Title: Clinical Utility of Plasma P-tau 181 and 217 Biomarkers for Alzheimer's Disease Diagnosis in a Hong Kong Cohort

Authors: Dr. Yat Fung Shea MBBS(HK), FRCP(Lond),^a,* Shun Shun So MSc,^b Hok Pui Chan BSc,^b Dr. Ka Keung Yam MBBS(HK), FRCP(Lond), ^a Dr. Felix Chi Kin Wong MBBS(HK), FRCPA,^b Dr. Tsz King Wong MBBS(HK),^a Whitney Chin Tung Ip, MBBS(HK), ^a Krystal Xiwing Yau MSc,^a Zifei Zhang BSC,^{a, c} Mr. Pak Lam Chan MSc,^b Dr. King Pui Florence Chan MBBS(HK), MD,^a Dr. Yi Wah Eva Cheung PhD,^d Dr. Henry Ka Fung Mak MD, FRCR,^d Prof. Sau Man Mary Ip MD, FRCP(Lond),^a Dr. Patrick Ka Chun Chiu MBBS(HK), FRCP(Glas),^a Prof. Ching Wan Lam PhD FRCP(Glas), ^{b, e,*} Prof. Koon Ho Chan MD, PhD ^*

Affiliations:

^a Department of Medicine, Queen Mary Hospital, Li Ka Shing Faculty of Medicine, University of Hong Kong

^b Division of Chemical Pathology, Department of Pathology, Queen Mary Hospital

^c Neuroimmunology and Neuroinflammation Research Laboratory, Li Ka Shing Faculty of Medicine, University of Hong Kong

^d Department of Diagnostic Radiology, Queen Mary Hospital, Li Ka Shing Faculty of Medicine, University of Hong Kong

^e Department of Pathology, Li Ka Shing Faculty of Medicine, University of Hong Kong

^* Authors contributed equally to the manuscript

 

Abstract Background: Plasma phosphorylated tau (P-tau181 and P-tau217) have not been extensively studied for AD diagnosis in Chinese populations in real-world clinics. Methods: We enrolled 306 consecutive memory/neurology patients at Queen Mary Hospital, Hong Kong (May 2025–Jan 2026). In a validation cohort (n=133 with amyloid PET), we determined Roche Elecsys cut-offs for P-tau181 and P-tau217 and estimated post-test probabilities using a Bayesian ROC approach. The cut-offs were applied to an application cohort (n=174) without PET to assess clinical impact. Results: Using P-tau181 > 1.08 pg/mL and P-tau217 > 0.291 pg/mL, sensitivity/specificity were 88.6%/93.0% and 79.3%/79.0% (AUC 0.87 and 0.90). A dual-cutoff for P-tau217 (0.30–0.418 pg/mL) yielded 91.2% sensitivity and 90.8% specificity. Bayesian post-test probabilities favored P-tau217 in patients with stage ≥3 disease (AUC 0.91 vs 0.82). In the application cohort, applying the P-tau217 cut-off reclassified 34.8% of clinically diagnosed AD as non-AD and 50% of non-AD as AD. Conclusions: Plasma P-tau181 and P-tau217 are useful for diagnosing AD in Chinese patients seen in memory/neurology clinics, across diverse etiologies of cognitive impairment.

Keywords: aged; Alzheimer’s disease; dementia; tau Protein; memory