Contrast-Enhancing Lesion Segmentation in Multiple Sclerosis: A Deep Learning Approach Validated in a Multicentric Cohort.

The detection of contrast-enhancing lesions (CELs) is fundamental for the diagnosis and monitoring of patients with multiple sclerosis (MS). This task is time-consuming and suffers from high intra- and inter-rater variability in clinical practice. However, only a few studies proposed automatic approaches for CEL detection. This study aimed to develop a deep learning model that automatically detects and segments CELs in clinical Magnetic Resonance Imaging (MRI) scans. A 3D UNet-based network was trained with clinical MRI from the Swiss Multiple Sclerosis Cohort. The dataset comprised 372 scans from 280 MS patients: 162 showed at least one CEL, while 118 showed no CELs. The input dataset consisted of T1-weighted before and after gadolinium injection, and FLuid Attenuated Inversion Recovery images. The sampling strategy was based on a white matter lesion mask to confirm the existence of real contrast-enhancing lesions. To overcome the dataset imbalance, a weighted loss function was implemented. The Dice Score Coefficient and True Positive and False Positive Rates were 0.76, 0.93, and 0.02, respectively. Based on these results, the model developed in this study might well be considered for clinical decision support.

Immunoglobulin A Antibodies Against Myelin Oligodendrocyte Glycoprotein in a Subgroup of Patients With Central Nervous System Demyelination.

Importance: Differential diagnosis of patients with seronegative demyelinating central nervous system (CNS) disease is challenging. In this regard, evidence suggests that immunoglobulin (Ig) A plays a role in the pathogenesis of different autoimmune diseases. Yet little is known about the presence and clinical relevance of IgA antibodies against myelin oligodendrocyte glycoprotein (MOG) in CNS demyelination. Objective: To investigate the frequency of MOG-IgA and associated clinical features in patients with demyelinating CNS disease and healthy controls. Design, Setting, and Participants: This longitudinal study comprised 1 discovery and 1 confirmation cohort derived from 5 centers. Participants included patients with suspected or confirmed demyelinating diseases and healthy controls. MOG-IgA, MOG-IgG, and MOG-IgM were measured in serum samples and cerebrospinal fluid (CSF) of patients, who were assessed from September 2012 to April 2022. Main Outcomes and Measures: Frequency and clinical features of patients who were seropositive for MOG-IgA and double-seronegative for aquaporin 4 (AQP4) IgG and MOG-IgG. Results: After the exclusion of 5 participants with coexisting AQP4-IgG and MOG-IgA, MOG-IgG, and/or MOG-IgM, 1339 patients and 110 healthy controls were included; the median follow-up time was 39 months (range, 0-227 months). Of included patients with isolated MOG-IgA, 11 of 18 were female (61%), and the median age was 31.5 years (range, 3-76 years). Among patients double-seronegative for AQP4-IgG and MOG-IgG (1126/1339; 84%), isolated MOG-IgA was identified in 3 of 50 patients (6%) with neuromyelitis optica spectrum disorder, 5 of 228 patients (2%) with other CNS demyelinating diseases, and 10 of 848 patients (1%) with multiple sclerosis but in none of the healthy controls (0/110). The most common disease manifestation in patients seropositive for isolated MOG-IgA was myelitis (11/17 [65%]), followed by more frequent brainstem syndrome (7/16 [44%] vs 14/75 [19%], respectively; P = .048), and infrequent manifestation of optic neuritis (4/15 [27%] vs 46/73 [63%], respectively; P = .02) vs patients with MOG-IgG. Among patients fulfilling 2017 McDonald criteria for multiple sclerosis, MOG-IgA was associated with less frequent CSF-specific oligoclonal bands (4/9 [44%] vs 325/351 [93%], respectively; P < .001) vs patients with multiple sclerosis who were MOG-IgG/IgA seronegative. Further, most patients with isolated MOG-IgA presented clinical attacks after recent infection or vaccination (7/11 [64%]). Conclusion and Relevance: In this study, MOG-specific IgA was identified in a subgroup of patients who were double-seronegative for AQP4-/MOG-IgG, suggesting that MOG-IgA may be a novel diagnostic biomarker for patients with CNS demyelination.

Association of Brain Atrophy With Disease Progression Independent of Relapse Activity in Patients With Relapsing Multiple Sclerosis.

Importance: The mechanisms driving neurodegeneration and brain atrophy in relapsing multiple sclerosis (RMS) are not completely understood. Objective: To determine whether disability progression independent of relapse activity (PIRA) in patients with RMS is associated with accelerated brain tissue loss. Design, Setting, and Participants: In this observational, longitudinal cohort study with median (IQR) follow-up of 3.2 years (2.0-4.9), data were acquired from January 2012 to September 2019 in a consortium of tertiary university and nonuniversity referral hospitals. Patients were included if they had regular clinical follow-up and at least 2 brain magnetic resonance imaging (MRI) scans suitable for volumetric analysis. Data were analyzed between January 2020 and March 2021. Exposures: According to the clinical evolution during the entire observation, patients were classified as those presenting (1) relapse activity only, (2) PIRA episodes only, (3) mixed activity, or (4) clinical stability. Main Outcomes and Measures: Mean difference in annual percentage change (MD-APC) in brain volume/cortical thickness between groups, calculated after propensity score matching. Brain atrophy rates, and their association with the variables of interest, were explored with linear mixed-effect models. Results: Included were 1904 brain MRI scans from 516 patients with RMS (67.4% female; mean [SD] age, 41.4 [11.1] years; median [IQR] Expanded Disability Status Scale score, 2.0 [1.5-3.0]). Scans with insufficient quality were excluded (n = 19). Radiological inflammatory activity was associated with increased atrophy rates in several brain compartments, while an increased annualized relapse rate was linked to accelerated deep gray matter (GM) volume loss. When compared with clinically stable patients, patients with PIRA had an increased rate of brain volume loss (MD-APC, -0.36; 95% CI, -0.60 to -0.12; P = .02), mainly driven by GM loss in the cerebral cortex. Patients who were relapsing presented increased whole brain atrophy (MD-APC, -0.18; 95% CI, -0.34 to -0.02; P = .04) with respect to clinically stable patients, with accelerated GM loss in both cerebral cortex and deep GM. No differences in brain atrophy rates were measured between patients with PIRA and those presenting relapse activity. Conclusions and Relevance: Our study shows that patients with RMS and PIRA exhibit accelerated brain atrophy, especially in the cerebral cortex. These results point to the need to recognize the insidious manifestations of PIRA in clinical practice and to further evaluate treatment strategies for patients with PIRA in clinical trials.