Adam S Bernstein1, Steve Z Rapcsak2, Michael Hornberger3, and Manojkumar Saranathan4
1College of Medicine, University of Arizona, Tucson, AZ, United States, 2Department of Neurology, University of Arizona, Tucson, AZ, United States, 3Department of Medicine, University of East Anglia, Norwich, United Kingdom, 4Department of Medical Imaging, University of Arizona, Tucson, AZ, United States
Synopsis
Using a novel multi-atlas segmentation technique, we studied the atrophy
of thalamic nuclei as a function of disease progression in Alzheimer’s disease.
We found statistically significant atrophy of the anteroventral, centromedian,
and mediodorsal nuclei, which are part of the limbic
system and known to play a known role in memory and cognitive function. We also found atrophy
of the medial geniculate nucleus and the pulvinar nucleus. The degree of atrophy increases from early MCI to full AD. These findings suggest that
a larger network of brain structures are affected in Alzheimer’s disease, which
together lead to the clinical presentation.
Introduction
Increasing evidence suggests that the thalamus, and particularly the
limbic thalamic nuclei may play a central role in the pathogenesis of
Alzheimer’s disease (AD).1 While whole thalamic volume atrophy and shape changes have been reported, only
limited research has been conducted to investigate atrophy in individual thalamic nuclei in AD, likely due to the paucity of accurate high
resolution thalamic nuclei segmentation methods. This work
utilizes a novel multi-atlas segmentation technique to parcellate the
thalamus into individual nuclei and explore changes in thalamic nuclear volumes
across the AD continuum (i.e. prodromal and clinical) using data from the ADNI database. We hypothesize that
there will be significant atrophy of specific thalamic nuclei, particularly those
associated with memory and the Papez circuit, including the anteroventral (AV)
nucleus, and the mediodorsal (MD) nucleus, with increasingly severity leading up to AD.Methods
3T T1-weighted Magnetization Prepared RApid Gradient Echo (MPRAGE) data from the ADNI database were used in
this study. Healthy controls (HC, n=119), early mild cognitive impairment (EMCI, n=208), late mild cognitive impairment (LMCI, n=116), and AD subjects (n=91) were
selected. The MPRAGE data were parcellated using a variant of the recently
proposed THalamus Optimized Multi-Atlas Segmentation (THOMAS) technique of Su et al2 to determine volumes of 11 individual thalamic nuclei on each side. For this work, THOMAS was modified to use majority
voting for label fusion (as opposed to joint fusion) to enable the use of standard CSF-nulled T1-weighted
MPRAGE data as input as opposed to white-matter-nulled (WMn) MPRAGE of Su et al.
The proposed segmentation pipeline is depicted in Figure 1. Briefly, a WMn template
image was generated by mutual registration and averaging of 20 WMn MPRAGE priors, each with manually labelled thalamic
nuclei. The template was first rigidly registered to the subject image using ANTs.3 Following rigid registration,
both the template and the subject image were cropped to improve registration
speed (by 10X) and accuracy. The cropped template image was then
nonlinearly registered to the subject image, using ANTs (SyN diffeomorphic). The nonlinear warps from each of the 20 priors to the template is pre-computed and available. Using a composite of prior-template and template-subject warps, manual nuclei labels from all 20 priors were warped into subject space
and then combined using a majority voting based label fusion scheme to produce subject-specific thalamic nuclei labels.
Thalamic nuclear volumes across the
different clinical subgroups were compared using ANCOVA after correcting for
intracranial volume and age. In addition to volume comparisons across groups, correlations of limbic nuclei volumes with neurocognitive
testing scores, clinical disease metrics including the clinical dementia rating
(CDR) and the Alzheimer’s disease assessment scale with 13 elements (ADAS13),
and phosphorylated tau protein (P-Tau) and beta amyloid (Aβ) levels were performed.Results
Thalamic nuclei segmentation labels from the modified THOMAS method overlaid on MPRAGE data in axial (top panel) and coronal (bottom panel) are shown in Figure 2 for a patient with AD (enlarged ventricles). Small nuclei such as anteroventral (AV), centromedian (CM), and the habenula (Hb) are clearly delineated.
Results of the volume analysis are shown in Figure 3. The volumes of the AV, Pul, MGN, CM, and MD nuclei were significantly smaller in subjects with AD when compared to HC subjects. Further, the AV, Pul, MGN, and MD nulcei were significantly smaller in subjects with LMCI when compared to HC subjects. No significant thalamic changes were found between HC subjects and EMCI patients.
Cohen’s d, a metric that measures effect size, is plotted in Figure 4 for all nuclei that demonstrated significant volume differences. There is a progressive
increase in effect size with increasing disease severity when volumes are
compared to those of the HC group. The correlation analysis, summarized in Table 1, demonstrates mild to moderate correlations that are statistically significant for the AV and MD nuclei with
neurocognitive scores (MoCA and RAVLT) and clinical assessments (CDR and
ADAS13). Of all the thalamic nuclei, the AV nuclei volumes were the only ones that demonstrated a significant correlation to P-tau levels. The MD showed a
significant correlation with Aβ.Conclusions and Discussion
While previous work has shown whole thalamic volume atrophy and thalamic
nuclei in AD4, this is the first work to systematically examine thalamic
nuclei atrophy across prodromal and clinical AD using a fast and accurate
thalamic nuclei segmentation method. This preliminary work highlights nuclei-specific atrophy within the
thalamus in subjects with LCMI and AD. This is consistent with the hypothesis
that the memory and cognitive changes in AD are mediated by damage to a large,
integrated neural network that extends beyond the medial temporal lobes,
(notably the hippocampus and related structures) which have been the
traditional focus of attention in AD research.Acknowledgements
Data collection and sharing for this project was funded by the Alzheimer's Disease Neuroimaging Initiative (ADNI) (National
Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense
award number W81XWH-12-2-0012). ADNI is funded by the National Institute on
Aging, the National Institute of Biomedical Imaging and Bioengineering, and
through generous contributions from the following: AbbVie, Alzheimer’s
Association; Alzheimer’s Drug Discovery Foundation; Araclon Biotech;
BioClinica, Inc.; Biogen; Bristol-Myers Squibb Company; CereSpir, Inc.;
Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company;
EuroImmun; F. Hoffmann-La Roche Ltd and its affiliated company Genentech, Inc.;
Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research
& Development, LLC.; Johnson & Johnson Pharmaceutical Research &
Development LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; Meso Scale
Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis
Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda
Pharmaceutical Company; and Transition Therapeutics. The Canadian Institutes of
Health Research is providing funds to support ADNI clinical sites in Canada.
Private sector contributions are facilitated by the Foundation for the National
Institutes of Health (www.fnih.org). The grantee organization is the Northern
California Institute for Research and Education, and the study is coordinated
by the Alzheimer’s Therapeutic Research Institute at the University of Southern
California. ADNI data are disseminated by the Laboratory for Neuro Imaging at
the University of Southern California.References
-
Aggleton JP. Multiple anatomical systems
embedded within the primate medial temporal lobe: Implications for hippocampal
function. Neurosci Biobehav Rev.; 36:1579-1596 (2012)
- Su JH, Thomas FT, Kasoff WS, et al.
Thalamus Optimized Multi Atlas Segmentation (THOMAS): fast, fully automated
segmentation of thalamic nuclei from structural MRI. Neuroimage; 194:272-282 (2019).
- Avants BB, Tustison NJ, Stauffer M, Song
G, Wu B, Gee JC. The Insight ToolKit image registration framework. Front
Neuroinform. 2014;8(APR). doi:10.3389/fninf.2014.00044
- Iglesias JE et al. A probabilistic atlas of the human thalamic nuclei combining ex vivo MRI and histology. Neuroimage; 183:314-326 (2018)