IC‐P‐105: Uncovering The Relationship Between β‐Amyloid and Glucose Metabolism in Mild Cognitive Impairment

Recent PET studies have interrogated the relationship between β-amyloid load, glucose metabolism, and apolipoprotein E ε4 (APOE ε4) genotype. It has been reported that APOE ε4, and not aggregated fibrillar β-amyloid, contributes to glucose hypometabolism in cognitively normal and MCI subjects. The major limitation of these studies is the representation of the overall β-amyloid burden by a single measurement taken from the mean SUVR from a composite region-of-interest (ROI) or dichotomization into low or high β-amyloid burden. We utilized a Singular Value Decomposition (SVD) approach to reveal patterns of cross-correlation structure between glucose metabolism, as measured by [18F]FDG PET, and β-amyloid, as measured by [18F]florbetapir PET, in 274 MCI subjects from the ADNI study. SVD yields a set of eigenimages and individual subject loadings (i.e. components) corresponding to both β-amyloid and glucose metabolism. The β-amyloid subject loadings for the first component represent the β-amyloid burden maximally related to metabolism. To identify regions where metabolism is statistically related to β-amyloid burden, we regressed β-amyloid subject loadings against FDG in a general linear model (GLM) that included age, gender, and APOE ε4 status as covariates. The first SVD component accounted for 86% of the total variability explained by the cross-correlation between β-amyloid and glucose metabolism. The stronger weights in the first β-amyloid eigenimage (Figure 1A) correspond to the medial prefrontal and posterior cingulate cortices, and inferior temporal and fusiform gyri. The highest negative values of first metabolic eigenimage (Figure 1B) are spatially located in regions that GLM identified as having significant negative correlations, which are not explained by the APOE ε4 effect, between the β-amyloid loadings and FDG, particularly in the angular gyrus and posterior cingulate cortex (Figure 2A). For comparison, no significant regions were observed using a whole cortex average of β-amyloid burden (Figure 2B). Multivariate, cross-correlation analyses can uncover complex brain patterns not found with univariate statistical analysis approaches. These results support the notion that it is the spatially distributed, rather than focal, accumulation of β-amyloid that is associated with metabolic dysfunction. Future work will expand this analysis to identify the pattern of β-amyloid maximally related to metabolic connectivity. Surface projections for β-amyloid (A) and glucose metabolism (B) eigenimages corresponding to first component in the SVD analysis. Negative values in the metabolic eigenimage (B) indicate a negative correlation with the β-amyloid subject loadings. The β-amyloid subject loadings are based on the spatial weights of the β-amyloid eigenimage (A). The highest values in the β-amyloid eigenimage correspond to regions where β-amyloid is maximally related to glucose metabolism. Surface projections of β-amyloid-glucose metabolism regression models. (A) Significant regression results for SVD-derived β-amyloid burden versus glucose metabolism. The significant regions include areas within the angular gyrus and the posterior cingulate cortex. Results are FDR-corrected (q<0.05). (B) Regression results for mean β-amyloid burden using a whole cortex ROI. No significant relationships were observed.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.