Abstract
INTRODUCTION: Plasma biomarkers are promising tools for Alzheimer's disease (AD) diagnosis, but comparisons with more established biomarkers are needed. METHODS: We assessed the diagnostic performance of p-tau(181) , p-tau(217) , and p-tau(231) in plasma and CSF in 174 individuals evaluated by dementia specialists and assessed with amyloid-PET and tau-PET. Receiver operating characteristic (ROC) analyses assessed the performance of plasma and CSF biomarkers to identify amyloid-PET and tau-PET positivity. RESULTS: Plasma p-tau biomarkers had lower dynamic ranges and effect sizes compared to CSF p-tau. Plasma p-tau(181) (AUC = 76%) and p-tau(231) (AUC = 82%) assessments performed inferior to CSF p-tau(181) (AUC = 87%) and p-tau(231) (AUC = 95%) for amyloid-PET positivity. However, plasma p-tau(217) (AUC = 91%) had diagnostic performance indistinguishable from CSF (AUC = 94%) for amyloid-PET positivity. DISCUSSION: Plasma and CSF p-tau(217) had equivalent diagnostic performance for biomarker-defined AD. Our results suggest that plasma p-tau(217) may help reduce the need for invasive lumbar punctures without compromising accuracy in the identification of AD. HIGHLIGHTS: p-tau(217) in plasma performed equivalent to p-tau(217) in CSF for the diagnosis of AD, suggesting the increased accessibility of plasma p-tau(217) is not offset by lower accuracy. p-tau biomarkers in plasma had lower mean fold-changes between amyloid-PET negative and positive groups than p-tau biomarkers in CSF. CSF p-tau biomarkers had greater effect sizes than plasma p-tau biomarkers when differentiating between amyloid-PET positive and negative groups. Plasma p-tau(181) and plasma p-tau(231) performed worse than p-tau(181) and p-tau(231) in CSF for AD diagnosis.