Aim: PET amyloid is considered a surrogate endpoint for clinical progression of Alzheimer's disease but has yet to be evaluated systematically using an established framework despite underpinning drug approval decisions. This study determined the surrogacy relationship between amyloid PET and clinical dementia rating sum-of-boxes (CDR-SB).
Methods: We identified all randomised, placebo-controlled trials of biologics developed to halt/reverse the progression of cognitive decline in Alzheimer’s disease that included amyloid PET and CDR-SB as primary/secondary outcomes (to 14/10/2024). Meta-regression modelled relative cognitive decline as a function of amyloid PET standard-uptake value ratio (SUVR). Surrogacy metrics included the R2trial and the surrogate threshold effect (STE) which is the minimum PET amyloid between trial difference required to predict clinical benefit with 95% confidence. The Biomarker-Surrogacy Evaluation Schema (BSES) (1) evaluated the status of amyloid PET as a surrogate endpoint.
Results: Fifteen qualifying trials represented 7 pharmacological drug classes: aducanumab, bapineuzumab, crenezumab, donanemab, gantenerumab, lecanemab and solanezumab. Subunit analyses of prespecified risk status and/or dosing increased the data set to 27 trial-level data sets. Median trial size was 783, trial duration 78 weeks, mean age 71 years, 56% female and 84% white. There was no STE but it approached -0.5 PET amyloid SUVR between trial reduction difference. R2trial was 0.49. Amyloid PET was good/excellent on BSES Study Design, Target Outcome, Generalisability domains but was poor/fair on Statistical Evaluation domain.
Conclusion: We provide the first systematic meta-regression evaluation of amyloid PET as a surrogate endpoint for attenuated cognitive decline for patients with Alzheimer’s disease.