Abstract
1576
Introduction: A quantitative, pathology-proven 18F-Florbetaben (FBB) abnormality cutoff has previously been developed in end of life subjects to discriminate Alzheimer’s disease (AD) from elderly Aβ-negative non-demented controls (NDC). However, current observational research studies and interventional clinical trials focus on earlier stages of Aβ deposition. In these situations, the use of previously established thresholds may be limited. In this study, we developed and validated FBB cutoffs to detect early Aβ deposition and to identify established amyloid pathology.
Methods: An SUVR threshold to detect early Aβ deposition was derived from a sample of young healthy controls (YHC) (n=70, 27.6±5.1 y) as 2 standard deviations above the mean SUVR value, and a second SUVR cutoff was derived using receiver operating characteristics curve analysis comparing visually Aβ-negative NDC (n=62, 67.8±6.8 y) and visually Aβ-positive AD dementia patients (n=62, 70.6±8.0 y). The obtained thresholds were then tested in two longitudinal studies: (1) Individuals with subjective cognitive decline (SCD) from the Fundació ACE Healthy Brain Initiative (FACEHBI) study (n=166; 64.9±7.2 y (Rodríguez-Gómez et al. 2017)) and (2) mild cognitively impaired (MCI) subjects (n=45, 72.7 ± 6.5 y) (Ong et al. 2013). The FBB PET scans were acquired in multiple PET scanners, and then converted to a standard image resolution (Joshi et al. 2009). They were then analyzed using the standard centiloid sampling methodology and cortex SUVR (CSUVR) values were derived using the standard centiloid regions of interest, where the whole cerebellum is used as a reference region (Klunk et al. 2015; Rowe et al. 2017).
Results: The mean CSUVR of the YHC sample was 1.03±0.03 resulting in a cutoff for early detection of amyloid load of 1.10 (13.5 CL). The derived CSUVR cutoff to identify established amyloid pathology was 1.24 (35.7 CL). These CSUVR cutoffs allowed defining three levels in the CSUVR continuum: Aβ-negative subjects (<13.5 CL), early Aβ accumulation (“gray zone”) (13.5-35.7 CL) and “established Aβ deposition” level (>35.7 CL). The rate of amyloid accumulation was significantly different from zero (p<0.05) in SCD and MCI participants with CSUVR in the “gray zone” (2.1±2.1 %/year (SCD), 2.6±1.5 %/year (MCI)). The accumulation rate was also different from zero (p<0.05) in the “established Aβ pathology” level (3.4±3.2 %/year (SCD), 1.4±1.8 %/year (MCI)). The rate of amyloid accumulation was not significantly different from zero (p>0.05) in Aβ-negative subjects (-0.1±1.1 %/year (SCD), 0.1±1.6 %/year (MCI)).
Conclusions: An interval between 13.5 and 35.7 CL is optimal for the detection of early Aβ deposition and to identify subjects that are likely accumulating Aβ.