A comprehensive head-to-head comparison of key plasma phosphorylated tau 217 biomarker tests

Abstract

Plasma phosphorylated-tau 217 (p-tau217) is currently the most promising biomarker for reliable detection of Alzheimer's disease pathology. Various p-tau217 assays have been developed, but their relative performance is unclear. We compared key plasma p-tau217 tests using cross-sectional and longitudinal measures of amyloid-β (Aβ)-PET, tau-PET and cognition as outcomes and benchmarked them against CSF biomarker tests. Samples from 998 individuals [mean (range) age 68.5 (20.0-92.5) years, 53% female] from the Swedish BioFINDER-2 cohort, including both cognitively unimpaired and cognitively impaired individuals, were analysed. Plasma p-tau217 was measured with mass spectrometry assays [the ratio between phosphorylated and non-phosphorylated (%p-tau217WashU) and p-tau217WashU] and with immunoassays (p-tau217Lilly, p-tau217Janssen and p-tau217ALZpath). CSF biomarkers included p-tau217Lilly, the US Food and Drug Administration-approved p-tau181/Aβ42Elecsys, and p-tau181Elecsys. All plasma p-tau217 tests exhibited a high ability to detect abnormal Aβ-PET [area under the curve (AUC) range: 0.91-0.96] and tau-PET (AUC range: 0.94-0.97). Plasma %p-tau217WashU had the highest performance, with significantly higher AUCs than all the immunoassays (Pdiff < 0.007). For detecting Aβ-PET status, %p-tau217WashU had an accuracy of 0.93 (immunoassays: 0.83-0.88), sensitivity of 0.91 (immunoassays: 0.84-0.87) and a specificity of 0.94 (immunoassays: 0.85-0.89). Among immunoassays, p-tau217Lilly and plasma p-tau217ALZpath had higher AUCs than plasma p-tau217Janssen for Aβ-PET status (Pdiff < 0.006), and p-tau217Lilly outperformed plasma p-tau217ALZpath for tau-PET status (Pdiff = 0.025). Plasma %p-tau217WashU exhibited stronger associations with all PET load outcomes compared with immunoassays; baseline Aβ-PET load (R2: 0.72; immunoassays: 0.47-0.58; Pdiff < 0.001), baseline tau-PET load (R2: 0.51; immunoassays: 0.38-0.45; Pdiff < 0.001), longitudinal Aβ-PET load (R2: 0.53; immunoassays: 0.31-0.38; Pdiff < 0.001) and longitudinal tau-PET load (R2: 0.50; immunoassays: 0.35-0.43; Pdiff < 0.014). Among immunoassays, plasma p-tau217Lilly was more associated with Aβ-PET load than plasma p-tau217Janssen (Pdiff < 0.020) and with tau-PET load than both plasma p-tau217Janssen and plasma p-tau217ALZpath (all Pdiff < 0.010). Plasma %p-tau217 also correlated more strongly with baseline cognition (Mini-Mental State Examination) than all immunoassays (R2: %p-tau217WashU: 0.33; immunoassays: 0.27-0.30; Pdiff < 0.024). The main results were replicated in an external cohort from Washington University in St Louis (n = 219). Finally, p-tau217NULISA showed similar performance to other immunoassays in subsets of both cohorts. In summary, both mass spectrometry- and immunoassay-based p-tau217 tests generally perform well in identifying Aβ-PET, tau-PET and cognitive abnormalities, but %p-tau217WashU performed significantly better than all the examined immunoassays. Plasma %p-tau217 may be considered as a stand-alone confirmatory test for Alzheimer's disease pathology, whereas some immunoassays might be better suited as triage tests where positive results are confirmed with a second test, which needs to be determined by future reviews incorporating results from multiple cohorts.

Keywords: Alzheimer’s disease; CSF biomarkers; p-tau217; plasma biomarkers.

Figure 1

Correlations between plasma p-tau217 biomarkers. The reported betas are standardized and across the full sample. Z-scores were based on cognitively unimpaired, CSF Aβ− participants (n = 364) as the reference group. The grey dots represent CSF Aβ− participants, the red dots represent CSF Aβ+ participants. Aβ = amyloid-β.

Figure 2

Comparisons of area under the curve and accuracy between plasma p-tau217 biomarkers for Aβ-PET and tau-PET status. (A) AUC comparisons for Aβ-PET status; (B) AUC comparisons for tau-PET status; (C) accuracy comparisons for Aβ-PET status; and (D) accuracy comparisons for tau-PET status. Dots and squares represent the AUC or accuracy, and bars represent the 95% confidence interval. The dashed line is drawn at CSF p-tau181/Aβ42_Elecsys to facilitate comparison of the other tests with the current US Food and Drug Administration-approved test. Significant differences between assays were assessed through bootstrapping, and all P-values were false discovery rate corrected. Models were corrected for age and sex. Aβ = amyloid-β.

Figure 3

Quantile grouping for Aβ-PET and tau-PET. Boxes show the interquartile range, and the horizontal lines represent the medians. Negative participants were defined as falling below the predefined cut-offs (Aβ-PET: <1.033; tau-PET: <1.362). Neighbouring quantiles were compared with Wilcoxon signed-rank tests. *P < 0.05. **P < 0.01. ***P < 0.001. Aβ = amyloid-β.

Figure 4

R ² comparisons with Aβ-PET and tau-PET load as outcome. (A) R² comparisons for associations with baseline Aβ-PET load; (B) R² comparisons for associations with baseline tau-PET load; (C) R² comparisons for associations with the rate of change in Aβ-PET load; and (D) R² comparisons for associations with the rate of change in tau-PET load. Dots and squares represent R², and bars represent the 95% confidence interval (95% CI). The dashed line is drawn at CSF p-tau181/Aβ42_Elecsys, to facilitate comparison of the other tests with the current US Food and Drug Administration-approved test. Significant differences between assays were assessed through bootstrapping, and all P-values were false discovery rate corrected. Cross-sectional: PET SUVR ∼ biomarker + age + sex. Longitudinal: individual rate of change in PET SUVR ∼ biomarker + age + sex. Aβ = amyloid-β; SUVR = standardized uptake value ratio.

Figure 5

R² comparisons with cognition as outcome. (A) R² comparisons for associations with baseline MMSE scores; (B) R² comparisons for associations with baseline mPACC performance; (C) R² comparisons for associations with the rate of change in MMSE scores; and (D) R² comparisons for associations with the rate of change in mPACC performance. Dots and squares represent R², and bars represent the 95% confidence interval (95% CI). The dashed line is drawn at CSF p-tau181/Aβ42_Elecsys, to facilitate comparison of the other tests with the current US Food and Drug Administration-approved test. Significant differences between assays were assessed through bootstrapping, and all P-values were false discovery rate corrected. For the MMSE models, participants with non-AD dementia were excluded. For the mPACC models, participants with dementia were excluded. Cross-sectional: cognition ∼ biomarker + age + sex + education. Longitudinal: individual rate of change in cognition ∼ biomarker + age + sex + education. Aβ = amyloid-β; MMSE = Mini-Mental State Examination; mPACC = modified Preclinical Alzheimer Cognitive Composite.

Figure 6

Replication in the Knight ADRC cohort. Dots and squares represent the area under the curve, accuracy or R², and bars represent the 95% confidence interval (95% CI). The dashed line is drawn at CSF p-tau181/Aβ42_Lumipulse, to facilitate comparison of the other tests with the current approved US Food and Drug Administration-approved test. The global cognitive composite was a composite of several cognitive tests, z-scored with cognitively unimpaired Aβ− individuals as reference group. Significant differences between assays were assessed through bootstrapping, and all P-values were false discovery rate corrected. Linear model Aβ-PET: Aβ-PET ∼ biomarker + age + sex. Linear model global cognitive composite: cognition ∼ biomarker + age + sex + education. Aβ = amyloid-β.