Evidence for progressive gray matter loss in patients with relapsing-remitting MS

Assessing brain volume changes on serial MRI scans is an established, reliable way to estimate the amount of progressive tissue loss known to occur in patients with multiple sclerosis (MS) as a result of neurodegeneration.¹ Brain atrophy is a consistent feature of all MS phenotypes,² and progressive brain tissue loss can be observed at the earliest clinical stage of the disease.³ Additional MRI work has shown that tissue loss and intrinsic damage^2,4 of cerebral gray matter (GM) might also contribute to the development of overall brain atrophy.¹

To our knowledge, only one preliminary study has explored the temporal evolution of GM loss in 13 patients with early relapsing–remitting MS (RRMS) followed up with MRI every 6 months for 18 months.⁵ In this study, we investigated the temporal profile of GM volume changes and how such changes relate to other MRI markers of MS evolution in a large sample of patients with more advanced RRMS who underwent monthly MRI scans over a period of 9 months as the placebo arm of a treatment trial.⁶

Methods.

The current analysis is based on MRI data from the patients enrolled in the placebo arm of the European/Canadian glatiramer acetate trial.⁶ All patients had RRMS and underwent physical and neurologic examination and brain MRI at screening, at baseline, and every month for 9 months. Conventional spin-echo sequences were used to obtain proton density (PD) and T2-weighted images. Two series of T1-weighted images were obtained before and after the injection of 0.1 mmol/kg of gadolinium (Gd). For each visit, T2-hyperintense, T1-hypointense, and Gd-enhancing lesion volumes (LVs) were calculated using a semiautomated technique. Additional information about trial design and MRI acquisition/postprocessing is reported elsewhere.⁶ On the precontrast T1-weighted images, normalized GM volumes were measured using SIENAx, the cross-sectional version of the SIENA (Structural Imaging Evaluation of Normalized Atrophy) software.⁷ SIENAx uses a brain extraction tool to perform segmentation of brain from nonbrain tissue in the head and to estimate the skull surface. Then, the extracted brain image is segmented into white matter (WM), GM, and CSF, yielding an estimate of the absolute volumes of brain tissue compartments. The original MRI image is registered to a canonical image in a standardized space (derived from the MNI152 standard space) to provide a spatial normalization scaling factor for each patient. The estimated absolute volumes for each subject are then multiplied by the normalization factor to yield a normalized parenchymal volume of these tissue compartments. To test the reproducibility of this technique, we repeated the measurements of normalized GM, WM, and normalized brain volumes on the whole set of scans belonging to 10 randomly selected patients. For each metric, the measurement variability was assessed by using the coefficient of variation (CoV), defined as the SD of a given variable divided by its mean value.

We investigated the time trend of GM volumes using a mixed random effect model, which accounted for the autocorrelation of repeated measures within each patient. This model was corrected with an adjustment for baseline GM volumes. The correlations between GM volume changes over the study period and other MRI parameters were assessed using the Spearman rank correlation coefficient.

Results.

Data for GM volume measurements were usable from 117 of 120 patients (97.5%). Three of the 120 patients originally assigned to the placebo arm did not contribute to the current analysis because of suboptimal image quality of 5 or more of the available 10 scans/patient, which prevented a reliable application of SIENAx to define an accurate temporal profile of GM volume changes. Clinical and MRI findings at baseline and follow-up are reported in the table. The mean values of the intraobserver CoVs were 0.01% (range = 0.007 to 0.03%) for normalized GM volume, 0.02% (range = 0.002 to 0.07%) for normalized WM volume, and 0.004% (range = 0.001-0.006%) for normalized brain volume measurements. The time–trend analysis revealed a progressive GM volume decrease over the study period (p < 0.001), with an estimated mean percentage change per month of −0.30%, whereas WM volume did not change significantly over the study period (figure). These results did not change when correcting for the corresponding LVs (data not shown). GM volume at baseline was correlated with the corresponding T2-hyperintense (r = −0.40, p < 0.001) and T1-hypointense (r = −0.22, p = 0.015) LVs. No significant correlation was found between the change of GM volume and those of other MRI parameters over the study period.

• Download figure
• Open in new tab
• Download powerpoint

Figure. Average normalized gray matter (A), white matter (B), and total brain volumes (C) and 95% CIs (bars) as detected by monthly scans over a 9-month follow up in 117 untreated patients with relapsing–remitting multiple sclerosis.

Discussion.

Our results are consistent with those of a previous longitudinal study⁵ of 13 patients with RRMS followed up for 18 months, with MRI obtained every 6 months. Although the duration of our follow-up was shorter than that of the previous report,⁵ we believe that the current study adds to the existing literature, because it was conducted on a much larger sample of patients, who also had a longer disease duration (mean of 8.3 vs 1.9 years). Another strength of the current study is the frequent sampling of MRI data, which allowed an accurate estimate of the temporal profile of GM damage. The results of these two studies taken together suggest that progressive GM loss is a consistent feature of MS in its relapsing–remitting phase. GM tissue loss might be preceded or accompanied by progressive intrinsic GM damage, as shown by a recent longitudinal study of RRMS followed for 18 months, with diffusion tensor MRI obtained every 3 months.⁵ In accord with previous evidence highlighting the clinical relevance of GM damage in MS,⁸ these results may also have important implications for the planning of trials of neuroprotective therapies in RRMS, where GM volume changes might serve as a meaningful measure of outcome.

White matter volume stability in RRMS might be the result of an earlier onset of WM atrophy, as shown by a study in patients with clinically isolated syndromes suggestive of demyelination,⁹ where WM but not GM atrophy was found. Another, but not mutually exclusive, explanation for WM volume stability might be that WM atrophy does also occur in RRMS patients but is masked by edema and cellular infiltration associated with inflammation, which is known to occur at a much lesser extent in GM.¹⁰

Two possible mechanisms (not mutually exclusive) can explain the presence of evolving GM pathology in MS: the progressive accumulation of GM lesions and neuronal damage due to retrograde degeneration of axons passing through WM lesions. The latter explanation is supported by our finding of a correlation between GM volume and MRI-visible lesion burdens at baseline, which is in keeping with the results of other studies.^2,5 Nevertheless, the lack of significant correlation between GM atrophy development and increased WM lesion loads also suggests that GM pathology cannot only be a consequence of WM damage, and that MS-related damage to the two brain tissue compartments may follow distinct patholobiogical pathways. Despite a previous post mortem study that found abundant axonal transection in active lesions from patients with MS,¹⁰ we did not find any significant correlation between evolving GM atrophy and enhancing LV changes. This might be explained by the relative paucity of enhancing lesions, which typically represent a small part of the overall brain tissue, or by the relatively short duration of the follow-up, which prevented us from investigating a possible long-term effect of inflammation on GM tissue loss.

Acknowledgment

The authors thank Teva for allowing us to conduct this natural history analysis using data from placebo patients enrolled in the European/Canadian MRI-monitored trial of glatiramer acetate.