Author Of 1 Presentation
P0557 - Characterizing 1-year development of cervical cord atrophy across different MS phenotypes: a voxel-wise, multicenter analysis (ID 1115)
Abstract
Background
In multiple sclerosis (MS) the cervical spinal cord is often affected by demyelination and neuro-axonal injury, leading to irreversible tissue loss.
Objectives
To use voxel-wise analysis to evaluate the distribution and changes over time of cervical cord atrophy in MS patients from a multicentre dataset acquired at 7 European sites.
Methods
Baseline and 1-year 3D T1-weighted cervical cord scans and clinical evaluation were obtained from 54 healthy controls (HC) and 110 MS patients (13 clinically isolated syndromes [CIS], 75 relapsing-remitting [RR] and 22 progressive [P]MS). A pipeline optimized for longitudinal analysis was used to co-register baseline and 1-year follow-up cervical cord scans to a cord template, obtained by averaging straightened HC images from all centers. Voxel-wise differences of cervical cord atrophy, their longitudinal changes and correlations with clinical variables were assessed using SPM12 and full factorial models (sex-, age-, center- and total cord volume-corrected).
Results
Compared to HC, MS patients exhibited significant (p<0.05, family-wise error [FWE] corrected) baseline cervical cord atrophy, mainly located in anterior, posterior and lateral cord regions at C1/C2, as well as in posterior regions between C4 and C6. While CIS patients showed a slight cord tissue expansion vs HC at posterior C4, RRMS presented significant clusters of cord atrophy vs CIS, mostly in lateral and posterior C2-C4 regions, and PMS showed widespread cord atrophy vs RRMS patients at C4-C5 and C7 levels. During the follow-up, a significant progression (p<0.05, FWE) of cord atrophy was detected in MS patients, predominantly in the posterior and lateral cord at C2, and between C4 and C6. Such pattern of cord atrophy progression was mainly driven by RRMS patients, while CIS patients did not show cord tissue loss at follow-up vs baseline, and PMS patients showed circumscribed tissue loss in posterior regions at C2 and C6. A strong relationship (p<0.05, FWE) was found between baseline clinical disability and baseline cord atrophy in the posterior and lateral cord at C2-C4. Also, baseline atrophy in the lateral cord at C3-C4 correlated with clinical disability at 1-year follow-up.
Conclusions
Voxel-wise analysis of cervical atrophy allowed to detect a differential involvement of cord levels and to characterize 1-year evolution of tissue loss across phenotypes. Cord atrophy was clinically relevant and contributed to explain follow-up clinical disability.