Skip to main content
NCBI home page
Alzheimer's & Dementia logoLink to Alzheimer's & Dementia
. 2023 Nov 6;20(2):1214–1224. doi: 10.1002/alz.13542

Plasma Aβ42/Aβ40 and phospho‐tau217 concentration ratios increase the accuracy of amyloid PET classification in preclinical Alzheimer's disease

Robert A Rissman 1,2,3,✉, Oliver Langford 2, Rema Raman 2, Michael C Donohue 2, Sara Abdel‐Latif 2, Matthew R Meyer 4, Traci Wente‐Roth 4, Kristopher M Kirmess 4, Jennifer Ngolab 2, Charisse N Winston 1,2, Gustavo Jimenez‐Maggiora 2, Michael S Rafii 2, Pallavi Sachdev 5, Tim West 4, Kevin E Yarasheski 4, Joel B Braunstein 4, Michael Irizarry 5, Keith A Johnson 6, Paul S Aisen 2, Reisa A Sperling 6; for the AHEAD 3‐45 Study team
PMCID: PMC10916957  NIHMSID: NIHMS1940743  PMID: 37932961

Abstract

INTRODUCTION

Incorporating blood‐based Alzheimer's disease biomarkers such as tau and amyloid beta (Aβ) into screening algorithms may improve screening efficiency.

METHODS

Plasma Aβ, phosphorylated tau (p‐tau)181, and p‐tau217 concentration levels from AHEAD 3–45 study participants were measured using mass spectrometry. Tau concentration ratios for each proteoform were calculated to normalize for inter‐individual differences. Receiver operating characteristic (ROC) curve analysis was performed for each biomarker against amyloid positivity, defined by > 20 Centiloids. Mixture of experts analysis assessed the value of including tau concentration ratios into the existing predictive algorithm for amyloid positron emission tomography status.

RESULTS

The area under the receiver operating curve (AUC) was 0.87 for Aβ42/Aβ40, 0.74 for phosphorylated variant p‐tau181 ratio (p‐tau181/np‐tau181), and 0.92 for phosphorylated variant p‐tau217 ratio (p‐tau217/np‐tau217). The Plasma Predicted Centiloid (PPC), a predictive model including p‐tau217/np‐tau217, Aβ42/Aβ40, age, and apolipoprotein E improved AUC to 0.95.

DISCUSSION

Including plasma p‐tau217/np‐tau217 along with Aβ42/Aβ40 in predictive algorithms may streamline screening preclinical individuals into anti‐amyloid clinical trials. ClinicalTrials.gov Identifier: NCT04468659

Highlights

  • The addition of plasma phosphorylated variant p‐tau217 ratio (p‐tau217/np‐tau217) significantly improved plasma biomarker algorithms for identifying preclinical amyloid positron emission tomography positivity.

  • Prediction performance at higher NAV Centiloid levels was improved with p‐tau217/np‐tau217.

  • All models generated for this study are incorporated into the Plasma Predicted Centiloid (PPC) app for public use.

Keywords: blood‐based biomarkers, mass spectrometry, plasma, tau

1. BACKGROUND

The most common form of dementia is Alzheimer's disease (AD), an age‐related and irreversible neurodegenerative disease. Within the United States, ≈ 6.5 million individuals are estimated to be living with AD, which is projected to increase to 7.2 million by 2025. 1 AD is characterized by increased brain amyloid beta (Aβ) and phosphorylated tau (p‐tau) that precede cognitive impairment. 2 Older, cognitively normal (CN) individuals with brain amyloidosis exhibit impaired performance on longitudinal neurophysiological and cognitive tests and possess a higher risk of progressing to mild cognitive impairment (MCI) and dementia. 3

AD therapeutics directed at slowing disease progression have the potential for greater efficacy when administered during the cognitively unimpaired stage, emphasizing the need for early detection. 4 Assays measuring both Aβ plaques and tau tangles through positron emission tomography (PET) and cerebrospinal fluid (CSF) biomarker detection allow the early detection of AD pathology prior to clinical symptoms. Elevated brain Aβ PET and lower CSF Aβ42/Aβ40 levels in cognitively unimpaired individuals are correlated with cognitive decline. 5 , 6 , 7 Studies measuring CSF tau phosphorylation suggest that specific hyperphosphorylation of the tau protein can further inform AD pathology and disease progression. 8 , 9 However, both the invasive nature and high investment of resources limits the broad implementation of current AD diagnostic procedures. 10 , 11 , 12 Blood biomarkers offer an attractive alternative as surrogates for brain Aβ and tau. The ease and familiarity of blood sample collection may additionally expand the detection and monitoring of AD progression to include under‐represented populations, including those in rural areas and diverse socioeconomic groups. 13 , 14

The detection of plasma Aβ has improved over the years with the implementation of mass spectrometry (MS). Multiple studies have demonstrated that plasma Aβ42/Aβ40 as measured by MS predicts brain amyloid status with a high degree of sensitivity and specificity, and moderately correlates with CSF Aβ42/Aβ40. 15 However, a qualified and validated plasma‐based MS assay for tau and p‐tau species is still lacking, due to the low plasma tau concentrations. 9 A recent study implemented MS to measure the various phosphorylated forms of plasma p‐tau, including p‐tau181, p‐tau217, and total tau across the stages of AD. 16 Although no relationship was observed for total tau, significant correlations were found between CSF and plasma p‐tau181 and p‐tau217 concentrations. While further supporting the use of p‐tau217 as an early marker for brain Aβ pathology and AD progression as indicated by previous immunoassay studies, 17 , 18 the small number of participants limited the scope of this initial study.

Other groups have previously demonstrated that plasma Aβ42/Aβ40 quantification by MS is a reliable biomarker for predicting elevated brain amyloid detected by PET. The predictive capability of plasma p‐tau217 along with Aβ42/Aβ40 has also been observed in two separate cognitively unimpaired cohorts, finding that p‐tau217 was highly indicative of MCI. 19 The current study samples originated from the AHEAD study, which reports quantitative amyloid PET using the Centiloid (CL) metric, normalizing raw amyloid PET to a 0 to 100 point scale. 20 , 21 While no consensus on absolute threshold exists, previous studies suggest that CL ≥ 20 reliably indicates the presence of amyloid plaque in the brain and is predictive of future accumulation. 22 , 23 , 24 , 25 To further validate plasma p‐tau species as early AD biomarkers, we used plasma samples from the first group of participants screened for AHEAD and used immunoprecipitation liquid chromatography‐tandem mass spectrometry (LC‐MS/MS) to measure phosphorylated and non‐phosphorylated forms of tau181 and tau217 alongside Aβ42 and Aβ40. The purpose of this study was to determine the utility of MS quantified plasma p‐tau217, np‐tau217, p‐tau181, np‐tau181, their concentration ratios, and Aβ42/Aβ40 ratio measurements to reliably predict brain amyloid PET status in cognitively unimpaired individuals.

RESEARCH IN CONTEXT

  1. Systematic review: The authors used PubMed to conduct their literature search. Studies indicate that plasma phosphorylated tau (p‐tau) improves plasma amyloid beta (Aβ)42/Aβ40 algorithms to correctly identify Alzheimer's disease (AD) status. Analysis of plasma p‐tau from cognitively unimpaired individuals exhibiting high amyloid positron emission tomography (PET) through a clinically approved mass spectrometry assay is lacking.

  2. Interpretation: Our retrospective analysis of blood samples obtained from 1080 cognitively unimpaired AHEAD 3–45 study participants found that including the phosphorylated variant p‐tau217 ratio (p‐tau217/np‐tau217) as a covariate along with Aβ42/Aβ40, age, and apolipoprotein E proteotype improved the area under the curve (AUC) from 0.891 to 0.95. These data complement previous AD plasma biomarker studies.

  3. Future directions: These findings demonstrate that adding plasma p‐tau217/np‐tau217 significantly improved plasma biomarker algorithms for identifying preclinical amyloid PET positivity. Incorporating plasma p‐tau217/np‐tau217 along with Aβ42/Aβ40 ratio into screening algorithms may improve recruitment of cognitively unimpaired participants into preclinical AD trials.

2. METHODS

2.1. Participant eligibility

The current study was approved by the ACTC Participant Advisory Board and ran in compliance with the ethical standards stated in the 1964 Declaration of Helsinki and subsequent amendments. The AHEAD study longitudinally assesses the efficacy of slowing cognitive decline with lecanemab in preclinical AD individuals, with a secondary outcome of measuring changes in both established biomarkers such as amyloid PET as well as exploratory biomarkers including plasma Aβ42/Aβ40 and p‐tau. Participant plasma samples (N = 1080) were collected before plasma Aβ42/Aβ40 testing was implemented for screening into AHEAD (Table 1). Amyloid PET levels were visualized with the NAV4694 tracer and processed into CL as previously described. 26 , 27 For this study, amyloid status by PET is defined as CL ≥20 as defined by the NAV4694 tracer. Plasma samples for these analyses consisted of both individuals with sufficient amyloid PET levels (N = 340; 31.5%) to be eligible for AHEAD and individuals who screen failed (N = 740; 68.5%). The study sample comprised participants screened into the A3 (20–40 CL) and the A45 (> 40 CL) portions of the AHEAD study.

TABLE 1.

Blood plasma sample demographics by AHEAD amyloid status.

AHEAD PET amyloid status
Negative (N = 740) Positive (N = 340) Total (N = 1080)
Age
Mean (SD) 66.66 (6.09) 69.76 (5.45) 67.64 (6.07)
Range 55–80 55–80 55–80
Education
N‐miss 0 1 1
Mean (SD) 16.023 (3.154) 16.482 (2.867) 16.167 (3.073)
Range 6.000–29.000 7.000–26.000 6.000–29.000
APOE
ε2/ε3 84 (11.4%) 7 (2.1%) 91 (8.4%)
ε2/ε4 19 (2.6%) 16 (4.7%) 35 (3.2%)
ε3/ε3 424 (57.3%) 77 (22.6%) 501 (46.4%)
ε3/ε4 188 (25.4%) 182 (53.5%) 370 (34.3%)
ε4/ε4 25 (3.4%) 58 (17.1%) 83 (7.7%)
Sex
Female 495 (66.9%) 224 (65.9%) 719 (66.6%)
Male 245 (33.1%) 116 (34.1%) 361 (33.4%)
Race
Asian 7 (0.9%) 2 (0.6%) 9 (0.8%)
Black or African American 33 (4.5%) 6 (1.8%) 39 (3.6%)
White 683 (92.3%) 328 (96.5%) 1011 (93.6%)
Multi 7 (0.9%) 3 (0.9%) 10 (0.9%)
Other 9 (1.2%) 1 (0.3%) 10 (0.9%)
Unknown or not reported 1 (0.1%) 0 (0.0%) 1 (0.1%)
Ethnicity
Hispanic or Latino/a 124 (16.8%) 22 (6.5%) 146 (13.5%)
Not Hispanic or Latino/a 616 (83.2%) 318 (93.5%) 934 (86.5%)
CFI—Participant
N‐Miss 3 0 3
Mean (SD) 1.55 (2.06) 1.86 (2.17) 1.65 (2.10)
Range 0–11 0–13 0–13
CFI—Study partner
N‐Miss 3 0 3
Mean (SD) 0.65 (1.21) 0.92 (1.45) 0.73 (1.29)
Range 0–8.5 0.000–8.5 0–8.5
MMSE
Mean (SD) 29.04 (0.99) 29.08 (0.99) 29.05 (0.99)
Range 26–30 26–30 26–30
Open in a new tab

Abbreviations: APOE, apolipoprotein E; CFI, Cognitive Function Instrument; MMSE, Mini‐Mental State Examination; PET, positron emission tomography; SD, standard deviation.

2.2. Sample processing and LC‐MS/MS

Participants’ blood samples were collected into K2 ethylenediaminetetraacetic acid tubes, processed into plasma on site, and 0.5 mL plasma aliquots were transferred into polypropylene cryovials that were frozen (−80°C) within 2 hours of phlebotomy. The resulting plasma samples were shipped on dry ice to C2N Diagnostics (St. Louis, Missouri) for processing and analysis. C2N Diagnostics used their LC‐MS/MS analytical platform to quantify plasma Aβ42, Aβ40, p‐tau181, np‐tau181, p‐tau217, np‐tau217 concentrations (all pg/mL), and calculated the plasma Aβ42/Aβ40, p‐tau181/np‐tau181, and p‐tau217/np‐tau217 concentration ratios as previously described. 28 , 29 , 30 Apolipoprotein E (APOE) proteotype was determined as previously described. 29 During plasma sample analysis, all C2N personnel were blinded to participant demographics and metadata. After analysis, plasma biomarker values were transferred from C2N to the AHEAD data coordinating center for statistical analysis. Interindividual differences in plasma tau concentrations were normalized by calculating the phosphorylated/non‐phosphorylated concentration ratios for each proteoform (p‐tau181/np‐tau181 and p‐tau217/np‐tau217, respectively).

2.3. Statistical analysis

Receiver operating characteristic (ROC) curve analysis was conducted to test the performance of blood plasma tau biomarkers to identify brain amyloid PET status in AHEAD screening participants. ROC curves were calculated for each biomarker concentration and their ratios were summarized using the area under the curve (AUC) and corresponding 95% confidence intervals (CI). Cut points were selected by optimizing the Youden index value 31 and the corresponding sensitivity, specificity, and accuracy are reported.

To determine whether the type of p‐tau marker (p‐tau181/np‐tau181 or p‐tau217/np‐tau217) has an impact on predicting amyloid status, we compared the AUC values between the two markers using a paired AUC test of DeLong and Clarke–Pearson via the Sun and Xu algorithm. 32 , 33

Prediction models were then developed to determine whether combining covariates improved the predictive performance of the plasma biomarkers. Of particular interest was the addition of the p‐tau217/np‐tau217 to a model that includes plasma Aβ42/Aβ40, age and APOE proteotype (ε2/ε3, ε2/ε4, ε3/ε3, ε3/ε4, ε4/ε4). To evaluate this, two separate models were developed: one with p‐tau217/np‐tau217 and one without p‐tau217/np‐tau217. These regression models were developed using a machine learning approach known as mixture of experts. 34 Mixture models aim to identify subpopulations, or clusters, from the data based on the outcome variable. The mixture of experts approach also incorporates covariates into the clustering process, making them useful for prediction models. A 10‐fold cross‐validation was used to evaluate the predictive performance of the model where the prediction and ROC analysis was done on each out‐of‐fold sample. AUC and corresponding 95% CIs are provided for each blood plasma biomarker. Thresholds for predicting amyloid PET positivity (defined as greater than 20 CL) were determined by using the maximum Youden index value. Other values such as corresponding accuracy, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were also calculated using the Youden index value. The sample size of the majority of predictors was similar (N = 1080), with the exception of the Amyloid Probability Score (APS; N = 1051). Statistical analyses were conducted using the statistical software R (version 4.1.0). 35 The models generated for this study have been incorporated into the Plasma Predicted Centiloid (PPC) app. 36

3. RESULTS

3.1. Sample demographics

This study included a total of 1080 participants, comprised of 67% females, 14% self‐reporting Hispanic/Latino/a ethnicity, 3.6% self‐reporting Black or African race, and a mean age of 67.64 (standard deviation [SD] = 6.07) years (Table 1). For this study, amyloid status by PET was pre‐defined as CL ≥ 20. Plasma samples included individuals > 20 CL on amyloid PET (amyloid positive, N = 340; 31.5%) and < 20 CL (amyloid negative, N = 740; 68.5%) amyloid PET levels to be eligible for AHEAD. Of all participants, 45.2% had at least one APOE ε4 allele. The ε3/ε3 APOE proteotype was highly represented among amyloid‐negative individuals (57.3%, Table 1), while 75.4% of individuals in the amyloid‐positive group had at least one ε4 allele (Table 1). The average body mass index (BMI) was similar between amyloid status groups (amyloid negative: 28.76, amyloid positive: 27.93). There were no significant differences in creatinine levels among amyloid status (amyloid negative: 67.05, amyloid positive: 67.28). Mean Mini‐Mental State Examination (MMSE) scores were 29.04 (SD = 0.99) for the amyloid negative group and 29.08 (SD = 0.99) for the amyloid positive group. Cognitive Function Instrument (CFI) participant scores for the amyloid‐negative group were 1.55 (SD = 2.06) and 1.86 (SD = 1.45) for the amyloid‐positive group. The CFI Study partner scores for the amyloid‐negative group were 0.65 (SD = 1.21) and 0.92 (SD = 1.45) for the amyloid‐positive group.

3.2. Univariate biomarker associations

The associations for the plasma Aβ and p‐tau concentration ratios with amyloid PET CL are shown in Figure 1. Aβ42/Aβ40 correlation with amyloid PET CL shows a curvilinear relationship (left), with p‐tau ratios indicating a more linear relationship (middle and right). Raw concentrations for Aβ and non‐phosphorylated tau species, as well as BMI, creatinine, and sex are presented in the supporting information (Figures S1 and S2, Table S1).

FIGURE 1.

FIGURE 1

Open in a new tab

Univariable analysis of plasma biomarker ratios. Scatter plots of biomarker ratio values for Aβ42/Aβ40 (left), p‐tau181/np‐tau181 (middle), and p‐tau‐217/np‐tau217 (right) plotted against PET Centiloid. Higher phosphorylation ratio at p‐tau181 and p‐tau217 plasma are associated with higher amyloid PET Centiloid levels. R value corresponds to the Pearson correlation coefficient and corresponding P‐value. Aβ, amyloid beta; PET, positron emission tomography; p‐tau, phosphorylated tau

3.2.1. Plasma p‐tau217/np‐tau217 had higher accuracy for amyloid PET CL than plasma Aβ42/Aβ40 and p‐tau181/np‐tau181 ratios

Each plasma biomarker was evaluated independently for accuracy in detecting amyloid status. The ROC curves show p‐tau217/np‐tau217 has the highest AUC value at 0.92 (95% CI: 0.90, 0.94), followed by plasma Aβ42/Aβ40 at 0.87 (95% CI: 0.84, 0.89) and then p‐tau181/np‐tau181 at 0.77 (95% CI: 0.74, 0.80; Figure 2A‐B, Table 2). Predictors listed in Table 2 as ratios indicate ROC analysis of each concentration ratio independent of other factors. The comparison of p‐tau markers p‐tau181/np‐tau181 and p‐tau217/np‐tau217 indicates statistical significance at an alpha level of 0.05, with p‐tau217/np‐tau217 overall performance being higher than p‐tau181/np‐tau181 (Table 3). The AUC values of p‐tau217/np‐tau217 are comparable to previous plasma predictive models such as the APS that uses Aβ42/Aβ40 and APOE genotype, and Amyloid Probability Score Plus (APS‐Plus) that incorporates tau, Aβ42/Aβ40, and APOE (Figure 2B, Table 2). Therefore, we decided to assess the value of adding p‐tau217/np‐tau217 into our AD status predictive algorithm.

FIGURE 2.

FIGURE 2

Open in a new tab

Increased diagnostic performance with the addition of phosphorylated ratio p‐tau217/np‐tau217. A. ROC curves for p‐tau217/np‐tau217 (light gray solid), p‐tau181/np‐tau181 (light blue, dashed), p‐tau217 (dark gray solid), p‐tau181 (blue broken), Aβ42/Aβ40 (dark red dashed), APS (tan broken), and APS‐Plus (yellow solid) for predicting brain amyloid PET status. B, The corresponding AUC and 95% confidence intervals for each biomarker. The phosphorylation at p‐tau217/np‐tau217 ratio exhibited high sensitivity and specificity for amyloid PET status. C, ROC curves from the raw p‐tau217 (dark gray dashed), p‐tau217/np‐tau217 (light gray solid), and prediction models incorporating covariates ptau217/np‐tau217, Aβ42/Aβ40, age and APOE status (full model, orange solid), p‐tau217/np‐tau217, age and APOE status (full model minus Aβ, teal solid), ptau217/np‐tau217, age and APOE status (full model minus ptau217, lavender dash), ptau217/np‐tau217, Aβ42/Aβ40 and age (full model minus APOE, pink broken), ptau217/np‐tau217 and age (full model minus Aβ & APOE, dark blue solid), along with APS (tan solid) and APS‐Plus (yellow broken). D, AUC and 95% confidence intervals for each modeled biomarker. Aβ, amyloid beta; APOE, apolipoprotein E; APS, Amyloid Probability Score; AUC, area under the receiver operating curve; PET, positron emission tomography; ROC, receiver operating characteristic

TABLE 2.

Performance values of plasma biomarker ratios.

Predictor AUC Cut point Accuracy Sensitivity Specificity PPV NPV Number of PET scans b Total number of screens b
Aβ42/Aβ40 0.87 (0.84, 0.89) 0.10 0.80 0.81 0.79 0.64 0.90 1565 3913
p‐tau181/np‐tau181 0.77 (0.74, 0.80) 16.46 0.74 0.69 0.77 0.58 0.84 1731 4615
p‐tau217/np‐tau217 0.92 (0.90, 0.94) 1.55 0.87 0.82 0.89 0.78 0.91 1288 3885
Full model 0.95 (0.94, 0.96) 18.65 a 0.90 0.88 0.91 0.81 0.94 1233 3600
Full model minus APOE 0.95 (0.93, 0.96) 16.31 a 0.88 0.88 0.88 0.77 0.94 1294 3612
Full model minus Aβ 0.93 (0.92, 0.95) 19.19 a 0.89 0.84 0.91 0.81 0.92 1235 3789
Full model minus p‐tau217 0.89 (0.87, 0.91) 23.15 a 0.83 0.79 0.84 0.70 0.90 1433 4030
APS c 0.90 (0.88, 0.92) 28.50 0.83 0.81 0.83 0.69 0.91 1451 3951
APS‐Plus 0.94 (0.93, 0.96) 14.50 0.87 0.91 0.85 0.74 0.95 1356 3495
Open in a new tab

Abbreviations: Aβ, amyloid beta; APOE, apolipoprotein E; APS, Amyloid Probability Score; AUC, area under the curve; NPV, negative predictive value; PET, positron emission tomography; p‐tau, phosphorylated tau; PPV, positive predictive value.

a

Cut point for the models are on the scale of predicted Centiloid.

b

Number needed to fill a trial of 1000 participants assuming a prevalence of 0.32.

c

Sample size = 1051.

TABLE 3.

Comparison of phosphorylated tau plasma ratios.

AUC
p‐tau181/np‐tau181 p‐tau217/np‐tau217 Statistic P‐value
0.77 (0.74, 0.80) 0.92 (0.90, 0.94) –8.79 <0.01
Open in a new tab

Abbreviations: AUC, area under the curve; p‐tau, phosphorylated tau.

3.2.2. PPC: An algorithm that included both plasma p‐tau217/np‐tau217 and Aβ42/Aβ40 had the highest accuracy

The ROC analysis (Figure 2A‐B) and mixture of experts results (Figure 2C‐D) are reported in Figure 2. Predictors in Table 2 prefaced with “Model” are the mixture of experts results of the biomarker listed along with the covariates of APOE proteotype and age. When p‐tau217/np‐tau217 was added to the model that included Aβ42/Aβ40, age, and APOE proteotype (full model), the AUC improved from 0.92 to 0.95 (95% CI: 0.94, 0.96; Table 3, Figure 2C‐D).

3.2.3. The addition of p‐tau217/ np‐tau217 improves prediction performance at higher NAV CL levels

Figure 3 plots the AUCs for the model with Aβ, model with p‐tau217, and model with Aβ and p‐tau217 against NAV PET cut point. CIs are depicted around each model line. At > 20 CL, the AUCs of the model with Aβ decline, while the AUCs of both the model with p‐tau217 and model with Aβ and p‐tau217 begin to decline at > 40 CL (Figure 3).

FIGURE 3.

FIGURE 3

Open in a new tab

The combination of Aβ42/40 of p‐tau217/np‐tau217 improves amyloid status prediction in individuals with CL > 20. AUC plotted against NAV PET cut point. Predictors plotted: full model (black), full model minus Aβ (yellow), and full model minus p‐tau217 (blue). Shaded regions signify confidence intervals (CI). Aβ, amyloid beta; AUC, area under the curve; CL, Centiloid; PET, positron emission tomography; p‐tau, phosphorylated tau

4. DISCUSSION

In this analysis of baseline samples from cognitively unimpaired AHEAD study participants, we assessed the performance of plasma p‐tau181 and p‐tau217 concentrations and ratios for correctly identifying participants with an amyloid burden of CL ≥ 20. We further evaluated whether the addition of p‐tau217/np‐tau217 to the existing plasma screening model (based on Aβ42/Aβ40 ratio, APOE proteotype, and participant age) improved the existing model's classification performance. The findings demonstrate an improved performance for identifying amyloid PET positive cognitively unimpaired participants when using plasma p‐tau217/np‐tau217, but the best performance as indicated by AUC was achieved when plasma p‐tau217 and Aβ42/Aβ40 ratios were combined in a model that predicted cerebral amyloid PET status. Furthermore, we identified that the predictability of the Aβ42/Aβ40 ratio is limited after CL > 20 and only with the addition of p‐tau217 is enhanced, suggesting that additional plasma biomarkers should be measured in individuals with higher amyloid PET levels.

Previous studies using analytical immunoassays highlight the potential for p‐tau using proteoforms to identify AD status. While elevated plasma p‐tau181 has been shown to correlate with amyloid status in MCI and AD cases, 9 , 37 our study, as well as other studies, report that plasma p‐tau217 performs better than p‐tau181 for predicting AD pathology (Figure 2A,B). 38 , 39 We show that p‐tau217/np‐tau217 can be used to predict presence of AD pathological hallmarks such as amyloid neuritic plaque as defined by CL ≥ 20. Interestingly, another study concluded that p‐tau217 was considered abnormally expressed at CL ≥ 35.4. 38 That study cohort comprised a high proportion of individuals with a family history of AD (47.4%) and a high percentage of APOE ε4 carriers (34.7%). In our study sample, 45.2% of all participants had at least one APOE ε4 allele, and more than 70% of APOE ε4 carriers were designated amyloid positive (Table 1). The high percentage of APOE ε4 carriers may influence which phosphorylated tau species more reliably detects brain amyloid status. We demonstrate that including plasma p‐tau217/np‐tau217 as a predictor variable along with age and APOE proteotype increased the overall predictive value of an algorithm that included only Aβ42/Aβ40 (Figure 2C,D). No differences between amyloid status were seen in other potential predictors associated with p‐tau217 elevation such as creatinine and BMI, although the sample size was reduced due to inconsistent reporting (Table 1, Figure S3 in supporting information). While both creatinine and BMI are associated with other AD biomarkers such as glial fibrillary acidic protein and p‐tau, the confounding effects on prediction status are minor. 40 The effects of sex were found to be minimal in the current cohort and sex was therefore not included as a covariate in our predictive models (Figure S3).

A previous study reported similar AUC values when adding MS‐based p‐tau217/np‐tau217 measures to their predictive models for AD pathology. 16 More than half of the individuals in the amyloid positive group of the previous study were clinically diagnosed with AD cognitive symptoms, indicating that peripheral p‐tau217 may be present throughout the disease stages of AD. Longitudinal studies have suggested that p‐tau217 increases with disease progression. 17 To summarize, plasma p‐tau217/np‐tau217 shows potential as an early biomarker for AD pathology and including p‐tau217/np‐tau217 into Aβ42/Aβ40 algorithms further improved screening accuracy for amyloid PET status. Although our study indicated that p‐tau181 performed worse than p‐tau217 and Aβ42/40 in predicting amyloid status, plasma p‐tau181 levels can potentially discriminate between AD and other tauopathies. 41 , 42 Future studies investigating the diagnostic value of other chemically modified tau proteoforms might further streamline screening for clinical trials and precisely diagnose neurodegenerative diseases.

The current study cohort is derived from cognitively unimpaired, high‐risk individuals eligible for the A3 (20–40 CL) and the A45 (> 40 CL) cohorts in the AHEAD study, and therefore a heterogenous population with regard to levels of brain amyloid. The AUCs for p‐tau217/np‐tau217 and modeled p‐tau217/np‐tau217 were higher than that of Aβ42/Aβ40 alone (AUC: 0.92 and 0.95 vs. 0.87; Table 1). Our observations are consistent with previous studies that observed that plasma Aβ42/Aβ40 did not improve predictability alongside a plasma tau biomarker in MCI samples. 43 However, the utility of plasma Aβ42/Aβ40 as a diagnostic may be sufficient at discerning amyloid status in individuals with low levels of amyloid PET (Figure 3). In individuals with higher levels of brain amyloid PET, the AUCs from the predictor model that included p‐tau217/np‐tau217 were greater than from the model using only Aβ42/Aβ40 (Figure 3), suggesting that additional biomarkers are needed to properly screen individuals with high amyloid PET.

The predictive models generated for this study (Table 3) along with the code are available for public use. 36 These models can be used to obtain predicted CL values based on the NAV4694 amyloid PET for C2N Aβ42/Aβ40, p‐tau217/np‐tau217, and other biomarkers. While ideally the PPC values would be applicable for all studies, caution should be taken when applying these CL values to studies using other tracers. Future studies using different tracers would serve to cross‐validate our findings and expand the application of the PPC.

Limitations of the current study include the sample participants' homogenous racial composition with a high proportion reporting being non‐Hispanic White (93.6%, Table 1). Extensive preclinical studies such as the A4 study have screen failed Hispanic and non‐White participants at a higher frequency after the first screening visit, despite a higher prevalence of AD in Blacks/African Americans and Hispanics/Latino(s) compared to non‐Hispanic Whites. 44 While the influence of demographic, lifestyle, and cognition‐based variables 5 , 45 , 46 can be factored into biofluid analyses, the impact of race and ethnicity remains unclear. Indeed, notable differences in the performance of AD biomarkers among different racial groups have been reported. 47 , 48 , 49 , 50 Therefore, expanding the racial and ethnic representation among clinical trial participants would provide crucial information to minimize AD health disparities. 51 As the AHEAD screening sample becomes more representative, work will continue to evaluate potential racial differences in biomarker performance.

Our findings demonstrate the utility of plasma biomarkers for further improving the screening process of AD status as defined by amyloid PET imaging. Inclusion of plasma p‐tau, whether raw concentration or normalized ratios, further improved the accuracy of models based on plasma Aβ42/Aβ40 for detecting amyloid PET positivity (Table 2). In our predictive algorithm, p‐tau217/np‐tau217 plus Aβ42/Aβ40 conferred the best classification performance at the preclinical stage of AD next to APS‐Plus (Table 3). These findings suggest that combining plasma p‐tau217 may be sufficient for models that predict amyloid PET status. Ongoing work will need to evaluate how these plasma biomarker measures can be applied to other PET tracers, including tau PET, with varying levels and regional distributions for amyloid and tau PET neuropathology. Additionally, longitudinal analysis would allow for a thorough assessment of the association of p‐tau217 and Aβ42/Aβ40 on amyloid PET. To further distinguish the specificity of blood plasma markers to AD, analysis from individuals with comorbid conditions such as synucleinopathy, TAR DNA‐binding protein 43–associated diseases, as well as traumatized brains need to be conducted. Our current priorities involve expanding these findings to more representative populations to determine whether the specific plasma Aβ and p‐tau217/np‐tau217, cutoff values, and their relation to amyloid and tau PET status are similar across different racial, ethnic, and other underrepresented groups.

CONFLICT OF INTEREST STATEMENT

Sara Abdel‐Latif, Jennifer Ngolab, and Keith A. Johnson have nothing to disclose. Oliver Langford is an employee of the Keck School of Medicine of USC. Matthew R. Meyer, Traci Wente‐Roth, Kristopher M. Kirmess, Tim West, Kevin E. Yarasheski, and Joel B. Braunstein are paid employees of C2N Diagnostics. Kevin E. Yarasheski has research grants from the NIH, BrightFocus Foundation, GHR Foundation, and Alzheimer's Drug Discovery Foundation. Pallavi Sachdev and Michael Irizarry are paid employees of Eisai. Robert A. Rissman, Rema Raman, Michael C. Donohue, and Charisse N. Winston have grants from the NIH. Rema Raman has received grants from Eisai, Eli Lilly, the Alzheimer's Association, and the American Heart Association. Michael C. Donohue serves as a consultant for Roche and has a spouse working at Janssen. Gustavo Jimenez‐Maggiora has received support from C2N Diagnostics, Eisai Co, Ltd, and the NIH. Michael S. Rafii serves as a consultant for AC Immune, Ionis, Keystone Bio, and Positrigo. Paul S. Aisen has research grants from Eisai, NIH, Lilly, the Alzheimer's Association, and Janssen and serves as a consultant for Merck, Bristol Myers Squibb, Switch Therapeutics, Roche, Arrowhead, ImmunoBrain, Checkpoint, Biogen, Abbvie, Genetech, and NewAmsterdam Pharma. Reisa A. Sperling has grants from the NIH, the Alzheimer's Association, the GHR Foundation, Eli Lilly, and Eisai, and has consulted for AC Immune, Acumen, Alector, Alnylam, Cytox, Genentech, Janssen, JOMDD, Nervgen, Neuraly, Neurocentria, Oligomerix, Prothena, Renew, Shionogi, Vigil Neuroscience, Ionis, Vaxxinity, and Bristol Myers Squibb. Author disclosures are available in the supporting information.

CONSENT STATEMENT

The current study was approved by the ACTC Participant Advisory Board and ran in compliance with the ethical standards stated in the 1964 Declaration of Helsinki and subsequent amendments.

Supporting information

Supporting Information

ALZ-20-1214-s002.docx (526.5KB, docx)

Supporting Information

ALZ-20-1214-s001.pdf (848.2KB, pdf)

ACKNOWLEDGMENTS

The authors would like to thank the Alzheimer's Clinical Trials Consortium (ACTC) and USC Alzheimer's Therapeutic Research Institute (ATRI) coordinating center staff and site PIs and site staff for AHEAD 3–45 Study. This work was supported by the NIH/NIA (Grant Numbers AG018440, AG058252, AG078109 AG058533 and AG073979 to RAR; NIA ACTC U24 AG057437, A3 R01 AG054029, and A45 R01AG061848 to PSA and RAS). C2N Diagnostics was supported by NIH (grant number R44 AG059489), BrightFocus Foundation (grant number: CA2016636), The Gerald and Henrietta Rauenhorst Foundation, and the Alzheimer's Drug Discovery Foundation (grant number: GC‐201711‐2013978).

COLLABORATORS

AHEAD 3–45 Study Team members

Paul Aisen, Shobha Dhadda, Mike Irizarry, Keith Johnson, Lynn Kramer, Reisa Sperling, Jeremy Pizzola, Cher Herh, Barbara Bartocci, Melissa Korba, Sarah Walter, Doug Middlesteadt, Gayatri Girwarr, Francesca Kelvy Robert Rissman, Sara Abdel‐Latif, June Kaplow, Akihiko Koyama, Jagadeesh Aluri, Froy Garcia, Christine Pham, Dominique Webster, Jiyoon Choi, Mike Donohue, Karin Ernstrom, Oliver Langford, David Li, Rema Raman, Gopalan Sethuraman, Claire Roberts, Misako Takei, Masaki Nakagawa, Jennifer Salazar, Alison Belsha, Maria Arampatzidou, Caileigh Zimmerman, Meagan Adams, John Petrera, Rusty Harris, Shilpa Kakad, Calvin Chen, Nao Hashida, Noboru Takizawa, Teng Fong Chin, Gustavo Jimenez‐Maggiora, Olusegun Adegoke, Veasna Tan, Olga Baryshnikava, Iris Sim Dacio, Jorge Corral, Michael Wolgast, James Cassidy, Malathi Reddy, Yuko Nakai, Yusuke Nakayama, Jason Karlawish, Josh Grill, Stefania Bruschi, Hongmei Qiu, Jia‐shing So, Chad Swanson, Michael Rafii, Dobri Baldaranov, Victoria Garcia, Jin Zhou, Takuya Yagi, Lindsey Hergesheimer, Ricky Cordova, Sid Raorane, Solana Leisher, Renarda Jones, Cliff Jack, Mike Weiner, Bret Borowski, Petrice Cogswell, Derek Flenniken, Arvin Forghanian‐Arani, Kejal Kantarci, Denise Reyes, Stephanie Rossie, Aaron Schultz, Duygu Tosun‐Turgut, Diana Truran, William Turke, Samantha Zuk, Cecily Jenkins, Kate Papp, Dorene Rentz, Steven Krause, Elizabeth Shaffer, Mark Hodgkinson, Stacie O'Sullivan, Taylor Clanton, Carl Hill, Dylan Kirn, Yakeel Quiroz, Jessica Langbaum, Laurie Ryan, Kristina McLinden.

AHEAD 3–45 Study Site PIs

David Weisman (Abington Neurological Associates), Ramon Rodriguez (Charter Research), Martin Farlow (Indiana University), Michael Plopper (Sharp Mesa Vista Hospital), Oscar Lopez (University of Pittsburgh), Gil Fernandez‐Yera (Advanced Clinical Research Network), Babak Tousi (Cleveland Clinic, Lou Ruvo, Cleveland), Paul Rosenberg (Johns Hopkins University), Sharon Sha (Stanford University), Anton Porsteinsson (University of Rochester Medical Center), Sanjiv Sharma (Advanced Memory Research Inst. of NJ), Jiong Shi (Cleveland Clinic Lou Ruvo, Las Vegas), Cherian Verghese (Keystone Clinical Studies, LLC), Scott Losk (Summit Research Network), Lon Schneider (University of Southern California), David Watson (Alzheimer's Research & Treatment Ctr.), Karen Bell (Columbia University Medical Center), Neill Graff‐Radford (Mayo Clinic, Jacksonville), Ira Goodman (Synexus Clinical Research, Orlando), Amanda Smith (University of South Florida), Mohammad Bolouri (AMC Research), Jonathan Liss (Columbus Memory Center), Jonathan Graff‐Radford (Mayo Clinic, Rochester) Ayesha Lall (Synexus Clinical Research, The Villages), Eduardo Marques Zilli (UT Health Science Ctr. at San Antonio), Pierre Tariot (Banner Alzheimer's Institute), Donald Marks (Donald S. Marks), Mary Sano (Mt. Sinai School of Medicine), Sharon Cohen (Toronto Memory Program), Brendan Kelley (University of Texas Southwestern), Alireza Atri (Banner Sun Health Research Institute), James Burke (Duke Health Center), John Scott (National Clinical Research, Inc.), Marissa Natelson Love (University of Alabama, Birmingham), Charles Bernick (Univ. of WA, Memory & Wellness Ctr.), Melissa Yu (Baylor College of Medicine), Hamid Okhravi (Eastern Virginia Medical School), Ian Grant (Northwestern University), David Sultzer (University of California, Irvine), Elaine Peskind (Univ. of WA, SBICR), Ronald Killiany (Boston University), Ihab Hajjar (Emory University School of Medicine), Douglas Scharre (Ohio State University), Julio Rojas Martinez (University of California, San Francisco), Cynthia Carlsson (University of Wisconsin), Mark Brody (Brain Matters Research), Raymond Scott Turner (Georgetown University), Aimee Pierce (Oregon Health & Science University), Ryan Townley (University of Kansas Medical Center), Paul Newhouse (Vanderbilt University Medical Center), Seth Gale (Brigham and Women's Hospital), Yaneicy Gonzalez Rojas (Gonzalez MD & Aswad Health), Alex White (Progressive Medical Research), Gregory Jicha (University of Kentucky), Suzanne Craft (Wake Forest School of Medicine), Stephen Salloway (Butler Hospital Memory and Aging Ctr.), William Shankle Hoag (Pickup Family Neurosciences Inst.), Anette Nieves (Renstar Medical Research), Judith Heidebrink (University of Michigan), Joy Snider (Washington University, St. Louis), Alan Lerner (Case Western Reserve University), Joseph Masdeu (Houston Methodist Neurological Inst.), Jonathan Drake (Rhode Island Hospital), Leigh Johnson (Univ. of North TX Health Science Ctr.), Ranjan Duara Wein (Center for Clinical Research), Michael Karathanos (Central States Research, LLC), Thomas Obisesan (Howard University), Neelum Aggarwal (Rush University), Sanjeev Vaishnavi (University of Pennsylvania), Christopher van Dyck (Yale University).

Rissman RA, Langford O, Raman R, et al. Plasma Aβ42/Aβ40 and phospho‐tau217 concentration ratios increase the accuracy of amyloid PET classification in preclinical Alzheimer's disease. Alzheimer's Dement. 2024;20:1214–1224. 10.1002/alz.13542

REFERENCES

  • 1. Rajan KB, Weuve J, Barnes LL, McAninch EA, Wilson RS, Evans DA. Population estimate of people with clinical Alzheimer's disease and mild cognitive impairment in the United States (2020‐2060). Alzheimer Dementia. 2021;17:1966‐1975. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Jack CR, Bennett DA, Blennow K, et al. NIA‐AA research framework: toward a biological definition of Alzheimer's disease. Alzheimer Dementia. 2018;14:535‐562. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Dubois B, Feldman HH, Jacova C, et al. Research criteria for the diagnosis of Alzheimer's disease: revising the NINCDS‐ADRDA criteria. Lancet Neurol. 2007;6:734‐746. [DOI] [PubMed] [Google Scholar]
  • 4. Sperling RA, Jack CR, Aisen PS. Testing the right target and right drug at the right stage. Sci Transl Med. 2011;3:111cm33‐111cm33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Donohue MC, Sperling RA, Petersen R, et al. Association between elevated brain amyloid and subsequent cognitive decline among cognitively normal persons. JAMA. 2017;317:2305‐2316. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Mormino EC, Papp KV, Rentz DM, et al. Early and late change on the preclinical Alzheimer's cognitive composite in clinically normal older individuals with elevated amyloid β. Alzheimer Dement. 2017;13:1004‐1012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Petersen RC, Wiste HJ, Weigand SD, et al. Association of elevated amyloid levels with cognition and biomarkers in cognitively normal people from the community. JAMA Neurol. 2016;73:85‐92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Barthélemy NR, Li Y, Joseph‐Mathurin N, et al. A soluble phosphorylated tau signature links tau, amyloid and the evolution of stages of dominantly inherited Alzheimer's disease. Nat Med. 2020;26:398‐407. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Mielke MM, Hagen CE, Xu J, et al. Plasma phospho‐tau181 increases with Alzheimer's disease clinical severity and is associated with tau‐ and amyloid‐positron emission tomography. Alzheimers Dement. 2018;14:989‐997. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Blennow K, Dubois B, Fagan AM, Lewczuk P, de Leon MJ, Hampel H. Clinical utility of cerebrospinal fluid biomarkers in the diagnosis of early Alzheimer's disease. Alzheimer Dement. 2015;11:58‐69. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Ray S, Britschgi M, Herbert C, et al. Classification and prediction of clinical Alzheimer's diagnosis based on plasma signaling proteins. Nat Med. 2007;13:1359‐1362. [DOI] [PubMed] [Google Scholar]
  • 12. Laske C, Sohrabi HR, Frost SM, et al. Innovative diagnostic tools for early detection of Alzheimer's disease. Alzheimers Dement. 2015;11:561‐578. [DOI] [PubMed] [Google Scholar]
  • 13. Henriksen K, O'Bryant SE, Hampel H, et al. The future of blood‐based biomarkers for Alzheimer's disease. Alzheimers Dement. 2014;10:115‐131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. O'Bryant SE, Edwards M, Johnson L, et al. A blood screening test for Alzheimer's disease. Alzheimer Dement. 2016;3:83‐90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Korecka M, Shaw LM. Mass spectrometry‐based methods for robust measurement of Alzheimer's disease biomarkers in biological fluids. J Neurochem. 2021;159:211‐233. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Barthélemy NR, Horie K, Sato C, Bateman RJ. Blood plasma phosphorylated‐tau isoforms track CNS change in Alzheimer's disease. J Exp Med. 2020:217. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Mattsson‐Carlgren N, Janelidze S, Palmqvist S, et al. Longitudinal plasma p‐tau217 is increased in early stages of Alzheimer's disease. Brain. 2020;143:3234‐3241. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Teunissen CE, Thijssen EH, Verberk IMW. Plasma p‐tau217: from ‘new kid’ to most promising candidate for Alzheimer's disease blood test. Brain. 2020;143:3170‐3172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Janelidze S, Palmqvist S, Leuzy A, et al. Detecting amyloid positivity in early Alzheimer's disease using combinations of plasma Aβ42/Aβ40 and p‐tau. Alzheimer Dement. 2022;18:283‐293. [DOI] [PubMed] [Google Scholar]
  • 20. Klunk WE, Koeppe RA, Price JC, et al. The Centiloid Project: standardizing quantitative amyloid plaque estimation by PET. Alzheimer Dement. 2015;11:1‐15. e4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Rafii MS, Sperling RA, Donohue MC, et al. The AHEAD 3‐45 study: design of a prevention trial for Alzheimer's disease. Alzheimer Dement. 2022;n/a. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Amadoru S, Doré V, McLean CA, et al. Comparison of amyloid PET measured in Centiloid units with neuropathological findings in Alzheimer's disease. Alzheimer Res Ther. 2020;12:22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Milà‐Alomà M, Shekari M, Salvadó G, et al. Cognitively unimpaired individuals with a low burden of Aβ pathology have a distinct CSF biomarker profile. Alzheimer Res Ther. 2021;13:134. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. La Joie R, Ayakta N, Seeley WW, et al. Multisite study of the relationships between antemortem [11C]PIB‐PET centiloid values and postmortem measures of Alzheimer's disease neuropathology. Alzheimer Dement. 2019;15:205‐216. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Salvadó G, Molinuevo JL, Brugulat‐Serrat A, et al. Centiloid cut‐off values for optimal agreement between PET and CSF core AD biomarkers. Alzheimer Res Ther. 2019;11:27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Rowe CC, Jones G, Doré V, et al. Standardized expression of 18F‐NAV4694 and 11C‐PiB β‐amyloid PET results with the centiloid scale. J Nucl Med. 2016;57:1233‐1237. [DOI] [PubMed] [Google Scholar]
  • 27. Aisen PS, Bateman RJ, Carrillo M, et al. Platform trials to expedite drug development in Alzheimer's disease: a report from the EU/US CTAD task force. J Prev Alzheimers Dis. 2021:1‐7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Nakamura A, Kaneko N, Villemagne VL, et al. High performance plasma amyloid‐β biomarkers for Alzheimer's disease. Nature. 2018;554:249‐254. [DOI] [PubMed] [Google Scholar]
  • 29. West T, Kirmess KM, Meyer MR, et al. A blood‐based diagnostic test incorporating plasma Aβ42/40 ratio, ApoE proteotype, and age accurately identifies brain amyloid status: findings from a multi cohort validity analysis. Mol Neurodegener. 2021;16:30. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Hu YKK, Meyer MR, Rabinovici GD, et al. Assessment of a plasma amyloid probability score to estimate amyloid positron emission tomography findings among adults with cognitive impairment. JAMA Netw Open. 2022;5(4):e228392. (in press). JAMA Network Open, 2022; 5(4): e228392 2022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Youden WJ. Index for rating diagnostic tests. Cancer. 1950;3:32‐35. [DOI] [PubMed] [Google Scholar]
  • 32. DeLong ER, DeLong DM, Clarke‐Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44:837‐845. [PubMed] [Google Scholar]
  • 33. Sun X, Xu W. Fast implementation of delong's algorithm for comparing the areas under correlated receiver operating characteristic curves. IEEE Signal Processing Letters. 2014;21:1389‐1393. [Google Scholar]
  • 34. Murphy K, Murphy TB. Gaussian parsimonious clustering models with covariates and a noise component. Adv Data Anal Classification. 2020;14:293‐325. [Google Scholar]
  • 35. Core Team R. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. 2021. [Google Scholar]
  • 36. Langford O, Donohue MC, Raman R, Plasma Predicted Centiloid. 2023. https://atrihub.shinyapps.io/ppc/
  • 37. Fowler C, Stoops E, Rainey‐Smith S, et al. Plasma p‐tau181/Aβ1‐42ratio predicts Aβ‐PET status and correlates with CSF‐p‐tau181/Aβ1‐42and future cognitive decline. medRxiv. 2022:22272320. 2022.03.13. [Google Scholar]
  • 38. Milà‐Alomà M, Ashton NJ, Shekari M, et al. Plasma p‐tau231 and p‐tau217 as state markers of amyloid‐β pathology in preclinical Alzheimer's disease. Nat Med. 2022;28:1797‐1801. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Palmqvist S, Janelidze S, Quiroz YT, et al. Discriminative accuracy of plasma phospho‐tau217 for Alzheimer disease vs. other neurodegenerative disorders. JAMA. 2020;324:772‐781. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Binette AP, Janelidze S, Cullen N, Mattsson‐Carlgren N, Hansson O. Creatinine and body mass index influence on plasma amyloid ratio, p‐tau217 and neurofilament light. Alzheimer Dement. 2021:17. [Google Scholar]
  • 41. Thijssen EH, La Joie R, Wolf A, et al. Diagnostic value of plasma phosphorylated tau181 in Alzheimer's disease and frontotemporal lobar degeneration. Nat Med. 2020;26:387‐397. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Janelidze S, Mattsson N, Palmqvist S, et al. Plasma P‐tau181 in Alzheimer's disease: relationship to other biomarkers, differential diagnosis, neuropathology and longitudinal progression to Alzheimer's dementia. Nat Med. 2020;26:379‐386. [DOI] [PubMed] [Google Scholar]
  • 43. Cullen NC, Leuzy A, Palmqvist S, et al. Individualized prognosis of cognitive decline and dementia in mild cognitive impairment based on plasma biomarker combinations. Nat Aging. 2021;1:114‐123. [DOI] [PubMed] [Google Scholar]
  • 44. Raman R, Quiroz YT, Langford O, et al. Disparities by race and ethnicity among adults recruited for a preclinical alzheimer disease trial. JAMA Netw Open. 2021;4:e2114364. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45. Sperling RA, Donohue MC, Raman R, et al. Association of factors with elevated amyloid burden in clinically normal older individuals. JAMA Neurol. 2020;77:735‐745. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46. Hedden T, Oh H, Younger AP, Patel TA. Meta‐analysis of amyloid‐cognition relations in cognitively normal older adults. Neurology. 2013;80:1341‐1348. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47. Howell JC, Watts KD, Parker MW, et al. Race modifies the relationship between cognition and Alzheimer's disease cerebrospinal fluid biomarkers. Alzheimers Res Ther. 2017;9:88. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48. Morris JC, Schindler SE, McCue LM, et al. Assessment of racial disparities in biomarkers for Alzheimer disease. JAMA Neurol. 2019;76:264‐273. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49. Garrett SL, McDaniel D, Obideen M, et al. Racial disparity in cerebrospinal fluid amyloid and tau biomarkers and associated cutoffs for mild cognitive impairment. JAMA Netw Open. 2019;2:e1917363. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50. Schindler SE, Karikari TK, Ashton NJ, et al. Effect of race on prediction of brain amyloidosis by plasma Aβ42/Aβ40, phosphorylated tau, and neurofilament light. Neurology. 2022;99:e245‐e257. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51. Manzine PR, Vatanabe IP, Peron R, et al. Blood‐based biomarkers of Alzheimer's disease: the long and winding road. Curr Pharm Des. 2020;26:1300‐1315. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information

ALZ-20-1214-s002.docx (526.5KB, docx)

Supporting Information

ALZ-20-1214-s001.pdf (848.2KB, pdf)

Articles from Alzheimer's & Dementia are provided here courtesy of Wiley

RESOURCES