Inverse Relationship between Plasma Neurofilament Light Chain Concentration and Symbol Digit Modalities Test Scores in the Sunbeam Trial of Ozanimod in RMS

Background: Cognitive dysfunction, including diminished cognitive processing speed (CPS), is a major source of impairment in patients with multiple sclerosis (MS). In the phase 3 SUNBEAM trial, the sphingosine 1-phosphate receptor 1 and 5 modulator ozanimod reduced plasma neurofilament light chain concentration (pNfL, a marker of neurodegeneration) and improved CPS based on the Symbol Digit Modalities Test (SDMT) compared with intramuscular interferon (IFN) β-1a.

Objectives: This post hoc analysis evaluated relationships between pNfL concentration and SDMT scores at baseline and on ozanimod treatment in SUNBEAM.

Methods: In the double-blind, double-dummy SUNBEAM trial (NCT02294058), patients with relapsing MS were randomized to once-daily oral ozanimod 0.92 or 0.46 mg or weekly intramuscular IFN β-1a 30 µg, continued for ≥12 months. pNfL (analyzed with Simoa technology) and SDMT (part of a secondary endpoint) were assessed at baseline and month 12. Relationships between pNfL and SDMT at baseline were evaluated with Kendall’s τ correlation, and changes on treatment were evaluated with linear regression and treatment-stratified bootstrap sampling.

Results: SUNBEAM included 1346 patients (mean [SD] age 35.6 [9.3] y; time since symptom onset 7.0 [6.2] y; Expanded Disability Status Scale 2.6 [1.1]): 447 received ozanimod 0.92 mg, 451 ozanimod 0.46 mg, and 448 IFN β-1a. At baseline, median (IQR) pNfL was 14.7 (10.2, 23.3) pg/mL and SDMT score was 48.0 (38.0, 56.0); Kendall’s correlation (95% CI) between these variables was -0.10 (-0.14 to -0.06), indicating a slight negative association. Based on 1000 bootstrap samples, greater median percent reduction in pNfL was associated with greater mean SDMT change from baseline at month 12; both ozanimod doses were associated with greater median reductions in pNfL and mean improvements in SDMT than IFN β-1a, with 0.92 mg showing the greatest differences from IFN β-1a. There were no relationships noted between baseline pNfL or change from baseline at month 12 in pNfL and any specific SDMT category (worsened/stable/improved based on ≥4-point change from baseline at month 12).

Conclusions: In this exploratory, post hoc analysis of the SUNBEAM trial, baseline pNfL correlated inversely with baseline SDMT. Ozanimod treatment decreased pNfL and improved SDMT to a greater extent than IFN β-1a. Future prospective analyses are warranted to determine the clinical utility of pNfL as a marker of CPS.