Prognostic value of blood-based biomarkers in multiple sclerosis patients in the absence of clinical relapses or new MRI lesions

Abstract

Background: In multiple sclerosis (MS), an increase in whole-brain lesion volume (LV) on MRI can be observed even in the absence of newly demarcated focal lesions or clinical relapses. However, it is unknown whether the presence of increasing LV alone is enough to justify changes in the therapeutic regimen. At this point, blood-based biomarkers may aid to identify patients at risk for progression. Objective: To determine the prognostic value of blood-based biomarkers (serum neurofilament (sNfL) and serum glial fibrillary acidic protein (sGFAP)) on disability progression in MS patients without newly demarcated lesions or clinical relapses. Design: Longitudinal cohort study. Methods: In total, out of 291 MS patients who were retrospectively screened for this study, 171 patients underwent a detailed clinical and MRI assessment and were finally included in the analysis: 100 patients with increasing LV (mean baseline Expanded Disability Status Scale (EDSS) = 1.5) and 71 with stable LV over 2 years (mean baseline EDSS = 1.0). Baseline blood-based measures (sNfL and sGFAP) and MRI metrics (total T2-weighted LV, gray matter (GM) volume) were acquired. EDSS worsening served as a clinical outcome measure and was determined through a 2-year follow-up. Receiver operator characteristic analyses were conducted to determine the predictive discriminative power of both blood-based biomarkers. Multivariate logistic regressions were performed to identify independent risk factors for EDSS progression in both cohorts. Results: MS patients with increasing LV had lower GM volume (p = 0.0109, q = 0.0490) and worse EDSS scores (p = 0.0065, q = 0.0650) at clinical follow-up compared to patients with stable LV. Patients with increasing LV and EDSS progression had significantly higher sNfL (p = 0.0049, q = 0.0196), but not sGFAP (p = 0.7425, q = 0.9900) levels. In the logistic regression model, sNfL levels remained an independent risk factor for EDSS progression in patients with increasing LV (odds ratio = 1.344, 95% confidence interval: 1.038–1.739, p = 0.025), but not in patients with stable LV. Finally, in patients with increasing LV, sNfL levels, but not sGFAP levels, discriminated progressive from non-progressive MS patients upon clinical follow-up (area under the curve = 0.67, p = 0.004; q = 0.016). Conclusion: sNfL enhances the prediction of disease progression in MS patients with merely increasing LV on MRI but no new T2 lesions or other signs of inflammatory activity. These findings may support treatment decisions in seemingly stable patients.