Identification of patients receiving amyloid-targeting therapies in observational studies using amyloid PET trajectories: Insights from LEADS.

INTRODUCTION

As amyloid-targeting therapies (ATTs) enter clinical care, observational cohorts must accurately ascertain ATT exposure. We developed an approach to flag mis/undocumented ATT in the Longitudinal Early-Onset Alzheimer's Disease Study using longitudinal amyloid positron emission tomography (PET).

METHODS

We analyzed 742 [18F]florbetaben PET from 270 participants with early-onset Alzheimer's disease. Using PET acquired before US Food and Drug Administration (FDA) approval of ATTs, we quantified Centiloid (CL) variability between two consecutive scans (PET "segments") to determine unusual CL decline thresholds that were then applied to post-FDA approval segments.

RESULTS

Pre-FDA approval segments increased by a median of 4.1 CL/year (whole cerebellum reference) and 3.6 CL/year (composite reference); unusual CL decline thresholds, defined as mean - 2 standard deviations of the pre-approval distributions were -15.8 and -9.4 CL/year, respectively. When applied to segments acquired post-FDA approval, extreme declines were observed in 59% to 78% of treated (n = 54) versus 4% to 8% of untreated segments (n = 344).

DISCUSSION: Longitudinal amyloid PET analyses can help identify ATT exposure in observational studies.