The evolution of tissue damage in paramagnetic rim lesions

Introduction Paramagnetic rim lesions (PRL) have gained interest as a diagnostic and prognostic multiple sclerosis (MS) biomarker, as they are associated with chronic white matter inflammation, ongoing neurodegeneration and higher clinical disability.1–3 We’ve recently shown that PRL-associated tissue damage appears correlated with lesion age, with older lesions featuring higher magnetic resonance imaging (MRI) T1 times and being associated with greater clinical disability.4 However, how tissue damage evolves during the process of lesion development and how this correlates with ongoing neuroaxonal damage as measured by serum neurofilaments light chain (sNfL) levels remains unclear. Objective To investigate the development of PRL-associated tissue damage using sNfL and serial 3T MRI. Material and Methods A cohort of 149 MS patients (97 females; 103 relapsing MS and 46 progressive MS) underwent 3T MRI using 3D FLAIR and MP2RAGE4 sequences. Lesions were manually segmented on FLAIR, and the presence of an acute or chronic lesional inflammatory activity was assessed respectively on 3D-EPI unwrapped phase5 and post-gadolinium (gad) T1 images. Lesional mean T1 was extracted using MP2RAGE-derived T1 maps and lesional volumes were calculated. Lesions were categorized based on their mean T1-times as previously described in short-T1 and long-T1 lesions.4 Lesion age was retrospectively calculated from initial time point of gad enhancement on clinical and/or study MRIs. Clinical status was assessed using the Expanded Disability Status Scale (EDSS) and patients were categorized based on treatment regimen. Relapse activity was defined as radiological or clinical activity within the last 90 days preceding clinical assessment. MRI-matched blood samples were collected and sNfL levels were measured using the Atellica Immunoassay System (Siemens Healthineers Atellica® IM). All sNfL levels were then z-score corrected.6 Lesion based statistical analyses included repeated measures mixed models and generalized additive models both with correction of random subject and lesion effects. Clinical analyses included t-tests, correlation analyses and regression analyses when appropriate. Results A total of 129 age-tracked lesions were analysed (58 PRL and 71 non-PRL) for a median time of 14 months (range 0-46). Both lesion types showed an initial increase in T1 times during gad enhancement, with PRL showing significantly higher T1 times (p = 0.02; Figure 1). After enhancement resolution T1 times decreased in both groups, suggesting reduction of oedema-associated T1 changes. While non-PRL lesions showed no further T1 increase, PRL lesions exhibited a secondary increase in T1 times, suggesting progressive tissue damage (Figure 1). After correction for random subject effects, long-T1 PRL were significantly older than short-T1 PRL (p=0.002; age long-T1 PRL = 2.6 years vs age short T1 PRL = 2 years), while no differences in lesion age were found between short-T1 and long-T1 non-PRL (p=0.82). PRL were also significantly larger than non-PRL (p=0.009; volume non-PRL = 66.55mm3 vs volume PRL = 148.41mm3), although no differences in volume during gad enhancement were found (p=0.076; Figure 2) In line with our previous findings,4 and after correcting for relapse activity and treatment status, only total PRL volume and long-T1 PRL volume were associated with EDSS scores (p=0.011 and p=0.002 respectively), while no associations were found between EDSS and short-T1 PRL volumes (p=0.3; Table 1). Regression analysis, adjusted for relapse activity, showed an association between sNfL levels and both total PRL and short-T1 PRL volumes (respectively p=0.016 and p=0.021), while long-T1 PRL volumes were not associated with sNfL levels (p=0.08, Table 2). Conclusion These results indicate that PRL undergo significant tissue damage over time, with long-T1 PRL representing older lesions associated with established tissue damage and greater clinical disability. In contrast, short-T1 PRL are younger and are associated with ongoing axonal damage as mirrored by elevated levels of sNfL. In line with our prior results,4 these findings highlight the importance of characterizing lesion heterogeneity and differentiating PRL subtypes when assessing disease activity. Future research should better clarify the physiopathological processes behind this lesional heterogeneity in MS, in order to refine its importance for the diagnostic and prognostic work-up of MS.