Methodology dependent sex differences in Aβ‐PET measured with SUVR

Abstract Background Determining Aβ‐PET status is crucial for Alzheimer’s disease trials. Standard uptake value ratio (SUVR) using a reference region is a common semi‐quantitative technique. Sex differences in regional blood flow and white matter (WM) could impact SUVR differentially depending on the reference region. It is important to understand how methodological factors can influence SUVR derived Aβ status. Method Individuals from Insight 46 (1946 British birth cohort) underwent PET/MR scanning with [18F]florbetapir (N = 425; mean age (SD) 70.6 (0.7); 49% female). Regions derived from T1‐weighted MRI with geodesic information flows (GIF) were resampled to PET space. SUVR was calculated (50‐60 mins post‐injection) using a cortical composite target normalised to whole cerebellum (WC) or eroded WM, with/without partial volume correction (PVC). Additionally, SUV was calculated normalising to injected dose and weight. Distribution volume ratio (DVR) from Logan graphical analysis (cerebellar grey matter reference) was calculated for N = 391 (51% female). Linear regression was used to investigate differences in Aβ‐PET measures by sex adjusting for region volumes. Aβ status was defined with Gaussian‐mixture modelling. Sex differences in SUVR and DVR were investigated in individuals rated concordantly Aβ‐ with all four SUVR methods. SUVR Aβ status discordance was examined in relation to sex and APOE e4 genotype. Result Females had significantly higher SUVR with a WC reference; and there were no sex differences with a WM reference or with dynamic DVR (Figure 1). Males had higher SUV in all regions and the difference was greater in the WC compared to WM (Figure 2). PVC decreased the influence of volume on SUV in the WC but not WM. 87% of individuals were classified concordantly Aβ+/‐ on SUVR measures with equal proportion of males/females. More females were Aβ+ only with WC reference with PVC, whereas more males were Aβ‐ only on WC without PVC (Figure 3).