PET AND MRI STUDIES IN ATYPICAL VARIANTS OF ALZHEIMER’S DISEASE

Abstract

Alzheimer's disease (AD) typically presents with amnestic deficits, but there is considerable cognitive heterogeneity and several atypical variants of AD have been described, including posterior cortical atrophy (PCA, “visual variant of AD”), logopenic variant primary progressive aphasia (lvPPA, “language variant of AD”) and the behavioral/dysexecutive variant of AD (bdAD). The remarkable diversity of these syndromes, as well as the relative paucity of confounding co-pathologies, provide an intriguing model for studying mechanisms that drive clinico-anatomic heterogeneity in AD. I will provide a comprehensive overview of the literature and additionally will present data from ongoing studies applying structural magnetic resonance imaging (MRI), [18F]fluorodeoxyglucose (FDG) positron emission tomography (PET), amyloid-β PET and tau PET in PCA, lvPPA and bdAD. Both neurodegenerative imaging markers (i.e. structural MRI and FDG PET) have consistently identified both shared and disease-specific affected regions in AD variants compared to amnestic-predominant AD. Regions with shared vulnerability typically involve wide regions of the temporoparietal cortex, including posterior cingulate, precuneus, medial temporal lobe structures and lateral temporal and parietal cortex. Disease-specific brain regions include occipital, biparietal and occipitotemporal cortices for PCA and left-hemisphere dominant language network hubs in temporoparietal cortex for lvPPA. Tau PET retention patterns largely overlap with these neurodegenerative regions and extend into regions that are not (yet?) atrophic or hypometabolic. Amyloid-β PET shows a much weaker regional correspondence with neurodegeneration and is distributed widely throughout the neocortex, largely irrespective of the clinical phenotype. For bdAD the clinico-anatomical correlates are less intuitive as we found a rather posterior pattern of brain atrophy with relative sparing of the frontal cortex, despite severe behavioral impairments in these patients. Tau and amyloid-β PET studies are currently ongoing in bdAD. These findings support a network model of neurodegeneration in which both the posterior default mode network (i.e. temporoparietal regions) and syndrome-specific vulnerable networks are affected in distinct AD variants. Furthermore, PET studies provided preliminary in vivo evidence that tau is more closely linked to neurodegeneration and symptomatology than amyloid-β. Identifying which protein is driving neurodegeneration and/or cognitive decline is essential for development of disease-modifying agents.