Cognitive Implications of Deep Gray Matter Iron in Multiple Sclerosis

Abstract

Background and purpose: Deep gray matter iron accumulation is increasingly recognized in association with multiple sclerosis and can be measured in vivo with MR imaging. The cognitive implications of this pathology are not well-understood, especially vis-à-vis deep gray matter atrophy. Our aim was to investigate the relationships between cognition and deep gray matter iron in MS by using 2 MR imaging-based iron-susceptibility measures.

Materials and methods: Forty patients with multiple sclerosis (relapsing-remitting, n = 16; progressive, n = 24) and 27 healthy controls were imaged at 4.7T by using the transverse relaxation rate and quantitative susceptibility mapping. The transverse relaxation rate and quantitative susceptibility mapping values and volumes (atrophy) of the caudate, putamen, globus pallidus, and thalamus were determined by multiatlas segmentation. Cognition was assessed with the Brief Repeatable Battery of Neuropsychological Tests. Relationships between cognition and deep gray matter iron were examined by hierarchic regressions.

Results: Compared with controls, patients showed reduced memory (P < .001) and processing speed (P = .02) and smaller putamen (P < .001), globus pallidus (P = .002), and thalamic volumes (P < .001). Quantitative susceptibility mapping values were increased in patients compared with controls in the putamen (P = .003) and globus pallidus (P = .003). In patients only, thalamus (P < .001) and putamen (P = .04) volumes were related to cognitive performance. After we controlled for volume effects, quantitative susceptibility mapping values in the globus pallidus (P = .03; trend for transverse relaxation rate, P = .10) were still related to cognition.

Conclusions: Quantitative susceptibility mapping was more sensitive compared with the transverse relaxation rate in detecting deep gray matter iron accumulation in the current multiple sclerosis cohort. Atrophy and iron accumulation in deep gray matter both have negative but separable relationships to cognition in multiple sclerosis.

Fig 1.

Sample axial images of the 3 MR imaging methods: T1-weighted image (A), R2* map (B), QSM map (C). D, Oblique view of 3D volume segmentation of the 4 deep gray matter nuclei. The ROIs within each section are shown for one side of the brain, with matching color to the 3D segmentation (caudate = green; putamen = blue; globus pallidus = yellow; thalamus = red).

Fig 2.

Group differences in age-, sex-, and intracranial volume-normalized deep gray matter volumes (A) and iron based on QSM (B) and R2* (C). Boxplots show ranges of the first-to-third quartiles, circles indicate means, lines inside the boxes indicate medians, dotted lines indicate the fence (1.5 interquartile ranges), and outliers are shown by dots outside the fence. Cau = caudate; Put = putamen; GP = globus pallidus; Tha = thalamus; double asterisks = P < .001; asterisk = P < .01.

Fig 3.

Prediction of cognition (NP_total) by age, sex, and intracranial volume-normalized putamen (A) and thalamus (B) volumes in patients with MS after correcting for age, sex, and education.

Fig 4.

Prediction of cognition (NP_total) by age- and sex-corrected QSM (A) and R2* (B) values in the globus pallidus in patients with MS after correcting for age, sex, education, and individual DGM volumes.