Axial sections from selected levels of the unsmoothed gray matter template in patients with benign multiple sclerosis. The regions of significantly reduced gray matter volume are shown in yellow. The z values refer to Montreal Neurologic Institute space coordinates.
Axial sections from selected levels of the unsmoothed gray matter template in patients with secondary progressive multiple sclerosis. The regions of significantly reduced gray matter volume are shown in yellow. The z values refer to Montreal Neurologic Institute space coordinates.
Regions with decreased gray matter concentration in patients with secondary progressive multiple sclerosis compared with benign multiple sclerosis (data obtained using statistical parametric mapping). A bilateral volume reduction of several cerebellar gray matter regions was found, with a preferential left-sided distribution.
Axial sections from selected levels of the unsmoothed gray matter template in patients with secondary progressive multiple sclerosis compared with those patients with benign multiple sclerosis without cognitive impairment. The regions of significantly reduced gray matter volume are shown in yellow. A predominant infratentorial localization of gray matter loss can be seen, with some other clusters in the right occipital lingual gyrus, left hypothalamus, and both thalami. The z values refer to Montreal Neurologic Institute space coordinates.
Mesaros S, Rovaris M, Pagani E, Pulizzi A, Caputo D, Ghezzi A, Bertolotto A, Capra R, Falautano M, Martinelli V, Comi G, Filippi M. A Magnetic Resonance Imaging Voxel-Based Morphometry Study of Regional Gray Matter Atrophy in Patients With Benign Multiple Sclerosis. Arch Neurol. 2008;65(9):1223-1230. doi:10.1001/archneur.65.9.1223
Evidence is accumulating that indicates that a selected assessment of gray matter (GM) damage is able to provide strong paraclinical correlates of multiple sclerosis (MS) severity.
To investigate the pattern of regional GM atrophy in patients with benign MS (BMS) vs those with secondary progressive MS (SPMS) to better elucidate the factors associated with a favorable status in patients with MS.
Cross-sectional survey from January 2006 to August 2007.
Referral, hospital-based MS clinics.
Sixty patients with BMS, 35 patients with SPMS, and 21 healthy volunteers.
Main Outcome Measures
Neuropsychological tests exploring memory, attention, and frontal lobe cognitive domains were administered to BMS patients. A voxel-based morphometry analysis of GM concentration was performed using statistical parametric mapping and a threshold of 0.05, corrected for multiple comparisons.
Twelve BMS patients (20%) had an abnormal performance on 3 or more neuropsychological tests. Compared with healthy individuals, BMS patients had a reduced GM volume in the subcortical and frontoparietal regions. Compared with BMS patients, those with SPMS had a significant GM loss in the cerebellum. No differences between BMS and SPMS patients were found when only BMS patients with cognitive impairment or those with shorter disease duration (15-19 years) and higher Expanded Disability Status Scale scores (>2.0) were considered.
Cerebellar GM atrophy seems to be a major determinant of irreversible locomotor disability in MS. The absence of cognitive impairment and a longer disease duration or lower Expanded Disability Status Scale score may identify those BMS patients with the potential for a favorable disease evolution.
Multiple sclerosis (MS) can follow a benign clinical course “in which the patient remains fully functional in all neurological systems 15 years after disease onset.”1(p909) However, the diagnosis of benign MS (BMS) can be done only retrospectively. Indeed, although numerous clinical features have been proposed as possible prognostic factors for BMS,2 follow-up studies3,4 have shown that many of these patients can still enter a disabling course after 15 years of disease duration. Moreover, cognitive dysfunction can be found in up to 45% of patients with BMS,5 indicating that the current definition of BMS, which considers the absence of locomotor disability as the main index of preserved functions, is likely to overestimate the prevalence of this condition.
Evidence is accumulating that indicates that a selected assessment of gray matter (GM) damage6,7 is able to provide strong paraclinical correlates of MS severity. Regional GM loss in MS has been described in several studies,8- 12 using different technical approaches. Atrophy of the GM can be related to markers of MS severity, such as locomotor disability,10- 12 cognitive impairment (CI),9 and magnetic resonance imaging (MRI) lesion load.8,12- 14
We performed the present voxel-based morphometry (VBM) study15,16 of BMS patients with the following aims: (1) to assess the topographic patterns of GM atrophy in BMS patients and (2) to better clarify the mechanisms that lead to a mild or absent neurologic impairment in BMS by comparing VBM findings in BMS patients with those from patients with secondary progressive MS (SPMS) with irreversible locomotor disability.
Patients were selected from the outpatient MS clinic populations of participating institutions from January 2006 to August 2007. To be included, patients had to be relapse and steroid free for at least 1 month. All patients underwent a complete neurologic examination, with rating of the Expanded Disability Status Scale (EDSS) score.17 Criteria for a diagnosis of BMS were an EDSS score of 3.0 or less and a disease duration of 15 years or more. Patients with MS, an SP course,1 and persistent walking limitations (ie, a confirmed EDSS score of ≥4.0) were enrolled to serve as controls. Twenty-one healthy controls (HCs) (11 women and 10 men; mean age, 45.7 years; age range, 25-66 years) were also studied.
Within 48 hours of MRI acquisition, BMS patients underwent neuropsychological tests exploring the following domains: (1) attention and information processing: Paced Auditory Serial Attention Test (3 seconds version),18 Trail Making Test,19 and Attentive Matrices Test20; (2) verbal and visual-spatial memory: Digit Span Test, Short Story Test, Corsi Span Test, Word List Test, and Rey-Osterrieth Complex Figure Test–recall task21; (3) abstract reasoning (Raven Test)22; (4) executive functions: Token Test,20 Verbal Fluency Test,20 and Wisconsin Card Sorting Test23; and (5) spatial cognition (Rey-Osterrieth Complex Figure Test–copy task).21 For each patient, the results from all neuropsychological tests were scored through a comparison with the percentile distribution of values from HCs.20 The individual test scores ranged from 0 to 4, where grade 4 means a normal performance. Only those patients with score of 0 in at least 3 tests were considered affected by CI.24 Local ethical committee approval and written informed consent from each individual were obtained before study initiation.
Brain MRIs were obtained using a 1.5-T scanner. The following sequences were collected during a single session: (1) dual-echo turbo spin-echo (repetition time [TR], 3300 milliseconds; echo time [TE], 16/98 milliseconds; echo train length, 5; 24 axial sections; thickness, 5 mm; matrix size, 256 × 256; field of view [FOV], 250 × 250 mm2), (2) sagittal 3-dimensional T1-weighted magnetization prepared rapid acquisition gradient echo (MP-RAGE) (TR, 9.7 milliseconds; TE, 4 milliseconds; flip angle, 15°; number of partitions, 128; section thickness, 1.5 mm; inversion time, 300 milliseconds; matrix size, 256 × 256; voxel size, 0.82 × 0.82 × 1.50; FOV, 210 × 210mm2). For dual-echo scans, the sections were positioned to run parallel to a line that joins the most inferoanterior and inferoposterior parts of the corpus callosum.25
Postprocessing MRI was performed by a single observer (S.M.) blinded to the study participants' identity. Brain MS lesions were identified on the dual-echo scans and lesion volumes (LVs) were measured.26 On MP-RAGE images, the intracranial volume (ICV) was measured using SIENAx software.27
Regional volumetry measurements were performed on MP-RAGE images, using an optimized VBM approach and statistical parametric mapping software.15,16 Full details of the steps involved in the optimized method of VBM analysis, as well as our application of this method in assessing GM atrophy, are extensively described elsewhere.14- 16 In the present study, the GM mask was thresholded at a value of 0.5 and then used as an explicit mask in the analysis.
An analysis of covariance was used to compare volumetry measurements among groups. Age, sex, and ICV were included as nuisance covariates. A multivariate analysis was used to assess the correlation between GM loss and clinical or MRI variables, with ICV, age, and sex as nuisance covariates. We considered P < .05 as statistically significant, after correction for multiple comparisons with the family-wise error method for all statistical analyses.
Sixty patients with BMS and 35 patients with SPMS were studied (Table 1). Twelve BMS patients (20%) were considered affected by CI.
The BMS patients had a lower brain T2-weighted LV than the SPMS patients (median values, 11.9 and 18.8 mL; P = .001). Infratentorial T2-weighted LV was lower in BMS than in SPMS patients (median values, 0.05 and 0.89 mL; P < .001). The ICV was lower in SPMS patients (median, 1212.0 mL) than in BMS patients (median, 1303.0 mL; P = .001) and in HCs (median, 1341.0 mL; P = .003). The ICV did not differ between the HCs and the BMS patients (P = .24).
Compared with HCs, a reduced GM concentration was found in BMS patients, including predominantly bilateral subcortical GM regions (Figure 1). For cortical regions, a preferential left-sided GM loss was found. The voxel number, involved Brodmann area, and statistical parametric mapping coordinates of each cluster are reported in Table 2. Compared with HCs, patients with SPMS showed bilateral clusters of significantly reduced GM concentration, affecting many cortical and subcortical regions and the cerebellum (Table 3 and Figure 2).
When compared with BMS patients, significant reductions in GM concentration in the SPMS patients were found only in the cerebellum (Table 4 and Figure 3). A similar pattern of differences in GM concentration was found when comparing BMS patients without CI with SPMS patients, with a predominant infratentorial localization of GM loss (Figure 4). No differences were found between BMS patients with CI and SPMS patients. No clusters of significant reduction in infratentorial white matter concentration were found when comparing SPMS patients with BMS patients.
When the subgroup of 35 BMS patients (58%) with long-lasting disease and mild impairment (ie, those with a disease duration of ≥20 years and an EDSS score of ≤2.0) was compared with SPMS, the latter group again showed significant symmetrical GM loss in the cerebellar cortex (posterior lobe). No differences in regional GM concentration were found between the remaining BMS patients and the SPMS patients.
In BMS patients, T2-weighted LV correlated with the severity of GM loss in the thalami (left, r = −0.71, P < .001; right, r = −0.59, P = .02), caudate bodies (right, r = −0.66, P = .001; left, r = −0.61, P = .01),and the right putamen (r = −0.59, P = .03). In SPMS patients, T2-weighed LV correlated with GM concentration in the right caudatus (r = −0.84, P < .001), left caudatus (r = −0.83, P < .001), right thalamus (r = −0.82, P = .001), left thalamus (r = −0.74, P = .03), right putamen (r = −0.80, P = .002), left putamen (r = −0.79, P = .006), left hypothalamus (r = −0.79, P = .005), and right postcentral gyrus (Brodmann area 2)(r = −0.79, P = .004). No correlation was found between infratentorial LV and the severity of cerebellar GM loss observed in SPMS patients compared with BMS patients. No correlations were found between regional GM atrophy and disease duration or EDSS score.
To our knowledge, this is the largest study ever performed to investigate the regional patterns of GM atrophy in patients with BMS. By comparing the distribution of GM concentration changes between BMS and SPMS patients, we also aimed at achieving a better understanding of the nature of MS-related disability. We focused our attention on GM damage, based on the increasing evidence that this aspect of MS plays a central role in determining locomotor disability.28
Compared with HCs, BMS patients displayed significant tissue loss mostly in the deep GM regions bilaterally but also in several frontoparietal regions with more pronounced left-sided differences. Significant GM loss in the thalamus has been described in other nondisabling forms of MS, including early relapsing-remitting MS (RRMS)11 and pediatric MS.14 The observed preferential left-sided GM loss is consistent with previous VBM findings in RRMS,8 but the localization of the affected areas is different. Cortical GM loss in RRMS seems to have a predominant frontotemporal localization,8 whereas in our BMS patients the frontoparietal regions were mostly affected. On the contrary, SPMS patients displayed a more widespread and bilateral GM loss. Nevertheless, the only significant difference between BMS and SPMS patients was a greater cerebellar GM loss in the latter group. Although we did not find any significant correlation between EDSS score and regional GM loss, our results indicate that the occurrence of irreversible damage in the cerebellum may be a major determinant of locomotor disability in MS, as also suggested by other quantitative MRI studies.29,30
The lack of significant differences between BMS and SPMS patients in terms of GM concentration in cortical and subcortical regions other than the cerebellum is in contrast with the different clinical profile of the 2 disease phenotypes and the notion that brain GM atrophy is associated with an unfavorable MS course.10,12 A possible explanation for our finding is the preserved ability of the BMS-affected brain to limit the clinical impact of GM disease by means of effective compensatory mechanisms, such as cortical reorganization. Another explanation may be the relative absence of spinal cord damage in BMS vs SPMS patients.31,32 On the other hand, however, the similarities between the 2 groups on a cross-sectional basis may also depend on the misclassification as BMS of some patients who may subsequently enter an SPMS phase.3,4
The BMS patients with CI did not show any differences in GM concentration compared with the SPMS patients. Conversely, those BMS patients without CI showed significant differences in the GM concentration of both supratentorial and infratentorial GM regions when compared with SPMS patients. This finding is consistent with the notion that BMS patients with preserved cognitive functions represent those with a “truly” benign course. In addition, SPMS patients compared with BMS patients with an EDSS score of 2.0 or lower after a 20-year or longer disease duration still showed significantly reduced cerebellar GM concentration, whereas no differences were found between SPMS patients and those BMS patients with higher EDSS scores and shorter disease durations.
The results of our correlation analysis suggest a significant role of T2-weighted lesion load in determining the severity of atrophy in deep GM structures, such as basal ganglia and thalami, whereas in only 1 cortical region (ie, the right postcentral gyrus of SPMS patients), GM loss was associated with T2-weighted LV. These findings are consistent with the described relationship between deep GM loss and T2-weighted lesion load in MS,11,13,14 suggesting that some features of MS, such as retrograde neuroaxonal degeneration or anterograde or transsynaptic changes from axonal transection in white matter lesions, may be possible determinants of such GM atrophy. Intriguingly, no relationship was found between the areas of more pronounced cerebellar GM loss in SPMS and the regional T2-weighted LV, suggesting that clinically relevant cerebellar atrophy may be independent of local white matter disease.
Correspondence: Massimo Filippi, MD, Neuroimaging Research Unit, Scientific Institute and University Ospedale San Raffaele, via Olgettina 60, 20132 Milan, Italy (email@example.com).
Accepted for Publication: March 11, 2008.
Author Contributions:Study concept and design: Filippi. Acquisition of data: Mesaros, Rovaris, Pulizzi, Caputo, Ghezzi, Bertolotto, Capra, Falautano, and Martinelli. Analysis and interpretation of data: Mesaros, Rovaris, Pagani, Comi, and Filippi. Drafting of the manuscript: Mesaros and Filippi. Critical revision of the manuscript for important intellectual content: Rovaris, Pagani, Pulizzi, Caputo, Ghezzi, Bertolotto, Capra, Falautano, Martinelli, Comi, and Filippi. Statistical analysis: Pagani. Obtained funding: Filippi. Administrative, technical, and material support: Mesaros, Rovaris, and Pagani. Study supervision: Filippi.
Financial Disclosure: None reported.
Funding/Support: This study was supported by grant 2005/R/18 from Fondazione Italiana Sclerosi Multipla. Dr Mesaros is supported by a Fellowship of the European Neurological Society.