Cross-Sectional and Longitudinal Diagnostic Performance of Plasma p-tau217 and p-tau217/Aβ42 in Alzheimer's Disease.
OBJECTIVE Recent studies suggest that combining plasma phosphorylated tau (p-tau) with β-amyloid (Aβ) may improve diagnosis accuracy for Alzheimer's disease (AD). However, the cross-sectional and longitudinal concordance of these markers with Aβ positron emission tomography (PET) positivity remains incompletely understood. This study aimed to evaluate the diagnostic performance of plasma p-tau, alone and in combination with plasma Aβ, in AD. METHODS We included 326 participants from the Alzheimer's Disease Neuroimaging Initiative and 357 Chinese older adults from the Greater-Bay-Area Healthy Aging Brain Study who underwent Aβ-PET imaging. Longitudinal data were available for 285 Alzheimer's Disease Neuroimaging Initiative participants. Plasma p-tau181, p-tau217, Aβ42, and Aβ40 were measured on different analytical platforms. Diagnostic performance for Aβ-PET positivity was assessed using a two-cutoff approach. RESULTS Combining plasma p-tau with Aβ42 or the Aβ42/40 ratio reduced the intermediate zone. Notably, p-tau217/Aβ42 showed stronger agreement with Aβ-PET positivity than p-tau217 alone. Among individuals classified as p-tau217/Aβ42 positive but p-tau217 intermediate, 57.1 to 83.3% were Aβ-PET positive. Longitudinally, most Stable Positive (88.0-96.1%) and Stable Negative (89.4-90.9%) cases defined by p-tau217 or p-tau217/Aβ42 were Aβ-PET positive and Aβ-PET negative, respectively. Critically, 69.7 to 75.8% of Non-positive to Positive cases defined by p-tau217/Aβ42 were Aβ-PET positive. INTERPRETATION These findings provide novel insights into the cross-sectional and longitudinal diagnostic performance of plasma p-tau217/Aβ42 in AD. To be specific, plasma p-tau217/Aβ42 can reduce the intermediate zone and improve agreement with Aβ-PET positivity, and longitudinal p-tau217/Aβ42 monitoring is particularly informative for identifying Aβ-PET-positive patients who were p-tau217/Aβ42 negative or intermediate at baseline and were misclassified as low risk of AD. ANN NEUROL 2026.