MEDIAL TEMPORAL ATROPHY AND ITS ASSOCIATION WITH AMYLOID STATUS AND NEUROPATHOLOGY

Abstract

Converging evidence suggests the neurodegenerative process underlying Alzheimer's disease (AD) begins years before the onset of dementia symptoms. Quantitative neuroimaging has been shown to be a useful in vivo biomarker of AD pathology. There is evidence that neurodegeneration—as measured by atrophy in MRI—correlates with clinical impairment and progression from Mild Cognitive Impairment (MCI) to dementia better than amyloid plaque burden. Medial temporal lobe atrophy, in particular, is predictive of rapid conversion from MCI to dementia (Heister et al., 2011). Here we examined the relationship between a quantitative neuroimaging measure of hippocampal atrophy and amyloid pathology—assessed through PET imaging and at autopsy—to evaluate its ability to predict underlying pathology and potential utility in identifying subjects with the greatest likelihood for decline. Two cohorts with high-resolution structural MRI including 989 (AD=171, MCI=494, NL=324) subjects from ADNI with Florbetapir (AV-45) PET and 130 subjects from 6 ADCs in the NACC (U01 AG016976) with neuropathology data were assembled to examine the relationship between hippocampal occupancy (HOC—a measure of hippocampal atrophy) and amyloid pathology. Volumetric MRI analysis was used to derive volumes of cortical regions and subcortical structures for each individual, and age- and sex-adjusted normative percentiles of HOC (HOCP) were generated (Brewer, 2009). The relationship between HOCP and amyloid positivity was examined by applying various percentile cutoffs. 47% of non-demented individuals in ADNI were amyloid positive. 69% of those who converted to dementia had an HOCP≤15. In all non-demented individuals, 32% of those with an HOCP≤15 converted to dementia, versus 2% of those with an HOCP>85. In the NACC cohort, 87% of subjects with an HOCP≤15 were clinically diagnosed with dementia while alive, versus 29% of those subjects with an HOCP>15. HOCP was associated with the amount of AD pathology at autopsy when controlling for the time between the MRI and death (p=0.0489). Even in subjects with dementia, HOCP was found to be significantly lower in amyloid positive subjects than amyloid negative subjects (p=0.0398). HOCP may be used as a predictor of decline and underlying AD pathology, and has potential use in clinical trial enrichment.