Background: Plasma p-tau217/Aβ42 has shown promising diagnostic accuracy for Alzheimer’s disease (AD), but its prognostic utility for AD risk and clinical phenotype remains unclear.
Methods: We analyzed 580 participants from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Diagnostic performance of p-tau217/Aβ42 for cognitively normal (CN) vs. AD was evaluated. Cox regression and linear mixed-effects models were used to investigate the relationships of plasma p-tau217/Aβ42 with AD risk, cognitive decline and brain atrophy. We also quantified the incremental predictive value of p-tau217/Aβ42 for AD risk, cognitive decline and brain atrophy beyond demographic data and the Mini-Mental State Examination (MMSE). Mediation analyses were conducted to test whether brain structure accounted for associations between p-tau217/Aβ42 and cognition.
Results: Plasma p-tau217/Aβ42 levels were elevated in individuals with amyloid and tau pathology. Combining plasma p-tau217/Aβ42 with MMSE discriminated AD from CN individuals (AUC = 0.988). Higher p-tau217/Aβ42 predicted faster cognitive decline and greater hippocampal atrophy (P < 0.001), increased explained variance for cognitive decline and brain atrophy (ΔmR² = 0.016–0.121). Hippocampal volumes partially mediated the relationships of p-tau217/Aβ42 with cognition (5.3–14.8%). In non-demented participants, the high-risk group of p-tau217/Aβ42 showed higher AD risk (HR = 7.45, 95% CI 4.36–12.74) and faster cognitive decline, remaining significant among those with preserved baseline cognition (MMSE > 28). The addition of p-tau217/Aβ42 improved AD risk prediction beyond MMSE and covariates (ΔC = 0.040), increased time-dependent AUC at 2 years (ΔAUC = 0.030), reduced prediction error (ΔBrier = − 0.017), improved reclassification (NRI/IDI; all P < 0.05).
Conclusion: Plasma p-tau217/Aβ42 is a strong prognostic biomarker for AD progression, adding value beyond demographics and brief cognitive screening. It supports a practical blood-first approach for risk stratification and prioritizing confirmatory testing, with additional utility for diagnostic support and trial enrichment.
Supplementary Information: The online version contains supplementary material available at 10.1186/s12967-026-07939-z.
Keywords: Alzheimer’s disease; Biomarker; Brain structure; Cognitive assessment; Plasma p-tau217/Aβ42; Predictive performance.