Background: High-efficacy disease-modifying therapies (HETs) have substantially improved multiple sclerosis (MS) management, yet ongoing cognitive decline remains a concern. This study aims to assess Symbol Digit Modalities Test (SDMT) changes in patients with stable relapsing-remitting MS (RRMS) treated with HETs and to evaluate the role of baseline MRI biomarkers as predictors of SDMT changes.

Methods: Consecutive patients with RRMS treated with HETs underwent clinical, SDMT and MRI assessment at baseline with SDMT and clinical re-evaluation after 24 months. Patients presenting relapses or MRI activity (new T2 and/or gadolinium-enhancing lesions) during follow-up were excluded. Cognitive changes were defined using the 90% CI regression-based reliable change index methodology accounting for sex, age, education and baseline score. Baseline MRI examination included three-dimensional-sagittal Fluid Attenuated Inversion Recovery (FLAIR), T1-Magnetization Prepared - RApid Gradient Echo (T1-MPRAGE) and quantitative susceptibility mapping (QSM) for paramagnetic rim lesions (PRLs) and QSM-isointense lesions (ISO) assessment. Univariate and multivariable regression analyses were performed to predict SDMT changes.

Results: 90 patients (mean age: 40.3 years, median Expanded Disability Status Scale: 2.0) were included. PRLs were present in 46 (51.1%) patients. After 24 months, 13 (14.4%) patients showed SDMT decline and 8 (8.9%) showed improvement. At multivariable analyses, PRLs were associated with higher risk of SDMT decline (β: 2.70, p: 0.02, OR: 14.82) while higher ISO lesion volumes were weakly associated with SDMT improvement (β: 0.07, p: 0.01, OR: 1.07).

Conclusions: SDMT decline and improvement are detectable in patients with RRMS without clinical or MRI activity over 2 years. PRLs seem to predict SDMT decline in MS, underscoring the critical role of compartmentalised chronic inflammation in disease progression.