Diagnostic performance of plasma pTau217/Aβ42 ratio and a three-zone threshold model for Alzheimer’s disease

Early and accurate diagnosis of Alzheimer's disease (AD) typically relies on invasive or expensive methods like cerebrospinal fluid (CSF) biomarkers and amyloid PET imaging. Blood-based biomarkers, particularly plasma phosphorylated tau (pTau₁₈₁, pTau₂₁₇) and amyloid-beta ratios (Aβ₄₂/₄₀), offer a more accessible diagnostic alternative. This study assessed the diagnostic accuracy of plasma biomarkers and developed a three-zone classification model to reduce reliance on invasive confirmatory tests. We retrospectively evaluated 109 participants referred to a tertiary memory clinic. Participants underwent cognitive assessments, brain MRI, CSF biomarker analyses (pTau₁₈₁, Aβ₄₂/₄₀), and plasma biomarker measurements (pTau₁₈₁, pTau₂₁₇, Aβ₄₂/₄₀, pTau₂₁₇/Aβ₄₂ ratio). Diagnostic performance was evaluated using ROC analyses, and thresholds achieving ≥ 95 % sensitivity and specificity were used to define low, intermediate and high-risk zones. Plasma biomarkers correlated significantly with CSF biomarkers. For identifying AD pathology (A+/T + vs. others), plasma pTau₂₁₇ and the pTau₂₁₇/Aβ₄₂ ratio demonstrated the highest accuracy (both AUC=0.95), outperforming plasma pTau₁₈₁ (AUC=0.88) and Aβ₄₂/₄₀ ratio (AUC=0.73). At optimal thresholds, plasma pTau₂₁₇ showed 87.5 % sensitivity and 93.4 % specificity, whereas the pTau₂₁₇/Aβ₄₂ ratio showed higher sensitivity (95.8 %) but lower specificity (85.2 %). Using the three-zone model, plasma pTau₂₁₇ enabled definitive classification in 80.7 % of patients, increasing to 84.4 % with the pTau₂₁₇/Aβ₄₂ ratio. Among patients with mild cognitive impairment, plasma pTau₂₁₇ achieved excellent accuracy (AUC=0.98). Plasma pTau₂₁₇, alone or combined with Aβ₄₂, provides highly accurate and scalable identification of AD pathology, substantially reducing the need for invasive diagnostic procedures.