Aims

The neuronal-cell-density-specific R2t* metric of the quantitative Gradient-Recalled-Echo (qGRE) MRI signal allows identification in brain tissue of two components – neuron-containing Viable Tissue (VT) and tissue void of neurons - Dark Matter (DM). This allows assessing brain microstructural neurodegeneration in AD pathology prior to detectable atrophy using structural MRI measurements. We report on associations between VT volume and DM fraction with PET-defined tauopathy and their relationships with cognitive performance.

Methods

47 participants (20 cognitively normal and 27 with mild AD - Clinical Dementia Rating® 0.5 or 1) recruited from the Knight Alzheimer Disease Research Center underwent qGRE MRI, 18F-flortaucipir (AV-1451) tau PET imaging and Free and Cued Selective Reminding Test (FcSRT). Correlation analysis (age-adjusted) was performed between the qGRE metrics (VT volume, and DM fraction) in the hippocampus, total hippocampal volume and the global PET metric, i.e., Tauopathy (averaging the Standard Uptake Value ratio of AV-1415 tracer in the inferior temporal, entorhinal, lateral occipital, and amygdala regions), and FcSRT.

Results

As shown in figure, the qGRE biomarkers in the hippocampus, DM fraction and VT volume, have stronger associations with Tauopathy (r=0.60; r=-0.58), and SRTfree cognitive score (r=-0.42; r=0.39) than hippocampal volume with Tauopathy (r=-0.43) and FcSRT (r=0.19). The association between FcSRT and Tauopathy was slightly higher (r=-0.5) than with DM.

Conclusions

The MRI-based qGRE measurements of tissue micro-neurodegeneration are significantly stronger associated with Tauopathy and cognitive decline than brain macro-neurodegeneration (atrophy), thus providing a better understanding of in vivo pathological changes of neurodegeneration in early onset of AD.

Aims

Biological heterogeneity in Alzheimer’s disease (AD) has been observed as subtypes in measures of neurodegeneration, predominantly in atrophy and more recently in hypometabolism. However, the multimodal relationships of subtypes inferred from the two modalities remain unclear. Thus, we aim to compare AD subtypes using these modalities, FDG-PET (hypometabolism) and MRI (atrophy).

Methods

Concurrent FDG-PET and MRI scans from 171 amyloid-positive AD individuals and 120 amyloid-negative healthy controls from the ADNI were included. FDG-PET uptake and MRI volumes in cortical and subcortical regions were used in separate but technically identical modality-specific clustering (random forest hierarchical) analyses for subtype identification. We compared (a) hypometabolism-based subtypes and atrophy-based subtypes across the two analyses, and (b) Spatial atrophy patterns of hypometabolism-based subtypes and vice-versa for atrophy-based subtypes.

Results

Subtyping revealed four hypometabolism-based (two relatively limbic-predominant: 45.5%, two relatively cortical-predominant: 54.3%) and three atrophy-based (diffuse: 18.7%, relatively limbic-predominant: 28.6%, minimal: 52.6%) AD subtypes with distinct spatial patterns and demographic/clinical characteristics. We observed that (a) hypometabolism-based subtypes shared cortical, but not subcortical, spatial correspondence with atrophy-based subtypes , (b) distinct hypometabolism-based subtypes always reflected comparable AD-typical atrophy patterns (involving the hippocampus, temporal regions) while distinct atrophy-based subtypes always reflected comparable AD-typical hypometabolism patterns (involving retrosplenial cortex).

Conclusions

Our findings suggest that hypometabolism-based and atrophy-based AD subtypes are not interchangeable. Despite being measures of neurodegeneration, hypometabolism and atrophy may not share the same extent of heterogeneity, bearing a complex multimodal relationship. Disentangling such relationships and combining complementary biomarkers to investigate heterogeneity could aid personalized treatment of AD.

Aims

The chronological progression of brain atrophy over decades, from presymptomatic to dementia stages, has never been formally depicted in Alzheimer’s disease (AD) due to the lack of large cohorts with long enough MRI follow-up in cognitively unimpaired young participants at baseline.

Methods

To describe such a spatiotemporal atrophy staging of AD at the whole-brain level, we build extrapolated lifetime volumetric models of healthy and AD brain structures by combining multiple large-scale databases (n=3512 quality controlled MRI from 9 cohorts of subjects covering the entire lifespan, including 415 MRI from ADNI1, ADNI2, and AIBL for AD patients). Then, we validated dynamic models based on cross-sectional data using external longitudinal data. Finally, we assessed the sequential divergence between normal aging and AD volumetric trajectories.

Results

Chronologically, five major stages described AD structural progression: (1) hippocampus and amygdala; (2) middle temporal gyrus; (3) entorhinal cortex, parahippocampal cortex, and other temporal areas; (4) thalamus and striatum and (5) middle frontal, anterior cingular, parietal, insular cortices and pallidum.

Conclusions

This MRI scheme of atrophy progression in AD is close but do not entirely overlap with Braak staging of tauopathy, with a “reverse chronology” between limbic and entorhinal stages. AD structural progression may be associated with local tau accumulation but may also be related to axonal degeneration in remote sites and to other limbic-predominant associated proteinopathies.

Aims

In Alzheimer’s disease patients, we combined the spatial information derived from fMRI-meta-analyses for episodic memory, language, executive functioning, and visuospatial abilities with tau-PET and examined whether estimates of tau pathology in functional critical regions allow precision-medicine prediction of domain-specific cognitive decline.

Methods

We tested 356 ADNI participants including cognitively normal, mild cognitively impaired and demented participants with ~2-year cognitive follow-up. All participants underwent baseline ¹⁸F-flortaucipir-tau-PET, amyloid-PET, and broad cognitive testing for determining summary scores on memory, language, executive function, and visuospatial ability. Meta-analytical masks from task-based functional MRI studies for memory, language, executive function, and visuospatial ability were downloaded from neurosynth and applied to tau-PET (“cognitive maps”). Cognitive change was calculated using linear mixed models. In bootstrapped linear regression we assessed the predictive accuracy of the respective cognitive map compared to global and temporal-lobe tau for future cognitive change. Models were controlled for age, sex, education, diagnosis, and baseline cognitive performance. In addition, a personalized global cognitive composite was computed (Fig.1).

Results

Bootstrapped linear regression revealed highest explained variance (partial R²) for longitudinal cognitive decline when using cognitive maps as predictors vs. global or temporal-lobe tau (Fig.2). A PET-informed personalized global cognitive composite enhanced the sensitivity to detect cognitive decline (Fig.3) resulting in lower sample sizes for detecting simulated intervention effects.

Conclusions

Tau distribution patterns encode critical information on domain-specific cognitive decline. We provide a personalized tau-PET-based estimation of future cognitive change, which may improve participant selection and determining personalized clinical endpoints in clinical trials.

Aims

The volume of the hippocampus decreases more slowly than the volume of the cortex during normal aging. We explored changes in the hippocampus-to-cortex volume (HV:CTV) ratio with increasing age in non-demented Parkinson’s disease (PD) patients as compared to healthy controls (HC). We also evaluated the association between the HV:CTV ratio and cognitive outcomes.

Methods

Altogether 130 participants without dementia aged 51 to 88 years were consecutively enrolled, including 54 PD patients (mean age 67, standard deviation (SD) 8 years) and 76 HC (mean age 69, SD 7 years). All participants underwent structural magnetic resonance examination and psychological evaluation. Hippocampal and cortex volumes were determined from T1 and FLAIR scans using FreeSurfer software, and the HV:CTV ratio was calculated. Regression lines for age dependance of the HV:CTV ratio for PD and HC groups were calculated. We further assessed the association between the HV:CTV ratio and cognitive tests examining hippocampus-related cognitive functions.

Results

PD patients and age-matched HC showed a significant difference in age dependance of HV:CTV ratio (p-value = 0.019), with a decreasing slope in PD and increasing slope in HC. In the PD group, a significant correlation (R = 0.561, p = 0.024) was observed between the HV:CTV ratio and the Digit Symbol-Coding test.

Conclusions

The reduction of HV:CTV ratio is accelerated in pathological aging due to PD pathology. The HV:CTV ratio was associated with impaired processing speed, i.e. the cognitive function that is linked to subcortical alterations of both associated basal ganglia circuitry and the hippocampus.

Aims

Dopaminergic projections to the striatum and prefrontal cortex (PFC) are damaged in Parkinson’s disease (PD) and contribute to motor and cognitive deficits. The underlying pathophysiology of diffusion MRI markers were only studied in animal models. Therefore, we investigated the integrity of nigrostriatal and nigro-PFC tracts in association with PD pathology in the substantia nigra (SN) of clinically-defined and pathologically-confirmed PD and non-neurological control donors.

Methods

We included 14 PD (7 non-demented and 7 demented) and 10 controls for post-mortem in-situ 3T-MRI: 3DT1 images and DTI-derived fractional anisotropy (FA) and mean diffusivity (MD), assessing the microstructural integrity of the SN and its tracts to the striatum and dorsolateral-PFC (DLPFC) with FSL. After autopsy, SN sections were immunostained for Tyrosine hydroxylase (TH) and pSer129 alpha-synuclein, and analyzed in Qupath. Group comparisons and MRI-pathology associations were assessed using general linear model and Spearman’s correlation.

Results

PD cases tend to have higher nigrostriatal tract MD compared to controls (p=.067), while no difference in SN FA and MD, or nigro-DLPFC tract FA and MD were observed. Nigrostriatal tract MD positively associated with SN Lewy Body (LB) density (r=0.6,p=.01), whereas nigro-DLPFC tract MD negatively associated with SN TH thread load (r=-0.55,p=.027). No associations were found for FA. Within our limited sample we could not address the differential contribution of dementia in these associations.

Conclusions

Our results indicate that microstructural disintegration, specifically measured by MD, of the nigrostriatal tract is linked with LB pathology in the SN, whereas the nigro-DLPFC tract disintegration is associated with dopaminergic loss in the SN.

Abstract Body

Advanced imaging techniques to study changes in the aging brain and dementia are potential biomarkers of dementing diseases. Among them are diffusion tensor imaging DTI) , magnetization transfer imaging (MTI) and iron mapping.

Measures derived from DTI allow reliable quantification of subtle white matter tissue alterations. A marker of special interest is the peak width of skeletonized mean diffusivity (PSMD) which combines DTI, skeletonization of white matter tracts, and the analysis of mean diffusivity histograms. PSMD increases with advancing age, but its differential diagnostic and prognostic role in Aalzheimer´s disease (AD) is widely unknown. Recently, advanced DTI metrics, such as diffusional kurtosis imaging (DKI), neurite orientation dispersion and density imaging (NODDI) and free water imaging (FWI) have been introduced. Most studies on advanced DTI models focused on FWI with increased FW content considered to represent a marker of neuroinflammation.

Magnetization transfer imaging (MTI), other than DTI offers information on tissue composition. The magnetization transfer ratio (MTR) associates with axonal attenuation and myelin content and in AD both white matter and gray matter MTR changes are important. Yet, the histopathologic correlates of MTR in the cortex are unknown.

Iron mapping may correlate to neurodegeneration. In AD, iron accumulation has been reported, but it is unclear if it is indeed causally linked to AD or represents a sole epiphenomenon of neurodegenerative disease. We only recently showed that in AD patients iron levels increase over a relatively short time period. Increasing iron in the temporal lobe was associated with cognitive decline.