# From Diagnosis to Disease Staging: Multisite Validation of Cerebrospinal Fluid Molecular Tests in Multiple Sclerosis

## TL;DR

This study shows that analyzing cerebrospinal fluid can accurately diagnose multiple sclerosis, distinguish its subtypes, and predict future disability, offering a more precise alternative to traditional methods.

## Contribution

The study externally validates CSF-based molecular tests for MS diagnosis, subtype differentiation, and prognosis in a real-world, multicenter setting.

## Key findings

- CSF molecular tests distinguished MS from other neurological diseases with an AUC of 0.94.
- The test differentiated relapsing-remitting from progressive MS with an AUC of 0.76.
- A prognostic classifier predicted future disability progression with statistical significance.

## Abstract

The growing demand for personalized treatment in multiple sclerosis (MS) highlights the need for more precise biomarkers that can outperform magnetic resonance imaging and clinical assessment in patient stratification. Advances in multiplex proteomic technologies suggest that cerebrospinal fluid (CSF) analysis at MS onset may not only improve diagnostic accuracy, but also offer prognostic and staging information, as well as insight into molecular therapeutic targets.

This multicenter study retrospectively analyzed cryopreserved CSF samples from 160 individuals undergoing diagnostic evaluation for possible neuroimmunological disorder, and among these, followed a cohort of 96 people with confirmed MS for at least 3 years. The goal was to externally validate previously published CSF‐based diagnostic and prognostic classifiers.

Upon unblinding, the CSF‐based molecular diagnostic test distinguished 96 people with confirmed MS from 30 individuals with other inflammatory neurological diseases, and 34 individuals with non‐inflammatory neurological diseases, achieving an area under the receiver operating characteristic curve of 0.94 (p = 4.7 × 10−21). The test also differentiated 65 individuals with relapsing–remitting MS from 31 individuals with progressive MS, with an area under the receiver operating characteristic curve of 0.76 (p = 1.4 × 10−5). The prognostic classifier predicted prospectively measured Expanded Disability Status Scale scores at follow up (rho = 0.43, p = 2.54 × 10−5).

This multicenter external validation study demonstrates that CSF‐based molecular tests can robustly distinguish MS from other neurological conditions, stratify MS subtypes, and predict future disability progression in real‐world settings. These results lay the groundwork for development of next‐generation molecular tools to personalize care in MS. ANN NEUROL 2026;99:328–340