Multi-Parametric Imaging Assessment of Alzheimer’s Disease Pathology
Eva-Maria Ratai1,2, Kimberly A. Stephens1,2, Alison E. Goldblatt1,2, Jean-Philippe Coutu1,2, Ciprian Catana1,2, Diana Rosas2,3, and David Salat1,2

1Radiology, Massachusetts General Hospital, Boston, MA, United States, 2A. A. Martinos Center for Biomedical Imaging, Charlestown, MA, United States, 3Neurology, Massachusetts General Hospital, Boston, MA, United States

Synopsis

The purpose of this study was to find associations between metabolites measures by MRS and glucose metabolism using FDG PET and cerebral blood flow (CBF) measured by ASL MRI in the assessment of Alzheimer’s Disease. Our study shows an association between increased myo-inositol, a marker of glial inflammation, and hypo-metabolism measured by FDG as well hypo-perfusion measured by ASL. Liner regression analysis revealed that creatine, a marker of altered energy metabolism positively correlated with increased glucose uptake by FDG PET. Increased levels of glutamate+glutamine (contributing to excitotoxicity) were related to decreased metabolic activity by PET and decreased CBF.

Purpose

The search for biomarkers that can detect and track the disease progression in individuals with Alzheimer’s disease (AD) has been a major pursuit. Regional hypo-metabolism measured by fluorodeoxyglucose positron emission tomography (FDG PET) has been found to scale with AD severity1. Several studies reported areas of hypo-perfusion measured by arterial spin labeling (ASL) MRI overlap considerably with hypo-metabolism. However, some studies have also noted increased regional cerebral blood flow (CBF) of individuals with early stage clinical AD, which may indicate an initial compensatory response to neurodegeneration2. Brain metabolites obtained through MRS are related to post-mortem neuropathological changes in individuals with Alzheimer’s disease; in particular decreases in N-Acetylaspartate/creatine (NAA/Cr), a marker of neuronal integrity and myo-inositol/Cr an index of gliosis/inflammation were found to be associated with higher Braak stage, and greater likelihood of AD3. A recent review of 705 distinct metabolite reports concluded that alterations in glutamate clearance from the synaptic cleft may contribute an excitotoxic component to AD pathology4. The purpose of this study was to find associations between metabolites measures by MRS and glucose metabolism using FDG PET and CBF measured by ASL MRI.

Methods

15 participants (8 men, 7 women, mean 79 age ± 9y/o) were consented and enrolled in this study. Of the 15 participants 6 had been diagnosed with AD, 5 with mild cognitive impairment (MCI) and 4 participants served as matched cognitively healthy older adults (CHOA). Imaging consisted of a 18F-FDG PET scan using a 3 Tesla mMR PET-MRI (Siemens, Erlangen), ASL MRI, and MR spectroscopy using a Trio (Siemens, Erlangen). Cerebral blood flow (CBF) was calculated from the ASL data. Single voxel 1H MR spectra were acquired from the posterior cingulate gyrus (VOI=2x2x2cm3) using a point resolved spectroscopy sequence with water suppression enhanced through T1 effects, and the following parameters: TE=30 ms, TR=1700 ms, and 128 acquisitions. In addition, water unsuppressed spectra were acquired from the same region to estimate ‘absolute’ concentrations. All spectra were processed offline using LCModel software to determine the quantities of the brain metabolites NAA, Cr, choline (Cho), mI, and glutamine+glutamate (Glx). Cortical reconstructions were performed on each individual from T1 data using Freesurfer (surfer.nmr.mgh.harvard.edu) and FDG and CBF measures were mapped to the cortical surface. General linear models tested for the association between metabolites and surface based CBF and FDG measures using the mri_glmfit tool provided in Freesurfer. ANOVA and Student t tests were used to compare metabolite concentrations between groups. Discriminant analysis was used to determine which variables discriminate between AD and MCI or CHOA. All analyses were conducted using JMP 11.0.

Results

ANOVA showed a trend toward significant differences in mI between the 3 groups (ANOVA p=0.08) with higher mI levels found in individuals with AD compared to CHOA (p=0.05) and MCI (p=0.05). None of the other MRS metabolic markers showed significant differences between groups which can be attributed to a limited sample size. There was an association between increased mI and regional hypo-metabolism in the brain measured by FDG (Figure 1) as well hypo-perfusion measured by ASL (Figure 2). Discriminant analysis showed a good separation between individuals with AD and MCI or CHOA when combining mI and FDG or CBF as variables. In contrast, other MRS metabolites such as Cr and Glx showed a non-group dependent association with cerebral FDG-uptake or CBF. Liner regression analysis across all participants revealed that Cr, a marker of altered energy metabolism positively correlated with increased glucose uptake by FDG PET (Figure 3). Increased Glx was related to decreased metabolic activity by PET (Figure 4) and decreased CBF (Figure 5). Of note, levels of NAA had the least associations with FDG PET or CBF in our population.

Discussion

A few studies have investigated multimodal approaches to establish biomarkers for AD5. Our study compared MRS markers of inflammation, (mI), altered energy metabolism (Cr) and excitotoxicity (Glx) to glucose metabolism (FDG PET) and perfusion. Both markers of energy metabolism, Cr measured by MRS and FDG uptake measured by PET showed associations across all groups. Combining markers of inflammation and hypo-metabolism or hypo-perfusion provided a good classification between individuals with AD and MCI or CHOA. Lastly, these data may support the idea that excitotoxic pathology may be a feature of AD which results in or is the result of hypometabolism and hypo-perfusion. However, these conclusions are highly speculative given the small sample size examined here and future work will be necessary to clarify these results.

Conclusion

Multi-parametric imaging may be used to detect multipathologic conditions in AD and MCI individuals varying in their combined pathology.

Acknowledgements

This works was supported by NIH grant R01 NR010827 and Biogen Idec.

References

1. Mosconi L. Glucose metabolism in normal aging and Alzheimer's disease: Methodological and physiological considerations for PET studies. Clin Transl Imaging 2013;1(4).

2. Wolk DA, Detre JA. Arterial spin labeling MRI: an emerging biomarker for Alzheimer's disease and other neurodegenerative conditions. Curr Opin Neurol 2012;25(4):421-428.

3. Kantarci K, Knopman DS, Dickson DW, Parisi JE, Whitwell JL, Weigand SD, Josephs KA, Boeve BF, Petersen RC, Jack CR, Jr. Alzheimer disease: postmortem neuropathologic correlates of antemortem 1H MR spectroscopy metabolite measurements. Radiology 2008;248(1):210-220.

4. Ellis B, Hye A, Snowden SG. Metabolic Modifications in Human Biofluids Suggest the Involvement of Sphingolipid, Antioxidant, and Glutamate Metabolism in Alzheimer's Disease Pathogenesis. J Alzheimers Dis 2015;46(2):313-327.

5. Zimny A, Bladowska J, Macioszek A, Szewczyk P, Trypka E, Wojtynska R, Noga L, Leszek J, Sasiadek M. Evaluation of the posterior cingulate region with FDG-PET and advanced MR techniques in patients with amnestic mild cognitive impairment: comparison of the methods. J Alzheimers Dis 2015;44(1):329-338.

Figures

Figure 1: Linear regression between myo-Inositol (mI) and FDG uptake in the left banks of the superior temporal sulcus (P = 0.006, R = -0.72).

Linear Discriminant analysis revealed a Misclassification of only 3/13 (23%).

Red symbols indicate individuals with Alzheimer’s disease (AD), blue symbols indicate individuals with mild cognitive impairment (MCI) and green symbols indicate matched cognitively healthy older adults (CHOA).


Figure 2: Linear regression between myo-Inositol (mI) and cerebral blood flow in the medial orbitofrontal cortex (P = 0.004, R = -0.73).

Linear Discriminant analysis revealed a Misclassification of only 2/13 (15%).

Red symbols indicate individuals with Alzheimer’s disease (AD), blue symbols indicate individuals with mild cognitive impairment (MCI) and green symbols indicate matched cognitively healthy older adults (CHOA).


Figure 3: Linear regression between creatine (Cr) and FDG uptake in the left caudal middle frotal gyrus (P = 0.002, R = 0.76).

Red symbols indicate individuals with Alzheimer’s disease (AD), blue symbols indicate individuals with mild cognitive impairment (MCI) and green symbols indicate matched cognitively healthy older adults (CHOA).


Figure 4: Linear regression between glutamine + glutamate (Glx) and FDG uptake in the posterior cingulate (P = 0.0008, R = -0.81).

Red symbols indicate individuals with Alzheimer’s disease (AD), blue symbols indicate individuals with mild cognitive impairment (MCI) and green symbols indicate matched cognitively healthy older adults (CHOA).


Figure 5: Linear regression between glutamine + glutamate (Glx) and cerebral blood flow in the left fusiform gyrus (P = 0.007, R = -0.71).

Red symbols indicate individuals with Alzheimer’s disease (AD), blue symbols indicate individuals with mild cognitive impairment (MCI) and green symbols indicate matched cognitively healthy older adults (CHOA).




Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
4032