. We also examined the effects of disease-modifying drugs (fingolimod [FTY720] and dimethyl fumarate [DMF]) on IgG-treated brain capillary endothelial cells. We used in vitro BBB models constructed with rat brain capillary endothelial cells (RBECs) to examine the effects on BBB function. The integrity of the RBECs was assessed by measuring transendothelial resistance (TEER) and cell viability. NMO or MOG-IgG treatment decreased TEER and cell viability in the endothelial monolayer model. Although FTY720 and DMF did not affect barrier function or cell viability under normal conditions, disease IgG-induced barrier dysfunctions were worsened by the presence of FTY720. These data indicate that circulating IgG in patients with NMO or MOG disease worsens BBB function. Furthermore, in patients with NMO or MOG disease treated with FTY720, changes in the integrity of the BBB were found to exacerbate the disease.

Differentiation between multiple sclerosis and neuromyelitis optica spectrum disorders by multiparametric quantitative MRI using convolutional neural network

Multiple sclerosis and neuromyelitis optica spectrum disorders are both neuroinflammatory diseases and have overlapping clinical manifestations. We developed a convolutional neural network model that differentiates between the two based on magnetic resonance imaging data. Thirty-five patients with relapsing-remitting multiple sclerosis and eighteen age-, sex-, disease duration-, and Expanded Disease Status Scale-matched patients with anti-aquaporin-4 antibody-positive neuromyelitis optica spectrum disorders were included in this study. All patients were scanned on a 3-T scanner using a multi-dynamic multi-echo sequence that simultaneously measures R1 and R2 relaxation rates and proton density. R1, R2, and proton density maps were analyzed using our convolutional neural network model. To avoid overfitting on a small dataset, we aimed to separate features of images into those specific to an image and those common to the group, based on SqueezeNet. We used only common features for classification. Leave-one-out cross validation was performed to evaluate the performance of the model. The area under the receiver operating characteristic curve of the developed convolutional neural network model for differentiating between the two disorders was 0.859. The sensitivity to multiple sclerosis and neuromyelitis optica spectrum disorders, and accuracy were 80.0%, 83.3%, and 81.1%, respectively. In conclusion, we developed a convolutional neural network model that differentiates between multiple sclerosis and neuromyelitis optica spectrum disorders, and which is designed to avoid overfitting on small training datasets. Our proposed algorithm may facilitate a differential diagnosis of these diseases in clinical practice.

Severe Relapses of Neuromyelitis Optica Spectrum Disorder During Treatment With Dimethyl Fumarate

ABSTRACT

Neuromyelitis optica spectrum disorder (NMOSD) is an autoimmune demyelinating disease of the central nervous system, characterized by a specific antibody that selectively binds aquaporin-4 channel.This is a report of an NMOSD case, with severe attacks of optic neuritis and myelitis after initiate of dimethyl fumarate (DMF).We suggested that DMF could deteriorate the neuromyelitis optica (NMO) disease course, which results in serious morbidity and mortality in patients. Thus, initiation of DMF should be avoided before ruling out NMOSD in patients experiencing demyelinating attacks, especially in the case of recurrent optic neuritis or myelopathy and concurrency of other rheumatologic diseases.

Update on neuromyelitis optica spectrum disorder

PURPOSE OF REVIEW

Neuromyelitis optica spectrum disorder is an autoimmune disease that causes optic neuritis and transverse myelitis. Attacks can cause severe neurological damage leading to blindness and paralysis. Understanding of the immunopathogenesis of this disease has led to major breakthroughs in diagnosis and treatment. In the past 18 months, three successful phase 3 clinical trials have been published using targeted approaches to preventing relapses.

RECENT FINDINGS

Updates in epidemiology, imaging, quality of life and treatment for acute relapse and prevention have been published in the past 18 months. Epidemiology studies are distinguishing patients based on their antigen specificity for aquaporin-4 and myelin oligodendrocyte glycoprotein, which are increasingly recognized as separate immunological conditions. Imaging by MRI and optical coherence tomography continue to be developed as tools to distinguish neuromyelitis optica spectrum disorders (NMOSD) from other diseases. This is especially relevant as the recent clinical trials showed differences in response between aquaporin-4 seropositive and seronegative patients. The three drugs that were tested for prevention of NMOSD relapses were eculizumab, inebilizumab, and satralizumab. All of the trials were worldwide, placebo-controlled, double-masked studies that demonstrated a clear benefit with each approach.

SUMMARY

Recent research in NMOSD has resulted in improved diagnosis and approved treatments.

Differential Effects of MS Therapeutics on B Cells-Implications for Their Use and Failure in AQP4-Positive NMOSD Patients

B cells are considered major contributors to multiple sclerosis (MS) pathophysiology. While lately approved disease-modifying drugs like ocrelizumab deplete B cells directly, most MS medications were not primarily designed to target B cells. Here, we review the current understanding how approved MS medications affect peripheral B lymphocytes in humans. These highly contrasting effects are of substantial importance when considering these drugs as therapy for neuromyelitis optica spectrum disorders (NMOSD), a frequent differential diagnosis to MS, which is considered being a primarily B cell- and antibody-driven diseases. Data indicates that MS medications, which deplete B cells or induce an anti-inflammatory phenotype of the remaining ones, were effective and safe in aquaporin-4 antibody positive NMOSD. In contrast, drugs such as natalizumab and interferon-β, which lead to activation and accumulation of B cells in the peripheral blood, lack efficacy or even induce catastrophic disease activity in NMOSD. Hence, we conclude that the differential effect of MS drugs on B cells is one potential parameter determining the therapeutic efficacy or failure in antibody-dependent diseases like seropositive NMOSD.

Role of Glutamatergic Excitotoxicity in Neuromyelitis Optica Spectrum Disorders

Neuromyelitis optica spectrum disorder (NMOSD) is an inflammatory disorder mediated by immune-humoral responses directed against central nervous system (CNS) antigens. Most patients are positive for specific immunoglobulin G (IgG) auto-antibodies for aquaporin-4 (AQP4), a water channel present in astrocytes. Antigen-antibody binding promotes complement system cascade activation, immune system cell infiltration, IgG deposition, loss of AQP4 and excitatory amino acid transporter 2 (EAAT2) expression on the astrocytic plasma membrane, triggering necrotic destruction of spinal cord tissue and optic nerves. Astrocytes are very important cells in the CNS and, in addition to supporting other nerve cells, they also regulate cerebral homeostasis and control glutamatergic synapses by modulating neurotransmission in the cleft through the high-affinity glutamate transporters present in their cell membrane. Specific IgG binding to AQP4 in astrocytes blocks protein functions and reduces EAAT2 activity. Once compromised, EAAT2 cannot take up free glutamate from the extracellular space, triggering excitotoxicity in the cells, which is characterized by overactivation of glutamate receptors in postsynaptic neurons. Therefore, the longitudinally extensive myelitis and optic neuritis lesions observed in patients with NMOSD may be the result of primary astrocytic damage triggered by IgG binding to AQP4, which can activate the immune-system cascade and, in addition, downregulate EAAT2. All these processes may explain the destructive lesions in NMOSD secondary to neuroinflammation and glutamatergic excitotoxicity. New or repurposed existing drugs capable of controlling glutamatergic excitotoxicity may provide new therapeutic options to reduce tissue damage and permanent disability after NMOSD attacks.