Introduction: The Challenge of Prognostication in Anti-NMDAR Encephalitis
Anti-N-methyl-D-aspartate receptor (anti-NMDAR) encephalitis represents one of the most common causes of autoimmune encephalitis, characterized by a complex constellation of psychiatric symptoms, cognitive deficits, seizures, and autonomic instability. While the disease is notoriously severe—often requiring prolonged intensive care—it is also uniquely treatable. However, the clinical trajectory of patients varies significantly. Some achieve full recovery with first-line immunotherapy, while others suffer from persistent cognitive impairment or require years of rehabilitation.
For clinicians, the ability to predict these outcomes early in the disease course is critical. Early identification of patients unlikely to respond to standard first-line therapies (corticosteroids, intravenous immunoglobulin, or plasma exchange) could facilitate the timely initiation of second-line agents like rituximab or cyclophosphamide. The newly developed NEOS2 score, validated in a large international cohort, aims to provide this predictive clarity using data available at the time of diagnosis.
Highlights of the NEOS2 Validation Study
The study introduces several pivotal findings for the management of anti-NMDAR encephalitis:
- The NEOS2 score demonstrates high accuracy (AUC 80-86%) in predicting functional recovery at one year and the resumption of school or work within three years.
- A modified version, the NEOS2-T, specifically predicts whether a patient will show clinical improvement within just two weeks of starting first-line immunotherapy.
- The score relies on five easily accessible clinical variables: age, treatment delay, presence of movement disorders, ICU admission, and CSF leukocyte count.
- By using the score as an ordinal measure, clinicians can move beyond binary outcomes to nuanced probability estimates for individual patients.
Study Design and Global Collaboration
This study represents a significant advancement in neuroimmunological research due to its scale and geographic diversity. Researchers analyzed data from 702 patients diagnosed with definite anti-NMDAR encephalitis between 2007 and 2022. The cohort spanned multiple specialized centers across France, Germany, Japan, the Netherlands, and Spain.
The primary objective was to refine the original Anti-NMDAR Encephalitis One-year functional Status (NEOS) score. The original score was a pioneering effort, but the NEOS2 iteration incorporates more granular data and extends the predictive window to include long-term social and vocational outcomes. The researchers used logistic regression to build multivariable models, ensuring that the predictors chosen were both statistically robust and clinically practical for bedside use.
The Five Pillars of the NEOS2 Score
The NEOS2 score is calculated based on five independent predictors identified through multivariable analysis. Each factor contributes to the likelihood of a poorer outcome:
1. Age at Diagnosis
The study found that higher age (OR 0.35) was a significant predictor of poorer functional outcomes. While anti-NMDAR encephalitis primarily affects young adults and children, older patients often face a more protracted recovery and higher mortality, potentially due to comorbidities or differences in immune system resilience.
2. Treatment Delay
Time is brain in autoimmune encephalitis. A longer interval between symptom onset and the initiation of immunotherapy (OR 0.49) was strongly associated with worse outcomes. This reinforces the clinical mandate for rapid diagnosis and early intervention.
3. Movement Disorders
The presence of prominent movement disorders—such as orofacial dyskinesias, choreoathetosis, or dystonia—at the time of diagnosis (OR 0.32) serves as a marker of severe, deep-brain involvement, correlating with more difficult recoveries.
4. ICU Requirement
Patients requiring intensive care for mechanical ventilation or autonomic instability (OR 0.34) naturally represent a more severe disease phenotype, which the NEOS2 score captures as a key prognostic weight.
5. Increased CSF Leucocyte Count
A higher leukocyte count in the cerebrospinal fluid (OR 0.65) indicates a more robust inflammatory response within the central nervous system, which may correlate with higher antibody titers and more extensive neuronal receptor internalization.
Key Findings: From Short-Term Response to Long-Term Integration
The study evaluated outcomes across three distinct timeframes, providing a comprehensive view of the patient journey.
Improvement After First-Line Immunotherapy (NEOS2-T)
In the cohort, 96% of patients received first-line immunotherapy, but only 38% showed significant clinical improvement (defined by a change in the modified Rankin Scale) within the first two weeks. The NEOS2-T model (which excludes age but includes the other four variables) predicted this early response with an AUC of 81-84%. This is perhaps the most clinically actionable aspect of the study, as it helps physicians identify “slow responders” who might benefit from immediate escalation to second-line therapy.
Functional Outcome at One Year
One year post-diagnosis, 80% of patients achieved a favorable functional outcome (mRS ≤ 2). The NEOS2 score effectively stratified these patients, showing that those with the lowest scores had a nearly 100% chance of a good outcome, whereas those at the highest end of the spectrum faced a high risk of persistent disability.
Resumption of Work and School (NEOS2-W)
A unique strength of this study was its focus on “real-world” recovery. At the three-year mark, 73% of patients had resumed their previous school or work activities. The NEOS2-W model predicted this outcome with an accuracy of 80%, highlighting that early clinical markers have long-lasting predictive value for a patient’s quality of life and social reintegration.
Expert Commentary and Clinical Implications
The development of the NEOS2 score marks a transition toward precision medicine in neuroimmunology. Historically, treatment for anti-NMDAR encephalitis has followed a stepped approach: start with first-line therapy, wait several weeks, and then escalate if no improvement is seen. However, the high accuracy of the NEOS2 score suggests that for patients with high-risk profiles (e.g., older age, severe ICU involvement, and significant treatment delay), the “wait and see” approach may result in lost opportunities for recovery.
One notable aspect of the NEOS2 score is its simplicity. Unlike some predictive models that require complex genetic testing or specialized neuroimaging software, the components of NEOS2 are part of the standard diagnostic workup for any patient suspected of having encephalitis. This makes the score highly generalizable, even in centers with limited resources.
However, the researchers and independent experts caution that while NEOS2 is a powerful tool, it should not replace clinical judgment. The study was retrospective in nature, and though the cohort was large and international, prospective validation is the necessary next step. Additionally, the score does not currently account for the presence of an underlying tumor (such as an ovarian teratoma), which is a known factor in both the etiology and prognosis of the disease.
Summary and Future Directions
The NEOS2 score provides a validated, evidence-based framework for predicting the trajectory of anti-NMDAR encephalitis. By integrating age, treatment timing, clinical severity, and inflammatory markers, it offers a window into the patient’s future—from their immediate response to steroids to their ability to return to the classroom or office years later. For the clinical community, the NEOS2 score is more than just a statistical exercise; it is a stratification tool that could potentially guide personalized treatment protocols and improve long-term outcomes for patients facing this challenging neurological condition.
Funding and References
This study was funded by Dioraphte (charity; project 2001 0403). The researchers involved represent the GENERATE Study Group and several leading academic medical centers worldwide.
Reference: Brenner J, Bastiaansen AEM, Guasp M, et al. Development and validation of the NEOS2 score for prediction of long-term outcomes and improvement after first-line immunotherapy in patients with anti-NMDAR encephalitis: an international cohort study. Lancet Reg Health Eur. 2025;62:101562. doi:10.1016/j.lanepe.2025.101562.
