Related Experiment Video
Updated: Jun 14, 2026
Bioluminescence and Near-infrared Imaging of Optic Neuritis and Brain Inflammation in the EAE Model of Multiple Sclerosis in Mice
Published on: March 1, 2017
Integrated Blood Inflammatory Ratios and Cerebrospinal Fluid Blood‒Brain Barrier Dysfunction Predict Relapse Risk in
Xingyue Zheng1,2, Jing Shi2, Hao Yin2
1Department of Neurology, The First Affiliated Hospital of Dalian Medical University, Dalian Medical University, Dalian, China.
Background:
Relapse is the primary driver of irreversible disability accumulation in neuromyelitis optica spectrum disorder (NMOSD). Although aquaporin-4 immunoglobulin G (AQP4-IgG) is central to disease diagnosis and pathogenesis, reliable tools for individualized relapse risk stratification remain limited. Given emerging evidence that blood-brain barrier (BBB) dysfunction and systemic immune activation play key and interconnected roles in NMOSD pathophysiology, we aimed to develop and internally validate an integrated prognostic model incorporating these dimensions to predict relapse risk.
Methods:
In this retrospective cohort study, 152 patients with NMOSD were enrolled and followed longitudinally for time to first relapse. Baseline peripheral inflammatory indices, including the neutrophil-to-lymphocyte ratio (NLR) and monocyte-to-lymphocyte ratio (MLR), as well as cerebrospinal fluid (CSF) parameters reflecting BBB integrity, were collected at study entry. Independent predictors of time to first relapse were identified using multivariable Cox proportional hazards regression. Three hierarchical prognostic models were constructed and compared: a clinical model, a clinical-CSF model, and an integrated clinical-CSF-blood inflammatory model. Model discrimination was evaluated using time-dependent receiver operating characteristic (ROC) analysis. Internal validation was performed using 1000 bootstrap resamples, and clinical utility was assessed by decision curve analysis (DCA).
Results:
Patients who experienced relapse exhibited more pronounced BBB dysfunction and higher systemic inflammatory activation at baseline. Model discrimination improved with the sequential incorporation of CSF indices and inflammatory markers, with AUCs of 0.693 for the clinical model, 0.744 for the clinical-CSF model, and 0.850 for the integrated model. The final nomogram demonstrated good discrimination (bootstrap-corrected C-index 0.811), good calibration, and favorable clinical utility. AQP4-IgG seropositive patients had a higher relapse risk; importantly, the integrated model retained predictive performance in this subgroup, indicating added prognostic value beyond serological status. Exploratory analyses further suggested that higher AQP4-IgG titers were associated with increased relapse risk.
Conclusions:
An integrated nomogram incorporating peripheral inflammatory ratios and BBB-related CSF indices enables individualized relapse risk prediction in NMOSD. The model provides incremental prognostic value beyond AQP4-IgG serostatus and may support risk-adapted clinical management.
