MRI graph parameters are longitudinal markers of neuronal integrity in multiple sclerosis

Purpose: We sought to determine if structural network parameters add to traditional markers of MS treatment response following immunoablation and autologous haemopoietic stem cell transplantation (IAHSCT). The postIAHSCT paradigm afforded us the opportunity to study MS patients after relapsing biology had been effectively suppressed, enabling us to study the cortical substrate of progressive MS in a less confounded manner. Methods: In this analysis of data from a phase 2 prospective study, associations between magnetic resonance graph parameters, N-acetylaspartate to creatine ratio (NAA/Cr), and serum neurofilament light chain (sNfL), amongst other markers, were assessed at 3 months pre-and 12 months post-IAHSCT. Correlations between graph parameter score changes and markers of brain health were calculated. Predictive factors of NAA/Cr or sNfL levels were calculated, adjusting for reference models. Model improvements were evaluated using the G2 likelihoodratio test. Results: 24 patients (aged 18-38) were evaluated. Post-IAHSCT, high NAA/Cr and low sNfL (both measures of neuronal injury) were respectively associated with more favourable degree, density, clustering and path lengths, and degree, gamma, and path length. Post-IAHSCT, absolute change in degree, path length and gamma were associated with NAA/Cr and sNfL. Multivariate analysis demonstrated that the relative change in network parameters after IAHSCT accounted for 14% and 35% more variance in NAA/Cr and sNfL levels respectively than the reference model alone. Conclusions: Cross-sectionally and longitudinally, network parameters demonstrate added utility as markers of disease severity in MS. These measures have the potential to capture cortical changes relevant to progressive nonrelapsing biology in MS.