Detection of neuronal antibodies for autoimmune encephalitis and paraneoplastic neurological syndromes relies on commercially available cell-based assays and lineblots. However, lineblots may reveal the presence of neuronal antibodies in patients with various non-autoimmune etiologies. Herein we describe patients with non-autoimmune etiologies (cohort B) and detectable neuronal antibodies and compare them to definite cases of autoimmune encephalitis (cohort A) for differences in clinical data. All patients positive for at least one neuronal antibody were retrospectively evaluated for autoimmune encephalitis and/or paraneoplastic neurological syndrome between 2016 and 2022. 39 cases in cohort B and 23 in cohort A were identified. In cohort B, most common diagnoses were neurodegenerative disorders in 9/39 (23.1%), brain tumors in 6/39 (15.4%) while most common detected antibodies were anti–titin (N10), anti-recoverin (N11), anti-Yo (N8) and all were detected in serum only. Differential aspects between cohort A and B were CSF pleocytosis (14/23 (60.8%) vs 11/35 (31.4%), p = 0.042, respectively), MRI features suggestive of encephalitis (6/23 (26.1%) vs 0 (0%), p = 0.002, respectively) and epilepsy restricted to temporal lobes (14/23 (60.9%) vs 2/30 (6.7%), p = 0.0003, respectively). A large proportion of lineblot results were non-specific when only serum was tested and were frequently found in non-autoimmune neurological conditions.

Background: Encephalitis is a severe and potentially life-threatening inflammatory disorder of the central nervous system. Without prompt diagnosis and appropriate treatment, it often results in poor clinical outcomes. The study aimed to develop an artificial intelligence-based model that distinguishes autoimmune encephalitis from infectious encephalitis, encompassing a broad spectrum of autoimmune encephalitis phenotypes, serostatuses, and neuroimmunological entities. Methods: We conducted a retrospective analysis of patients diagnosed with autoimmune encephalitis, including paraneoplastic neurological syndromes and/or infectious encephalitis, at Vilnius University Hospital Santaros Klinikos from 2016 to 2024. Supervised machine learning techniques were used to train the models, and Shapley Additive Explanations analysis was applied to improve their interpretability. Results: A total of 233 patients were included in the study. The Random Forest model demonstrated the best performance in differentiating the etiology of encephalitis, achieving an AUROC of 0.966. Further analysis revealed that laboratory, electroencephalography, and clinical data were the most influential predictors, whereas imaging data contributed less to classification accuracy. Conclusions: We developed a machine learning model capable of distinguishing infectious encephalitis from both seropositive and seronegative autoimmune encephalitis. Since autoimmune cases may be misdiagnosed as infectious in the absence of detectable antibodies, our model has the potential to support clinical decision-making and reduce diagnostic uncertainty.

Tumefactive multiple sclerosis (TmMS) is a rare subtype of a demyelinating disease that develops over time. Cases of hyperacute presentations mimicking cerebrovascular disorders have been reported; however, detailed clinical and demographic data are lacking. this study aimed to systematically review the literature on tumefactive demyelinating disorders presenting as strokes. After screening the PubMed, PubMed Central, and Web of Science databases, 39 articles describing 41 patients were identified, including 2 historical patients from our center. 23 (53.4%) patients were diagnosed with multiple sclerosis variants (vMS), 17 (39.5%) with inflammatory demyelinating variants (vInf), and 3 with tumors; however, only 43.5% of cases were verified histologically. In subgroup analysis, vMS differed from vInf in several aspects. Inflammatory cerebral spinal fluid parameters, including pleocytosis, proteinorachia was more commonly observed in vInf [11 (64.7%) vs. 1 (5.2%), P = 0.001 and 13/17 (76.4%) vs. 6/23 (31.5%), P = 0.02] than that in vMS. Neurological deterioration and fatal outcomes were more commonly observed in vInf [13/17 (76.4%) vs. 7/23 (30.4%), P = 0.003, and 11/17 (64.7%) vs. 0/23 (0%), P = 0.0001] than that in vMS. Clinicodemographic data might aid in recognizing different subtypes of TmMS and warrant consideration of unconventional therapies because outcomes may be poor in the vInf of TmMS.

Background: Chimeric antigen receptor T (CAR-T) cell therapy is an effective treatment for treatment-refractory hematological disorders with an acceptable safety profile. In contrast, preliminary reports suggest good efficacy for treatment-refractory autoimmune disorders, including autoimmune nervous system disease, but their safety profile is largely unknown. Objective: To describe the first case of glutamic acid decarboxylase-65 (GAD65) antibody-mediated cerebellar ataxia (CA) successfully treated with CD19 CAR-T cells. Results: A 33-year-old male was diagnosed with GAD65 antibody mediated CA in 2023. Despite treatment with Rituximab and Cyclophosphamide, the patient’s condition worsened with new-onset recurrent falls and increasing vertigo. Ambulation was maintained. CD19 CAR-T cells at a dose of 1 × 10 ⁶ cells per kilogram of body weight were infused after administration of standard lymphodepleting chemotherapy, resulting in a good serological response with reduction of GAD65 serum titers by 95% at day +90, significant clinical improvement in ataxia at day +30 and no evidence of disease progression at day +270 clinically, radiologically and laboratory-wise. The toxicity was limited to cytokine release syndrome grade 1. Discussion: The favorable clinical response observed in our patient, along with other reports demonstrating preliminary efficacy and limited toxicity, supports further study of CD19 CAR-T cell therapy in GAD65 neurological disorders.

Autoimmune encephalitis is an increasingly recognized cause of encephalitis. Detection of disease-specific antibodies is the cornerstone of diagnosis. Despite growing clinical awareness and the routine availability of antibody assays in most centers, diagnosis remains challenging due to diverse clinical presentations, the low diagnostic yield of commercial antibody kits, difficulties in interpreting antibody results, and a substantial proportion of paraclinically silent patients. This narrative review summarizes current diagnostic approaches to autoimmune encephalitis, with particular emphasis on antibody detection strategies, the diagnostic yield of different techniques, serum vs. cerebrospinal fluid testing, and the diagnostic value of supportive cerebrospinal fluid biomarkers. In addition, we discuss patients with seronegative or paraclinically silent disease, in whom diagnosis relies primarily on clinical criteria and the exclusion of alternative etiologies. Finally, we outline future perspectives, including advanced immunological techniques and machine learning-based diagnostic models, which may facilitate earlier identification and more accurate classification of autoimmune encephalitis. Improved integration of clinical assessment with cerebrospinal fluid biomarkers and novel analytical tools may reduce diagnostic delay and support the timely initiation of immunotherapy, ultimately improving neurological outcomes.