Paramagnetic rim lesions are associated with pathogenic CSF profiles and worse clinical outcomes in multiple sclerosis: a retrospective cross-sectional study

Background and Objective: Paramagnetic rims have been observed as a feature of some MS lesions on susceptibility-sensitive MRI and indicate ongoing inflammation, principally consisting of compartmentalized activated microglia/macrophages. We investigated clinical, MRI, and intrathecal (cerebrospinal fluid, CSF) associations of paramagnetic rim lesions (PRL) using 3T MRI in MS. Methods: This is a retrospective, cross-sectional analysis of patients at a single neuroimmunology clinic. All patients had standardized 3T MRI using a multiecho T2*-weighted sequence with susceptibility postprocessing (SWAN protocol, GE) as part of the inclusion criteria. SWAN-derived filtered phase maps and corresponding T2-FLAIR images were manually reviewed by one expert rater blinded to clinical data, and PRL were determined based on qualitative assessment of hypointense paramagnetic edges on corresponding T2-hyperintense lesions. Descriptive statistics, t-tests, ANOVA, and linear regression determined demographic, clinical, MRI, and intrathecal profile associations with the presence of one or more PRL. Results: One hundred and forty-seven (147) MS patients were included in this analysis (2 clinically isolated syndrome, 118 relapsing-remitting, 14 secondary progressive, 13 primary progressive). Baseline mean age was 48.8 years, disease duration 12.8 years, and median EDSS 2, with 79% women. Seventy-five percent of patients were receiving a disease-modifying therapy, and 79 patients (54%) had available cerebrospinal fluid (CSF) analysis. Sixty-three patients (43%) had at least 1 PRL. PRL status (presence or absence) did not vary by sex or EDSS but was associated with younger age (51 vs 46 years; p=0.01) and shorter disease duration (14.5 vs 10.5 years; p=0.01). PRL status was also associated with worse disease (MS severity score: 2.8 vs 3.7; p=0.05) and blood-brain barrier disruption as determined by higher protein and pathologically elevated albumin quotient, as well as the presence of CSF oligoclonal bands (all p≤0.05); there was no association with immunoglobulin index or synthesis rate. PRL status was additionally associated with higher burden of T2-hyperintense cerebral lesion volume (T2LV), higher age-adjusted cerebral brain volume loss (especially of gray matter), and poorer performance on multiple clinical measures, including the 9-hole peg test and symbol digit modalities test (but not timed 25-foot walk speed). Clinical and intrathecal profiles remained associated with PRL after adjustment for age and in many cases T2LV as well. Sensitivity analyses limited to subgroups of patients without disease activity at the time of CSF sampling remained supportive of results. Patients with PRL were being treated with higher-efficacy disease-modifying therapies at the time of the data query. Conclusions: PRL, an emerging noninvasive biomarker of chronic cerebral neuroinflammation in MS, are confirmed to be associated with greater disease severity and newly shown to be associated with intrathecal inflammation and blood-brain-barrier disruption.