Permanent tissue damage in multiple sclerosis lesions is associated with reduced pre-lesion myelin and axon volume fractions

Abstract

Background: The use of advanced magnetic resonance imaging (MRI) techniques in MS research has led to new insights in lesion evolution and disease outcomes. It has not yet been determined if, or how, pre-lesional abnormalities in normal-appearing white matter (NAWM) relate to the long-term evolution of new lesions.

Objective: To investigate the relationship between abnormalities in MRI measures of axonal and myelin volume fractions (AVF and MVF) in NAWM preceding development of black-hole (BH) and non-BH lesions in people with MS.

Methods: We obtained magnetization transfer and diffusion MRI at 6-month intervals in patients with MS to estimate MVF and AVF during lesion evolution. Lesions were classified as either BH or non-BH on the final imaging visit using T₁ maps.

Results: Longitudinal data from 97 new T₂ lesions from 9 participants were analyzed; 25 lesions in 8 participants were classified as BH 6-12 months after initial appearance. Pre-lesion MVF, AVF, and MVF/AVF were significantly lower, and T₁ was significantly higher, in the lesions that later became BHs (p < 0.001) compared to those that did not. No significant pre-lesion abnormalities were found in non-BH lesions (p > 0.05).

Conclusion: The present work demonstrated that pre-lesion abnormalities are associated with worse long-term lesion-level outcome.

Keywords: MRI; T2 lesions; biomarkers; demyelination; multiple sclerosis; quantitative MRI.

Figure 1.

Representative lesion masks. New T₂-w lesion masks (green) and spatially matched contralateral NAWM masks (blue) are overlaid on the pre-lesion (6 months prior to lesion appearance (T-6)) T₁-w anatomical (top) and T₂-w FLAIR (bottom) images. The anterior lesion (open arrow) is a discrete de novo lesion arising entirely from NAWM with no voxels spatially connected to existing lesions. The second lesion (open arrowhead) is an expanding/enlarging lesion, and the spatially matched contralateral NAWM ROI was manually adjusted to avoid an existing T₂ lesion while maintaining total ROI volume.

Figure 2.

Parametric maps showing lesion evolution. Parametric maps demonstrating lesion evolution in the de novo (open arrow) and expanding/enlarging (open arrowhead) lesions from Figure 1. AVF change is not visually appreciable in this slice at T₀ in the de novo lesion, where axonal loss follows acute inflammation and demyelination. AVF, MVF, and T₁ worsen progressively over time in both lesions, ultimately resulting in black-hole fate at T+12.

Figure 3.

Time course plots. Time course plots represent mean differences between lesion ROIs and matched contralateral NAWM ROIs. Contralateral ROIs were masked to remove non-WM tissue (i.e. cortex, lesion, cerebrospinal fluid) and were subsequently manually adjusted to maintain consistent ROI volume. Time courses were shifted in time such that Time 0 corresponds to the first time each lesion was observed on MRI. Lesions containing clusters of at least 8 connected voxels with T₁ > 1600 ms at the last imaging timepoint were classified as black-hole lesions (plotted in green). Bars represent the standard errors. Pre-lesion and post-acute values are calculated for each lesion by averaging all available timepoints in the gray- and orange-shaded boxes, respectively.

Figure 4.

Pre-lesion boxplots. Boxplots comparing pre-lesion values (calculated for each ROI by averaging all available timepoints in the Figure 3 gray-shaded boxes) between lesion ROIs and contralateral NAWM ROIs matched in size and homologous location to each lesion ROI. Pre-lesion MVF, AVF, MVF/AVF, and T₁ in black-hole lesions are significantly different from contralateral NAWM. Mixed-effects models were used to calculate p-values and significance was set at the 0.05 level. Outliers are indicated by red + symbols. ***p < 0.001.

Figure 5.

Parameter change boxplots. Boxplots illustrating parameter change for each lesion type and NAWM (i.e. MVF_drop = post-acute MVF − pre-lesion MVF). On average, NAWM ROIs showed no significant change over time. Aside from AVF in non-BH lesions, all other signals showed longitudinal changes that were significantly greater than in contralateral NAWM ROIs matched in size and homologous location to each lesion ROI. Recovery in BH lesions was substantially attenuated compared to non-BH lesions: MVF and AVF were significantly decreased, and T₁ was significantly increased (p < 0.001 for all comparisons). Interestingly, (MVF/AVF)_drop was not different between BH and non-BH lesions. Mixed-effects models were used to calculate p-values and significance was set at the 0.05 level. Outliers are indicated by red + symbols. *p < 0.05; ***p < 0.001.