Advertisement

THRESHOLDS FOR BRAIN AMYLOID CONCENTRATION IN RELATION TO DISEASE STAGE, COGNITION, BRAIN METABOLISM, BRAIN ATROPHY, AND APOE GENOTYPE IN THE ADNI COHORT

      Background: To determine vulnerability of cognitive, functional, cerebral metabolic rate of glucose (CMRglu) and brain volume variables to fibrillar amyloid burden, in relation to APOE genotype and disease stage in the ADNI Cohort.
      Methods: Data used in this study were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). Baseline demographic, clinical, neuropsychological, and volumetric MRI data for 385 subjects [55 diagnosed with mild Alzheimer's disease (AD), 91 with late MCI (LMCI), 114 early MCI (EMCI), and 125 cognitively normal (CN) cases] were explored in relation to APOE genotype (ApoE4+ (n=136) versus ApoE4- (n=211)). Piecewise Linear Models were used to assess the concurrent relationships between mean brain standardized uptake value ratios (SUVRs) of [18F]-florbetapir (AV45) uptake, auditory verbal learning test (AVLT), CDR sum of boxes (CDRsb) scores, regional CMRglu values and brain volumes on MRI, as a function of e4 status. Inflection points for SUVR scores in relationship to each of these regression curves were identified by computing points on the linear models at which the second derivative changes sign.
      Results: Overall, the association of SUVR scores was stronger for ApoE4+ than ApoE4- subjects (p<0.01). The coefficients of determination were: entorhinal cortex volume (R 2 for ApoE4+ = 18.14% and ApoE4- = 9.91%), mean CMRglu (R 2 for ApoE4+ = 13.89% and ApoE4- = 13.38%), RAVLT (R 2 for ApoE4+ = 9.36% and ApoE4- = 5.79%), and Hippocampal volume (R 2 for ApoE4+ = 16.71% and ApoE4- = 7.06%). Inflection points were identifiable for all the variables (Figure 1), and were at higher SUVR scores in ApoE4+ (μ =1.298, ƒ = 0.13) than ApoE4- (μ =1.072, ƒ = 0.08) subjects (p<0.001). The SUVR at the inflection points among cognitive (μ =1.270, ƒ = 0.13), atrophy (μ =1.230, ƒ = 0.16) and CMRglu (μ =1.201, ƒ = 0.20) variables were different.
      Conclusions: For a given level of neurodegeneration (as reflected in cognitive scores, CMRglu values and regional brain atrophy in AD affected regions), the corresponding AV45 SUVRs and inflection points are strongly influenced by factors such as APOE genotype. These findings indicate the absence of a fixed threshold for “positive” or “negative” amyloid PET s cans.
      Figure thumbnail fx1
      Figure 1Fibrillar amyloid burden mean AV45 SUVR versus (A) Cerebral Metabolic Rate of glucose (CMRglu), (B) Rey Auditory Verbal Learning Test (RAVLT), (C) Normalized Hippocampus Volume and (D) Normalized Entorhinal Cortex Volume showing variations in mean AV45 SUVR for the inflection points of among cognitive, volumetric and functional measures. The inflection points are listed below each corresponding graph.