Converging evidence for a "gray‐zone" of amyloid burden and its relevance: The value of amyloid PET beyond dichotomization. (7th December 2020)

• Record Type: Journal Article
• Authors: Bullich, Santiago; Salvadó, Gemma; Alves, Isadora Lopes; Marquié, Marta; Stephens, Andrew W; Gispert, Juan Domingo; Molinuevo, Jose Luis; Buckley, Christopher; Boada, Mercè; Barkhof, Frederik
• Is Part Of: Alzheimer's & dementia. Volume 16(2020)Supplement 5
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-12-07
• Subjects: Alzheimer's disease -- Periodicals; Alzheimer Disease -- Periodicals; Dementia -- Periodicals; Démence; Maladie d'Alzheimer; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 616.83
• DOI: 10.1002/alz.044786
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: Traditional quantitative cut‐offs for amyloid PET positivity have been defined to discriminate Alzheimer's dementia (AD) subjects from elderly Aβ‐negative non‐demented controls. Currently, observational and interventional studies focus on earlier stages of Aβ deposition, where established cut‐offs might not be appropriate. Here, we review recent developments on early pathology identification, and provide supporting evidence from three cohorts for the establishment of a "gray‐zone" of amyloid burden. Method: Data from three cohorts of cognitively unimpaired individuals were included, namely ALFA+ (n=357), FACEHBI (n=228) and the AMYPAD Prognostic Study (n=122). PET scans ([ 18 F]flutemetamol and [ 18 F]florbetaben) were analyzed in Centiloid (CL) units. Gaussian Mixture Modelling was used to fit two/three Gaussians to the global CL values, and the cut‐off for early pathology detection was defined as two standard deviations above the mean of the left Gaussian. Then, the data was merged to derived a joint cut‐off. Finally, these data‐driven cut‐offs were compared to previously established thresholds in the context defining a range of CL values where pathology is emerging. Result: GMM identified a cut‐off of CL=11 for FACEHBI, CL=14 for ALFA+, CL=17 for AMYPAD and CL=11 for the cohorts combined (Figure 1). Similarly, Salvadó et al identified two cut‐offs based on a direct comparison with established CSFAβ42 thresholds: CL=12 to rule out‐amyloid pathology and CL=29 to denote established pathology. With a different approach, the FACEHBI working group finds CL=13 as an early threshold defined based on young healthy controls, and CL=36 as an established cut‐off to separate healthy elderly from Aβ‐positive AD‐dementia subjects. In AMYPAD, where a more comprehensive sampling of the AD continuum is present, fitting 3 Gaussians through the data allows the identification of two data‐driven cut‐offs: CL=13 and CL=29. Conclusion: A number of recent and current reports converge to the utility of two cut‐offs for amyloid PET abnormality, an early cut‐off around CL=11‐17 where pathology may be emerging, and a second around CL=29‐36 where amyloid burden levels greatly correspond to neuropathology findings. Together, these create a gray‐zone of CL values pre‐AD dementia levels of amyloid burden, which can improve the detection of emerging pathology in observational and secondary prevention trials.