CSF/serum IL-6 ratio and its correlation with mri severity in mog antibody-associated cerebral cortical encephalitis: an exploratory pilot study

Objective We aimed to quantify the relationship between CSF IL-6 compartmentalization and MRI lesion extent in patients with MOG-CCE, and to evaluate the CSF/serum IL-6 ratio as a biomarker of CNS-compartmentalized inflammation. Methods Eight MOG-CCE patients (≥14 years old) were retrospectively enrolled. CSF and serum cytokines (IL-6, IL-17, IFN-γ) and MRI scans were collected. A semi-quantitative MRI scoring system (0–9) was developed; inter-rater reliability was assessed by two independent radiologists. Associations were evaluated using Spearman correlation and simple linear regression. Results All patients presented with seizures at onset. Median CSF IL-6 (284.53 pg/mL) markedly exceeded paired serum levels (2.25 pg/mL; Z = 2.521, p = 0.012; Cohen’s d = 1.084). The median CSF/serum IL-6 ratio was 220.33 (IQR: 64.26–690.89), exceeding 1 in all patients. The MRI scoring system showed good inter-rater reliability (87.5% concordance; P = 0.893). The total MRI score correlated significantly with CSF IL-6 (r = 0.819, p = 0.013) and the CSF/serum IL-6 ratio (r = 0.807, p = 0.015). Simple linear regression confirmed that the Ln-transformed ratio predicted MRI score (R² = 0.573, p = 0.030, B = 0.659, β = 0.757). Crucially, the ratio showed no significant correlation with conventional systemic inflammatory markers. Conclusion CSF IL-6 levels correlated significantly with MRI severity in MOG-CCE. The Ln-transformed CSF/serum IL-6 ratio explained approximately 57.3% of MRI variance and was uncorrelated with systemic inflammatory indices, supporting its specificity for CNS-compartmentalized immune activation. These findings position the CSF/serum IL-6 ratio as a candidate quantitative biomarker of local CNS inflammatory intensity and lesion burden in MOG-CCE. As an exploratory, hypothesis-generating pilot study (n = 8), these results require prospective validation in larger cohorts with Reiber graph analysis before the ratio can be adopted as a clinical biomarker.