Table 1.
Demographic and clinical characteristics of the study population.
Fig 1.
Lesion masking for ELs and NELs as seen on pre- and post-contrast T1w images, T1 map, and ΔT1 map.
A: Lesioned areas (indicated by blue arrows) appeared hypointense on pre-contrast T1-weighted image. B: EL area (indicated by a red arrow) on post-contrast T1-weighted image while NELs remains unenhanced. C: White matter lesions seen on T1 map (EL in red, NEL in blue). D: White matter lesions seen on ΔT1 map (EL in red, NEL in blue, units: s). The color bar displays the full range of ΔT1 (units: s) values only in this participant. EL = enhancing lesion; NEL = non-enhancing lesion; T1w = T1-weighted.
Table 2.
T1 metrics from pre- and post-contrast T1 maps.
Fig 2.
Magnitude of ΔT1 in non-enhanced lesion areas.
Shown are images from three separate participants (A: SPMS/female/53, B: RRMS/female/52, C: SPMS/female/61). Color maps of ΔT1 (units: s) in NEL’s were overlaid upon T1-weighted images. Each color bar (not scaled) displays the full range of ΔT1 only within NEL of the corresponding participant. Three representative participants shown for display purposes (analysis was in all subjects). Each row shows three consecutive image slices of a single participant. SPMS = secondary progressive multiple sclerosis; RRMS = relapsing-remitting multiple sclerosis; BBB = blood-brain barrier; NEL = non-enhancing lesion.
Fig 3.
ΔT1 and Pre-contrast T1 in NEL, NAWM, cGM of entire cohort.
A: Overlaid kernel density estimations (KDEs) of ΔT1 in NEL. Each colored line represents the smoothed histogram using KDE for all voxels in one subject. Blue vertial line indicates voxel-wise mean ΔT1 value (-0.116 s), and green line denotes voxel-wise median ΔT1 value (-0.105 s). B: Per voxel Spearman’s rho coefficient for the relationship of pre-contrast T1 and ΔT1 of all NEL voxels of the entire cohort shows a weak correlation, which indicates that observed ΔT1 is not solely driven by initial T1 before contrast agent administration. C: Overlaid KDEs of pre-contrast T1 in NEL D: Overlaid KDEs of ΔT1 in NAWM. E: Per voxel Spearman’s rho coefficient in NAWM shows a stronger correlation than NEL (blue line, mean ΔT1 of 0.011 s; green line, median ΔT1 of 0.002 s). F: Overlaid KDEs of pre-contrast T1 in NAWM. G: Overlaid KDE of ΔT1 in cGM (blue line, mean ΔT1 of -0.107 s; green line, median ΔT1 of -0.091 s). H: Per voxel Spearman’s rho coefficient in cGM also shows a stronger correlation than NEL. I: Overlaid KDEs of pre-contrast T1 in cGM. NEL = non-enhancing lesion, NAWM = normal-appearing white matter, cGM = cortical gray matter.
Table 3.
Correlation of ΔT1 with WML volume and count.
Table 4.
Correlation of per subject ΔT1 with clinical measures (Part I).
Table 5.
Correlation of per subject ΔT1 with clinical measures (Part II).
Fig 4.
Group differences in median ΔT1 in non-enhancing white matter lesions.
A: NEL median ΔT1 comparison for EDSS scores that were > median (3.0) vs. ≤ median. Gadolinium induced greater changes in T1 in MS participants with higher EDSS scores. B: NEL median ΔT1 comparison for MS Type, PMS vs. RRMS. Gadolinium induced greater changes in T1 in participants with PMS. C: NEL median ΔT1 comparison for multiple sclerosis immunomodulatory treatment, Yes vs. No. Gadolinium induced greater changes in T1 in participants not on MS treatment. Per subject median ΔT1 values were used in the Mann-Whitney U-test. NEL = non-enhancing lesion; EDSS = Expanded Disability Status Scale; MS = multiple sclerosis; PMS = progressive multiple sclerosis (PPMS + SPMS); PPMS = primary progressive multiple sclerosis; SPMS = secondary progressive multiple sclerosis; RRMS = relapsing-remitting multiple sclerosis.
Table 6.
Group differences in ΔT1 metrics.
Table 7.
Multiple logistic regression analysis of significance of treatment effect on ΔT1.