Diagnostic Performance of Plasma Phospho‐tau217 for Alzheimer's Disease in a Large Memory Clinic Cohort

Abstract Background Phospho‐tau217 ( p ‐tau217) has emerged as a leading blood‐based biomarker for Alzheimer's disease (AD) and has been recommended by an Alzheimer's Association Workgroup as a biomarker for AD pathology. However, its translational performance in clinical settings remains to be explored. The objective of this study was to evaluate the concordance between p ‐tau217 predicted AD pathology and the clinical diagnosis of AD in a large memory clinic cohort to address this question. Methods Participants at the University of Pittsburgh Alzheimer's Disease Research Center (Pitt‐ADRC) provided blood samples and underwent ADRC classification scheme‐based clinical diagnosis. APOE genotyping was conducted using TaqMan assays. Plasma p ‐tau217 levels were quantified using the Quanterix ALZpath Single Molecule Array (Simoa) assay. A sub‐cohort with amyloid PET data was used to establish the p ‐tau217 cutpoints. Both a single‐cutpoint approach based on the Youden index and a two‐cutpoint approach, with low and high cutpoints determined by 99% sensitivity and 95% specificity, respectively, were used to evaluate the agreement between p ‐tau217 and clinical diagnosis status. Logistic regression was used to statistically test the interaction between p ‐tau217 and demographic and genetic covariates. Results This study included 3,788 participants (59.6% female; 47.8% APOE ε4 carriers), of whom 2,014 had a clinical diagnosis of AD. Using a single‐cutoff approach with a threshold of 0.547 pg/mL, p ‐tau217 demonstrated 84.3% sensitivity, 76.0% specificity, and an overall accuracy of 82.0% in interpreting AD status. The two‐cutpoint approach, with low and high cutpoints of 0.321 and 0.935, improved the diagnostic accuracy to 88.0%, with sensitivity and specificity of 89.9% and 91.7%, respectively, and 1360 (35.9%) in the intermediate zone. With the single‐cutpoint approach, individuals with high p ‐tau217 were associated with an odds ratio of 17.1 of having AD compared to those with low p ‐tau217. Sex, but not race, and APOE ε4 carrier status significantly moderated the association between p ‐tau217 status and clinical diagnosis, with women having an odds ratio of 20.9 compared to men of 12.1 ( p = 0.010). Conclusion Our results demonstrated that using p ‐tau217 as a surrogate biomarker for AD pathology shows strong concordance with clinical diagnosis and can be a promising tool for enhancing clinical decision‐making.