Development and Interrelation of Spatiotemporal Patterns of Brain Atrophy and Lesions in Patients with a First Clinical Demyelinating Event in the REFLEX/ION Study

Background: Treatment with subcutaneous interferon beta-1a (sc IFN β-1a) significantly reduced MRI outcomes in patients (pts) with a first clinical demyelinating event (FCDE) in the 24-month REFLEX study. However, the spatiotemporal relation between atrophy and lesion processes remains unclear.

Objectives: Study regional spatiotemporal development of brain atrophy and white matter (WM) lesions in FCDE pts, how these processes are related, and how this relates to treatment and conversion to clinically definite multiple sclerosis (CDMS).

Methods: Yearly brain images from the 5-year REFLEX/ION trial were analyzed in 392 pts presenting with FCDE. Pts received early (from baseline; N=262; ET) or delayed (from Month-24; N=130; DT) sc IFN β-1a treatment; 162 pts converted to CDMS and 230 did not. Yearly SIENA images and maps of growing/new/shrinking/disappearing lesions were obtained and non-linearly registered on a study template (voxel size=8mm³) using FSL tools. For each year, 3 different voxelwise analyses were used to assess the relation between: 1) atrophy and conversion/treatment; 2) lesion activity and conversion/treatment; and 3) atrophy and total lesion volume change (TLVC) in the prior year. Analyses were corrected for age, sex, and site (p<0.05, cluster-corrected); when testing for conversion/treatment, treatment/conversion were used as covariates. Volume (mm³) of significant voxels V was reported.

Results: Pts converted to CDMS had higher posterior lateral ventricle atrophy versus non-converters in Years 1, 2, and 4 (V=6880, V=1016, and V=8152, respectively). In Year 1, greater ventricular atrophy was found in ET versus DT pts (V=28,768). In Year 2, increased frontal lobe atrophy was found in DT versus ET pts (V=10,568). In corona-radiata, increased growing lesion activity was found in CDMS converters versus non-converters in Year 5 (V=2864) and DT versus ET pts in Year 1 (V=448). There was a significant positive relationship between periventricular atrophy and previous TLVC in Years 3, 4, and 5 (V=34,408, V=25,888, and V=7656, respectively).

Conclusions: Periventricular atrophy and lesion activity increased faster in pts who converted to CDMS during the study. Pseudoatrophy was detected immediately following treatment onset, but DT pts showed greater atrophy in Year 2 and increased lesion activity in Year 1. Periventricular atrophy is predicted by TLVC in the prior year.