Cleveland Clinic Foundation
Mellen Center for Multiple Sclerosis Treatment and Research

Author Of 1 Presentation

Imaging Poster Presentation

P0606 - MRI changes over the disease course in a large multiple sclerosis clinical cohort (ID 1318)

Presentation Number
Presentation Topic



Quantitative MRI measures are proposed as biomarkers of disease course and therapeutic response. Understanding the evolution of these metrics is key for interpretation of change in clinical practice.


To describe longitudinal changes in T2 lesion volume (T2LV), whole brain (WBF) and gray matter (GMF) fraction, and thalamic volume (TV) over the disease course in a large multiple sclerosis (MS) cohort.


Demographics, disease history, and MRI were collected from MS patients at a single site. Patients with ≥2 MRI assessments were included. T2LV, WBF, GMF, and TV annualized rate of change and raw values compared to the first available scan were analyzed. Multivariate mixed-effects models were used to evaluate longitudinal MRI changes, adjusting for age at disease onset, sex, and patient-determined disease steps category (PDDS) with a random intercept for patient and an autoregressive covariance structure. For each outcome, three models were generated: a linear model, a second-order B-spline model, and a third-order B-spline model were tested for nonlinearity in the relationship between MRI outcome and disease duration and were compared based on Akaike Information Criterion.


1012 patients were included (69.2% female, 72.9% relapsing-remitting MS, mean ± SD age at disease onset 34.4±10.3, age at baseline MRI 43.8±11.1, disease duration 9.4±5.8 years, mean number of MRIs 3.1±1.2, median [IQR] PDDS 1.0 [0.0-3.0]). Male sex (B=4.9) and PDDS>3 (B=7.0) were associated with greater T2LV accumulation over the disease course (best fit: linear model). T2LV annualized rate of change peaked at 5-6 years of disease duration (rate 9%/year) (best fit: third-order B spline). Male sex, older age, and PDDS>3 were associated with lower WBF, TV (best fit: linear model), and GMF (best fit: second-order B spline), all p<0.05. No non-linear effect of disease duration on WBF, TV, and GMF were observed. There was no statistically significant change in the annualized rate of change of WBF, TV, and GMF over the disease course.


The dynamics of T2LV accumulation are variable throughout the disease course, whereas the rate of change of WBF, TV, and GMF were more stable. These results suggest T2LV accumulation reflecting focal lesion activity predominates early in the disease while WBF, TV, and GMF loss reflecting underlying neurodegeneration is present at disease onset and continues throughout the course.