Explore the vast range of titles available.

Artificial intelligence (AI) has meant a turning point in data analysis, allowing predictions of unseen outcomes with precedented levels of accuracy. In multiple sclerosis (MS), a chronic inflammatory-demyelinating condition of the central nervous system with a complex pathogenesis and potentially devastating consequences, AI-based models have shown promising preliminary results, especially when using neuroimaging data as model input or predictor variables. The application of AI-based methodologies to serum/blood and CSF biomarkers has been less explored, according to the literature, despite its great potential. In this review, we aimed to investigate and summarise the recent advances in AI methods applied to body fluid biomarkers in MS, highlighting the key features of the most representative studies, while illustrating their limitations and future directions.

BackgroundMultiple sclerosis (MS), neuromyelitis optica spectrum disorder (NMOSD) and myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD) can share similar features, posing diagnostic challenges. In this study, we identified sets of conventional MRI lesion distribution criteria proposed for disease differentiation and investigated their clinical utility.MethodsWe searched five electronic databases for English-written and peer-reviewed diagnostic accuracy studies that included brain MRI at least. Hierarchical and univariate random-effects logistic regression models were employed for diagnostic accuracy meta-analysis. Heterogeneity was explored with subgroup analyses. Certainty of evidence was assessed using the GRADEpro tool.ResultsThree sets of criteria (‘Matthews’, ‘Cacciaguerra’, ‘MS lesion checklist’) were investigated in 11 studies (2008 patients; MS, n=1037; NMOSD, n=842; MOGAD, n=129), with low applicability concerns. Overall pooled sensitivity and specificity of the Matthews brain MRI criteria (MS vs seropositive-NMOSD differentiation) were 0.92 (0.86 to 0.96) and 0.85 (0.79 to 0.90), respectively, with higher diagnostic values in non-Caucasian populations and during follow-up. Pooled sensitivity and specificity of the Cacciaguerra brain-spinal cord criteria (seropositive-NMOSD vs MS differentiation) were 0.96 (0.76 to 0.99) and 0.83 (0.71 to 0.90), respectively. The MS lesion checklist (MS vs NMOSD/MOGAD differentiation) had lower diagnostic accuracy measures (sensitivity, specificity: 0.74, 0.79, respectively). The Matthews criteria provided the strongest moderate certainty evidence and also showed high pooled diagnostic accuracy for MS versus seronegative-NMOSD (sensitivity: 0.93 (0.84 to 0.97)); specificity: 0.90 (0.80 to 0.95)) and for MS versus MOGAD differentiation (sensitivity: 0.86 (0.81 to 0.90); specificity: 0.87 (0.76 to 0.93)).ConclusionsLesion distribution criteria can accurately discriminate between MS, NMOSD and MOGAD. Further optimised validation studies, and revisions or extensions may support sustained implementation.PROSPERO registration numberCRD42023472178.

Myelin oligodendrocyte glycoprotein (MOG) antibody-associated disease (MOGAD) is an immune-mediated demyelinating disease that is challenging to differentiate from multiple sclerosis (MS), as the clinical phenotypes overlap, and people with MOGAD can fulfil the current MRI-based diagnostic criteria for MS. In addition, the MOG antibody assays that are an essential component of MOGAD diagnosis are not standardized. Accurate diagnosis of MOGAD is crucial because the treatments and long-term prognosis differ from those for MS. This Expert Recommendation summarizes the outcomes from a Magnetic Resonance Imaging in MS workshop held in Oxford, UK in May 2022, in which MS and MOGAD experts reflected on the pathology and clinical features of these disorders, the contributions of MRI to their diagnosis and the clinical use of the MOG antibody assay. We also critically reviewed the literature to assess the validity of distinctive imaging features in the current MS and MOGAD criteria. We conclude that dedicated orbital and spinal cord imaging (with axial slices) can inform MOGAD diagnosis and also illuminate differential diagnoses. We provide practical guidance to neurologists and neuroradiologists on how to navigate the current MOGAD and MS criteria. We suggest a strategy that includes useful imaging discriminators on standard clinical MRI and discuss imaging features detected by non-conventional MRI sequences that demonstrate promise in differentiating these two disorders.Myelin oligodendrocyte glycoprotein antibody-associated disease is an immune-mediated demyelinating disease that is distinct from multiple sclerosis but shares some of its characteristics. This Expert Recommendation, based on a Magnetic Resonance Imaging in MS workshop, proposes a diagnostic algorithm for the differential diagnosis of myelin oligodendrocyte glycoprotein antibody-associated disease and multiple sclerosis, using serological, imaging and clinical features.

Background and objectivesDifferent treatment response scoring systems in treated MS patients exist. The objective was to assess the long-term predictive value of these systems in RRMS patients treated with self-injectable DMTs. MethodsRRMS-treated patients underwent brain MRI before the onset of therapy and 12 months thereafter, and neurological assessments every 6 months. Clinical and demographic characteristics were collected at baseline. After the first year of treatment, several scoring systems [Rio score (RS), modified Rio score (MRS), MAGNIMS score (MS), and ROAD score (RoS)] were calculated. Cox-Regression and survival analyses were performed to identify scores predicting long-term disability. ResultsWe included 319 RRMS patients. Survival analyses showed that patients with RS > 1 and RoS > 3 had a significant risk of reaching an EDSS of 4.0 and 6.0 The score with the best sensitivity (61%) was the RoS, while the MRS showed the best specificity (88%). The RS showed the best positive predictive value (42%) and the best accuracy (81%). ConclusionsThe combined measures integrated into different scores have an acceptable prognostic value for identifying patients with long-term disability.Thus, these data reinforce the concept of early treatment optimization to minimize the risk of long-term disability.

IntroductionTo evaluate the impact of the COVID-19 pandemic on (1) number of clinical visits, (2) magnetic resonance (MR) scans, and (3) treatment prescriptions in a multiple sclerosis (MS) referral centre. MethodsRetrospective study covering January 2018 to May 2021.ResultsThe monthly mean (standard deviation [SD]) of visits performed in 2020 (814[137.6]) was similar to 2018 (741[99.7]; p = 0.153), and 2019 (797[116.3]; p = 0.747). During the COVID-19 period (2020 year), 36.3% of the activity was performed through telemedicine. The number of MR scans performed dropped by 76.6% during the “first wave” (March 14 to June 21, 2020) compared to the mean monthly activity in 2020 (183.5[68.9]), with a recovery during the subsequent two months. The monthly mean of treatment prescriptions approved in 2020 (24.1[7.0]) was lower than in 2019 (30[7.0]; p = 0.049), but similar to 2018 (23.8[8.0]; p = 0.727). Natalizumab prescriptions increased in the “first wave” and onwards, whereas anti-CD20 prescriptions decreased during the COVID-19 period.ConclusionMaintenance of the number of clinical visits was likely due to telemedicine adoption. Although the number of MR dramatically dropped during the “first wave”, an early recovery was observed. Treatment prescriptions suffered a slight quantitative decrease during 2020, whereas substantial qualitative changes were found in specific treatments.

Purpose Brain volume changes (BVC) on therapy in MS are being considered as predictor for treatment response at an individual level. We ought to assess whether adding BVC as a factor to monitor interferon-beta response improves the predictive ability of the (no) evidence of disease activity (EDA-3) and Río score (RS-3) criteria for confirmed disability progression in a historical cohort. Methods One hundred one patients from an observational cohort treated with interferon-beta were assessed for different cutoff points of BVC (ranged 0.2–1.2%), presence of active lesions (≥ 1 for EDA/≥ 3 for RS), relapses, and 6-month confirmed disability progression (CDP), measured by the Expanded Disability Status Scale, after 1 year. Sensitivity, specificity, and positive and negative predictive values for predicting confirmed disability progression at 4 years in original EDA (EDA-3) and RS (RS-3) as well as EDA and RS including BVC (EDA-4 and RS-4) were compared. Results Adding BVC to EDA slightly increased sensitivity, but not specificity or predictive values, nor the OR for predicting CDP; only EDA-3 showed a trend for predicting CDP (OR 3.701, p = 0.050). Adding BVC to RS-3 (defined as ≥ 2 criteria) helped to improve sensitivity and negative predictive value, and increased OR for predicting CDP using a cutoff of ≤ − 0.86% (RS-3 OR 23.528, p < 0.001; RS-4 for all cutoffs ranged from 15.06 to 32, p < 0.001). RS-4 showed areas under the curve larger than RS-3 for prediction of disability at 4 years. Conclusion Addition of BVC to RS improves its prediction of response to interferon-beta.

We aimed to investigate the ability of natalizumab (NTZ)-treated patients to assume treatment-associated risks and the factors involved in such risk acceptance. From a total of 185 patients, 114 patients on NTZ as of July 2011 carried out a comprehensive survey. We obtained disease severity perception scores, personality traits' scores, and risk-acceptance scores (RAS) so that higher RAS indicated higher risk acceptance. We recorded JC virus status (JCV+/-), prior immunosuppression, NTZ treatment duration, and clinical characteristics. NTZ patients were split into subgroups (A-E), depending on their individual PML risk. Some 22 MS patients on first-line drugs (DMD) acted as controls. No differences between treatment groups were observed in disease severity perception and personality traits. RAS were higher in NTZ than in DMD patients (p<0.01). Perception of the own disease as a more severe condition tended to predict higher RAS (p=0.07). Higher neuroticism scores predicted higher RAS in the NTZ group as a whole (p=0.04), and in high PML-risk subgroups (A-B) (p=0.02). In low PML-risk subgroups (C-E), higher RAS were associated with a JCV+ status (p=0.01). Neither disability scores nor pre-treatment relapse rate predicted RAS in either group. Risk acceptance is a multifactorial phenomenon, which might be partly explained by an adaptive process, in light of the higher risk acceptance amongst NTZ-treated patients and, especially, amongst those who are JCV seropositive but still have low PML risk, but which seems also intimately related to personality traits.