A nomogram for predicting viral encephalitis based on cerebrospinal fluid biomarkers

Area of Science:

• Neurology
• Infectious Diseases
• Biostatistics

Background:

• Viral encephalitis (VE) is a significant CNS infection posing a considerable disease burden.
• Early identification and intervention are critical for managing VE and improving patient outcomes.

Purpose of the Study:

• To develop and validate a predictive model for the early detection of viral encephalitis (VE) patients.
• To identify key clinical factors associated with VE risk and protection.

Main Methods:

• Retrospective analysis of clinical data from 160 VE and 131 non-VE patients.
• Development and validation of a predictive model using logistic regression, ROC curves, and calibration plots.
• Construction of a nomogram for visualizing predictive results and assessing clinical utility via decision curve analysis (DCA).

Main Results:

• Six covariates (ALB, β2-MG, IFN-γ, CSF WBC, CSF LYM, CSF P) identified as independent VE risk factors.
• IFN-α identified as a protective factor against VE.
• The nomogram demonstrated high discrimination (AUC 0.86 in training, 0.85 in validation) and good calibration, with significant clinical utility.

Conclusions:

• The developed predictive model reliably identifies VE patients.
• The model provides a scientific basis for clinical decision-making in VE diagnosis.
• Implementation of this model can lead to improved patient outcomes through timely interventions.