Comparative performance of plasma pTau181/Aβ42, pTau217/Aβ42 ratios, and individual measurements in detecting brain amyloidosis

Abstract

Background: Early detection of brain amyloidosis (Aβ+) is crucial for diagnosing Alzheimer' disease (AD) and optimizing patient management, especially in light of emerging treatments. While plasma biomarkers are promising, their combined diagnostic value through specific ratios remains underexplored. In this study, we assess the diagnostic accuracy of plasma pTau isoform (pTau181 and pTau217) to Aβ42 ratios in detecting Aβ+ status.

Methods: This study included 423 participants from the multicenter prospective ALZAN cohort, recruited for cognitive complaints. Aβ+ status was determined using cerebrospinal fluid (CSF) biomarkers. The confirmatory cohort comprises 1176 patient samples from the Alzheimer's Disease Neuroimaging Initiative (ADNI), with Aβ+ status determined by positron emission tomography (PET) imaging. Plasma biomarkers (pTau181, pTau217, Aβ40, Aβ42) were measured using immunoassays and mass spectrometry, with specific ratios calculated. In the ALZAN cohort, the impact of confounding factors such as age, renal function, ApoE 4 status, body mass index, and the delay between blood collection and processing was also evaluated to assess their influence on biomarker concentrations and diagnostic performance. The primary outcome was the diagnostic performance of plasma biomarkers and their ratios for detecting Aβ+ status. Secondary outcomes in the ALZAN cohort included the proportion of patients classified as low, intermediate, or high risk for Aβ+ using a two-cutoff approach.

Findings: In ALZAN the pTau181/Aβ42 ratio matched the diagnostic performance of pTau217 (AUC of 0.911 (0.882-0.940) vs. 0.909 (0.879-0.939), P = 0.85). The pTau217/Aβ42 ratio demonstrated the highest diagnostic accuracy, with an AUC of 0.927 (0.900-0.954). Both ratios effectively mitigated confounding factors, such as variations in renal function, and were also efficient in identifying Aβ+ status in individuals with early cognitive decline. Diagnostic accuracy of ratios vs. individual measurement was confirmed in the ADNI cohort. In ALZAN, using two-cutoff workflows with pTau217/Aβ42 instead of pTau217 alone reduced the intermediate-risk zone from ∼16% to ∼8%, enhancing stratification for clinical decision-making.

Interpretation: The pTau217/Aβ42 ratio demonstrated improved diagnostic performance for detecting Aβ+ compared to individual biomarkers, potentially reducing diagnostic uncertainty. These findings suggest that plasma biomarker ratios could be useful; however, further validation in independent and diverse clinical settings is necessary before broader clinical implementation.

Funding: Fondation Research Alzheimer (ALZAN projet), AXA Mécénat Santé (INTERVAL Project), Fondation pour la Recherche Médicale (FRM, team Proteinopathies).

Keywords: Alzheimer; Biomarkers; Blood; Diagnosis.

Fig. 1

ROC curves of plasma biomarkers according to amyloid status in ALZAN and ADNI. Receiver operating characteristic curve curves for Aβ+ detection in ALZAN (a–c) and ADNI (d–f) using Aβ40, Aβ42, and Aβ42/40 in combination with pTau181 (a) or pTau217 (b, d, e). The corresponding areas under curve (AUC) with 95% confidence intervals are represented in (c) and (f). In ADNI, biomarkers were measured using either an immunoassay approach (Fuji) (d) or with mass spectrometry (C2N) (e). AUC comparisons are reported in Supplementary Tables S4 and S5.

Fig. 2

Association of plasma biomarkers and their ratios with ALZAN cohort characteristics. Forest plots of associations in the ALZAN cohort with age, BMI, blood delay and eGFR, using linear regression, for: (a) Aβ40, (b) Aβ42, (c) Aβ42/40, (d) pTau181, (e) pTau181/(Aβ42/40), (f) pTau181/Aβ42, (g) pTau217, (h) pTau217/(Aβ42/40) and (i) pTau217/Aβ42. Regression beta coefficients with 95% confidence intervals are illustrated. Abbreviations: BMI, body mass index; eGFR, estimated glomerular filtration rate.

Fig. 3

Distribution of the ALZAN population between high-, intermediate-, and low-risk brain Aβ+ groups. Histograms of the ALZAN population, distributed among high-, intermediate-, and low-risk brain amyloidosis groups, are shown as defined by the cut points of pTau181, pTau217, pTau181/Aβ42, and pTau217/Aβ42 (see Table 2). Cutoffs were selected to achieve over 90% specificity/sensitivity (panel a) or over 90% predictive value (with Aβ+ prevalence of 57.5%) (panel b). Panel c: ROC curves for Aβ+ detection using pTau217 (blue) and pTau217/Aβ42 (cyan) were plotted. Respective brackets link the two cut-points used to define low and high predictive risk for Aβ+ using 90% specificity/sensitivity (blue brackets) or 90% predictive value (with Aβ+ prevalence of 57.5%) (green brackets). Note that the brackets are smaller on the pTau217/Aβ42 curve. Abbreviation: ROC, receiver operating characteristic curve.