Impact of Focal White Matter Damage on Localized Subcortical Gray Matter Atrophy in Multiple Sclerosis: A 5-Year Study

Article Figures & Data

Download figure
Open in new tab
Download powerpoint
Fig 1.
If WM damage in connected tracts is a primary driver of local SGM atrophy, then pathology represented by (C) ought to explain local SGM atrophy, beyond what is otherwise predicted by whole-brain atrophy and overall lesion growth (A and B) alone. On the other hand, if SGM atrophy largely progresses from independent pathology, we would still expect it to be correlated with overall lesion growth due to shared underlying overall disease activity (via A and B) but we would not expect (C) to explain any additional variance in atrophy. We therefore controlled for overall progression of disease-related pathology (A and B) in our analysis to determine the independent contribution of (C).
Download figure
Open in new tab
Download powerpoint
Fig 2.
The Network Modification tool measures the proportion of disrupted WM tracts connected to a given GM region. An individual WM abnormality mask is referenced with respect to a data base of HC tractograms to probabilistically determine the proportion of connected tract streamlines that pass through a WM abnormality and are therefore considered disrupted.
Download figure
Open in new tab
Download powerpoint
Fig 3.
Mean annualized atrophy (A) and change in WM tract disruption (B) for each SGM region across 5 years. Each sphere represents an SGM region. The size of each sphere is proportional to the magnitude of longitudinal atrophy and increased disruption of connected WM tracts.

Table 1:

Demographic and clinical characteristics of study participants

Note:—CIS indicates clinically isolated syndrome; IQR, interquartile range; RRMS, relapsing-remitting MS; EDSS, Expanded Disability Status Scale; –, not applicable/not evaluable; PPMS, primary-progressive MS; SPMS, secondary-progressive MS.

Table 2:

Annualized atrophy and change in tract disruption^{^a}

	% Atrophy, MS	% Atrophy, HC	P (ANCOVA)	Δ Tract Disruption, MS	P (1-Sample)
Left thalamus	−1.22 ± 1.01	−0.78 ± 0.77	.006^{^b}	0.0051 ± 0.0236	.005^{^b}
Right thalamus	−1.06 ± 0.93	−0.81 ± 0.80	.090	0.0052 ± 0.0204	.001^{^b}
Left caudate	−0.92 ± 1.33	−0.29 ± 2.10	.011^{^b}	0.0224 ± 0.0614	<.001^{^b}
Right caudate	−0.73 ± 1.58	−0.15 ± 2.11	.04^{^b}	0.0285 ± 0.0652	<.001^{^b}
Left putamen	−1.13 ± 1.46	−0.72 ± 1.01	.065	0.0115 ± 0.0307	<.001^{^b}
Right putamen	−0.98 ± 1.67	−0.68 ± 0.85	.245	0.0119 ± 0.0318	<.000^{^b}
Left globus pallidus	−1.90 ± 2.36	−0.72 ± 1.31	.001^{^b}	0.0030 ± 0.0165	.017^{^b}
Right globus pallidus	−1.85 ± 1.93	−0.60 ± 1.11	<.001^{^b}	0.0019 ± 0.0113	.029^{^b}
Left hippocampus	−1.27 ± 1.75	−0.69 ± 1.21	.040^{^b}	0.0015 ± 0.0095	.032^{^b}
Right hippocampus	−0.99 ± 1.80	−0.63 ± 1.52	.232	0.0047 ± 0.0162	<.00^{^b}
Left amygdala	−1.87 ± 3.03	−0.55 ± 2.11	.005^{^b}	0.0014 ± 0.0088	.040^{^b}
Right amygdala	−2.03 ± 3.22	−0.51 ± 2.99	.003^{^b}	0.0025 ± 0.0114	.004^{^b}
Left accumbens	−1.57 ± 4.76	−0.58 ± 3.29	.203	0.0011 ± 0.0079	.073
Right accumbens	−1.52 ± 5.10	0.20 ± 5.21	.051	0.0021 ± 0.0090	.002^{^b}

Note:—Δ Tract Disruption indicates longitudinal change in the disruption of connected WM tracts.
↵a Data are means unless otherwise indicated.
↵b Significant.

Table 3:

Longitudinal analysis controlling for T2LV and WBV^{^a}

Region	Model 1			Model 2			Change
Region	R²	P	P^{^b}	R²	P	P^{^b}	ΔR²	P	P^{^b}
Left thalamus	0.156	.000	.000	0.160	.000	.000	0.004	.181	.984
Right thalamus	0.141	.000	.000	0.147	.000	.000	0.006	.137	.984
Left caudate	0.067	.007	.028	0.074	.005	.020	0.007	.135	.984
Right caudate	0.084	.002	.014	0.092	.001	.008	0.008	.123	.984
Left putamen	0.146	.000	.000	0.164	.000	.000	0.018	.034	.340
Right putamen	0.055	.016	.036	0.130	.000	.000	0.075	.000	.000
Left globus pallidus	0.140	.000	.000	0.164	.000	.000	0.024	.017	.187
Right globus pallidus	0.152	.000	.000	0.161	.000	.000	0.009	.101	.909
Left hippocampus	0.102	.000	.000	0.097	.001	.008	−0.005	.729	1.00
Right hippocampus	0.098	.0001	.000	0.094	.001	.008	0.004	.590	1.00
Left amygdala	0.059	.012	.036	0.062	.012	.024	0.003	.207	.984
Right amygdala	−0.013	.711	.711	0.028	.106	.106	0.041	.005	.065
Left accumbens	0.070	.005	.025	0.104	.008	.024	0.034	.008	.096
Right accumbens	0.084	.002	.014	0.086	.002	.010	0.002	.240	.984

Note:—ΔWBV indicates change in whole-brain volume; ΔT2LV, change in T2 lesion volume.
↵a Hierarchic regression analyses predicting regional SGM 5-year atrophy. Predictors included in each model are as follows: model 1, age, sex, disease duration, ΔWBV, ΔT2LV, baseline tract disruption; model 2, age, sex, disease duration, ΔWBV, ΔT2LV, baseline tract disruption, longitudinal change in disruption of connected WM tracts.
↵b Holm-Bonferroni–corrected P values. All P values are significant.