Alzheimer's & Dementia

INTRODUCTIONPlasma biomarkers for Alzheimers disease (AD) pathology promise scalable diagnostic access, yet their performance in diverse, population-representative cohorts remains uncharacterized. We evaluated equity and transportability of plasma amyloid-tau-neurodegeneration (ATN) biomarkers in a nationally representative U.S. aging cohort. METHODSCross-sectional analysis of 4,427 adults aged [&ge;]50 years from the 2016 Health and Retirement Study Venous Blood Study. Plasma biomarkers (A{beta}42/40, pTau181, NfL, GFAP) were classified using established ATN criteria. Survey weights produced population-representative estimates. Outcomes included biomarker-cognition associations, fairness metrics (sensitivity, specificity, predictive values) stratified by race/ethnicity and sex, and education-stratified analyses. RESULTSAmong 4,427 participants representing 36.6 million U.S. adults (weighted: 68 years, 55% female, 79% White), survey-weighted analysis revealed tau as the only biomarker maintaining robust cognitive associations ({beta}=-0.74, p<0.001), while amyloid ({beta}=0.11, p=0.43) and neurodegeneration ({beta}=-0.27, p=0.08) lost significance. White participants demonstrated 12-percentage-point higher sensitivity than Black participants (23.4% vs. 11.4%), with Black women showing lowest sensitivity (8.8%). Educational attainment modified biomarker effects: low-education groups showed paradoxical positive amyloid associations ({beta}=0.74, p=0.01) and amplified neurodegeneration effects ({beta}=-1.02, p=0.006). Race-specific optimal cutpoints differed by 40%. Vascular comorbidity burden was higher in Black (82%) and Hispanic (73%) versus White (65%) participants, yet associations persisted after vascular adjustment. DISCUSSIONPlasma ATN biomarkers demonstrate significant equity gaps and differential transportability across demographic subgroups. The 12-percentage-point sensitivity disparity and education-dependent effect modification highlight barriers to equitable implementation. Population-based validation with fairness metrics should be prerequisite for clinical deployment.

INTRODUCTIONThis study uses longitudinal amyloid biomarker and cognitive data to generate sample size estimates for two-armed, pre-clinical amyloid clearance clinical trials. METHODSPET PiB DVR ranges defined three amyloid groups (positive, "A+"; sub threshold/low positive, "subA+"; and negative, "A-") in cognitively unimpaired Wisconsin Registry for Alzheimers Prevention participants. Amyloid group trajectories estimated from mixed effects models informed per-treatment-arm sample size estimates to detect plausible treatment effects over 3-year (biomarker) or 6-year (cognition) study windows (80% power). RESULTSTo detect [&ge;]60% slowing in PiB accumulation, [&le;]40 may be needed per arm for both SubA+ and A+; to detect the same effect sizes in plasma p-tau217 trajectories, [~]50-1700 are needed, depending on assay and amyloid subgroup. Among cognitive outcomes, Digit Symbol Substitution and a 5-test Preclinical Alzheimers Cognitive Composite consistently required fewest (<2000) per arm. DISCUSSIONEarly intervention study planning will benefit from selection of outcomes that are most sensitive to AD biomarker-related preclinical change.

INTRODUCTIONElevated tau is temporally proximal to dementia onset but less is known about factors influencing T+ onset age and time to dementia following T+ in Alzheimers disease. We used sampled iterative localized approximation (SILA) estimated T+ onset age (ETOA) to investigate factors associated with T+ age and time from T+ to dementia onset in ADNI. METHODSUsing SILA-estimated A+ and T+ onset ages derived from 18F-Flortaucipir, 18F-Florbetapir, and 18F-Florbetaben PET and Cox proportional hazards and accelerated failure time models, we analyzed APOE, sex, amyloid burden, age, educational attainment, and literacy associations with ETOA and time from T+ to dementia. RESULTSHigher amyloid, APOE-{varepsilon}4, lower education, and lower literacy associated with younger ETOA. Older ETOA and higher amyloid associated with shorter time from T+ to dementia. DISCUSSIONThis work highlights the prognostic value of ETOA and the need to better characterize factors contributing to ETOA and dementia onset in AD.

BackgroundDetection of cerebral amyloid pathology currently requires amyloid PET imaging ($5,000-$8,000) or cerebrospinal fluid analysis via lumbar puncture, procedures that are inaccessible for population-level screening. The FDA-cleared Lumipulse G pTau217/A{beta}1-42 plasma ratio test (May 2025) represents the first approved blood-based alternative; however, single-ratio approaches cannot distinguish Alzheimers disease (AD) from non-AD neurodegeneration or provide multi-dimensional disease characterization. MethodsWe developed Virtual Spectral Decomposition (VSD), a framework that decomposes plasma biomarker profiles into biologically interpretable diagnostic channels. Four plasma biomarkers--phosphorylated tau-217 (pTau217), amyloid-{beta}42/40 ratio, neurofilament light chain (NfL), and glial fibrillary acidic protein (GFAP)--were measured in 1,139 Alzheimers Disease Neuroimaging Initiative (ADNI) participants. Each biomarker was mapped to a VSD channel representing a distinct pathophysiological axis: tau/amyloid phosphorylation, amyloid clearance, neurodegeneration, and astrocytic activation. Channel weights were calibrated via logistic regression, and performance was evaluated against amyloid PET (UC Berkeley) using 10x5-fold repeated cross-validation. ResultsVSD 4-channel fusion achieved AUC = 0.900 ({+/-}0.018), exceeding pTau217 alone (0.888{+/-}0.022). Optimal sensitivity was 89.7% with 78.1% specificity (NPV = 90.8%). The NfL channel received a negative weight ({beta} = -1.1), functioning as a disease-exclusion signal: elevated neurodegeneration without amyloid-tau coupling actively reduces the AD probability, distinguishing AD from non-AD neurodegeneration. Complementary CSF proteomics analysis (7,008 proteins, 533 participants) identified 17 amyloid-specific proteins (0.24% of the proteome), revealing a 49:1 tau-to-amyloid asymmetry that explains why blood-based tau markers outperform amyloid markers. ConclusionsBlood-based VSD provides an interpretable, multi-channel framework for amyloid detection that incorporates explicit disease-exclusion logic unavailable to single-biomarker approaches. The architecture extends to multi-disease screening, where the same blood specimen could be routed through disease-specific modules for AD, Parkinsons disease, and cancer.

BackgroundThe ATN (Amyloid/Tau/Neurodegeneration) framework provides a theory-driven approach to Alzheimers disease (AD) classification using binary biomarker cutoffs, while unsupervised machine learning offers data-driven phenotyping. The concordance between these approaches in population-representative samples remains incompletely characterized. ObjectiveTo compare plasma ATN classification with data-driven clustering methods and evaluate their associations with cognitive outcomes in a nationally representative cohort. MethodsWe analyzed plasma biomarkers (A{beta}42/40 ratio, p-tau181, NfL, GFAP) from 4,465 participants aged [&ge;]51 years in the Health and Retirement Study 2016 Venous Blood Study. ATN profiles were classified using literature-based cutoffs. We applied k-means clustering, Gaussian mixture modeling, and variational autoencoder (VAE) dimensionality reduction to identify data-driven biomarker-defined subgroups. Agreement between ATN and clustering was quantified using adjusted Rand index (ARI) and normalized mutual information (NMI). Longitudinal analyses examined associations with cognitive decline over 4 years (2016-2020). ResultsThe analytic sample included 4,465 individuals (mean age 69.7{+/-}10.4 years; 58.7% female; 75.8% non-Hispanic White). ATN classification yielded 14 profiles, with A+/T-/N-(27.4%) and A-/T-/N-(22.6%) most prevalent (Figure 2). K-means clustering identified 4 optimal clusters with distinct biomarker signatures. Agreement between ATN and clusters was modest (ARI=0.119, NMI=0.113). Sensitivity analysis excluding GFAP from clustering reduced agreement substantially (ARI=0.03 vs 0.119 with GFAP, -74.5% decrease), demonstrating that GFAP accounts for most of the observed concordance between clustering and ATN classification, with only one-third arising from the shared three biomarkers.[Table S12] Additional sensitivity analyses confirmed that k=4 provides finer biomarker resolution than k=3 by retaining biomarker-extreme subgroups[Table S13], and that Cluster 4 represents a stable biological structure across distance metrics[Table S14] despite its small size. Cluster 1 (n=51, 1.2%) showed severe pathology; Cluster 3 (n=3,479, 78.6%) represented the largest and most heterogeneous group, encompassing the broad spectrum of minimal to moderate pathology across all ATN profiles; Cluster 4 (n=14, 0.3%) represented a small but stable non-AD biomarker-defined subgroup (Jaccard=0.779). The VAE revealed a localized nonlinear structure. Silhouette values in the latent space are not directly comparable to clustering silhouettes, but the VAE embedding showed clearer local separation, whereas PCA explained more variance (67.1%). Both ATN and clusters predicted 4-year cognitive decline (ATN R{superscript 2}=0.024, p<0.001; Clusters R{superscript 2}=0.019, p<0.001). O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=118 SRC="FIGDIR/small/26343331v4_fig2.gif" ALT="Figure 2"> O_LINKSMALLFIG WIDTH=200 HEIGHT=118 SRC="FIGDIR/small/26343331v4_fig2a.gif" ALT="Figure 2"> View larger version (31K): org.highwire.dtl.DTLVardef@16e5f3eorg.highwire.dtl.DTLVardef@12eea70org.highwire.dtl.DTLVardef@1215faborg.highwire.dtl.DTLVardef@febb9_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOFIGURE 2.C_FLOATNO Plasma Biomarker Distributions and ATN Profile Prevalence Panel A: Violin plots showing the distribution of log-transformed plasma biomarkers across the entire sample (N=4,465). Each violin displays the probability density of values, with embedded box plots indicating median (center line), interquartile range (box), and 1.5xIQR (whiskers). From left to right: NfL (log): Median=2.996 (equivalent to 20.0 pg/mL); GFAP (log): Median=5.095 (equivalent to 163.2 pg/mL); A42/40 ratio (log): Median=-2.900 (equivalent to 0.055); p-Tau181 (log): Median=0.788 (equivalent to 2.2 pg/mL). All biomarkers show right-skewed distributions in the original scale, with log-transformation achieving approximate normality suitable for clustering analyses (Q-Q plots in Supplementary Figure S1). Panel B: Bar chart showing the distribution of ATN profiles in the sample. The x-axis displays 14 observed ATN profiles ordered by prevalence; y-axis shows percentage of sample. The most common profile is A+/T-/N-(27.4%, n=1,224), followed by A-/T-/N-(22.6%, n=1,009), A+/T+/N+ (13.0%, n=579), A+/T-/N+ (11.2%, n=500), A-/T+/N+ (9.4%, n=419), A-/T-/N+ (7.3%, n=325), A+/T+/N-(5.1%, n=227), and A-/T+/N-(3.2%, n=144). Rare profiles (A+/NA/N+, A-/NA/N+, A+/NA/N-, NA/T-/N+, NA/T-/NA) each comprise <1% of the sample. Abbreviations: A, amyloid-; GFAP, glial fibrillary acidic protein; IQR, interquartile range; NfL, neurofilament light. C_FIG ConclusionsTheory-driven ATN classification and data-driven biomarker phenotyping capture partially overlapping but largely distinct information. Modest concordance (ARI=0.119) reflects GFAPs contribution to shared structure, with most alignment arising from GFAP rather than from the three ATN biomarkers alone (ARI=0.03). The primary source of discordance remains the binary versus continuous representation of biomarker variation. Sensitivity analyses showed that k=4 provides finer biomarker resolution than k=3, and that Cluster 4 represents a small but reproducible biomarker-defined subgroup. Both approaches predicted cognitive decline with modest effect sizes (R{superscript 2}=1.9-2.4%), consistent with population-based studies. Integrating theory-driven and data-driven frameworks may support a more comprehensive characterization of AD-related pathology in population research.

BackgroundThe mechanisms underlying cognitive resilience to Alzheimers disease (AD) pathology are under active investigation. The association between dietary micronutrients with neuroprotective properties and cognitive resilience is currently unknown. ObjectiveTo investigate whether plasma L-ergothioneine (ET), its metabolite L-hercynine (HC), as well as their ratio (HC:ET, as an index of ET metabolism) affect the association between plasma markers of brain amyloid pathology (phosphorylated tau species) and cognitive decline. Design, setting and participantsThis study consisted of 259 dementia-free participants (mean age [SD] = 72 [8] years; 50% females) recruited from memory clinics and the community in Singapore from August 2010 to July 2019 as part of a larger, ongoing longitudinal cohort study on dementia, with analyses performed in February 2025. MeasurementsAll participants of this study were dementia-free at baseline (median Clinical Dementia Rating-Sum of Boxes (CDR-SB) [IQR] scores = 0 [1]), had blood collected for measurements of plasma ET, HC, HC:ET, p-tau181 and p-tau217, and underwent annual neuropsychological assessments for up to 5 years (mean follow-up [SD] = 52 [15] months). The main cognitive outcomes were cognitive decline, defined by annual change of CDR-SB, and risk of incident cognitive decline, defined as Global CDR scores (CDR-GS) increments of [&ge;] 0.5 at follow-up. Linear regression analyses tested for interaction effects of plasma ET, HC and HC:ET on the association between plasma p-tau and cognitive decline, while Cox proportional hazards models were fitted to estimate hazard ratios (HRs) of incident cognitive decline. ResultsPlasma HC:ET significantly moderated associations between plasma p-tau181 and cognitive decline, whereby only High HC:ET attenuated the detrimental effects of plasma p-tau181 on cognitive decline (High HC:ET {beta} = 0.0976; 95% Confidence Interval [CI] = -0.444 to 0.639 vs. Low HC:ET {beta} = 0.849; 95% CI, 0.318 to 1.38). Among participants with high risk of amyloid pathology, those with Low HC:ET had approximately twofold increased risk of cognitive decline (Hazards ratio [HR] = 1.96; 95% CI = 1.01 to 3.79) compared to participants with High HC:ET (HR=0.87; 95% CI = 0.33 to 2.26) over the follow-up period. ConclusionsIdentification of plasma HC:ET as a biomarker of cognitive resilience to amyloid pathology suggests potential beneficial effects of ET metabolism. Further studies are needed to elucidate ET-mediated neuroprotective mechanisms as potential therapeutic targets in delaying or moderating AD-associated cognitive decline.

Introduction: Participant- and study partner-reported Cognitive Function Index scores may provide complementary information, but it remains unclear whether Alzheimer's disease biomarkers are associated with CFI scores across reporters, reporter-specific imbalance, or both. Methods: Using A4/LEARN screening data (screening sample, N = 1,686; primary dyadic analytic sample, N = 1,682; CDR global score = 0), we jointly modeled participant-reported (CFI-PT) and study partner-reported (CFI-SP) scores in long format to evaluate biomarker associations with CFI scores and biomarker x reporter interactions. As a secondary analysis, we compared tau PET with plasma p-tau217. Results: In an A4/LEARN-adapted regional extent model, reporter balance varied across amyloid regional extent categories, with the largest participant-leading contrast observed in the exploratory restricted early cortical subgroup. Tau PET was associated with higher CFI scores, but this association did not differ detectably between reporters. Plasma p-tau217 showed no clear CFI association or reporter-specific interaction in the A4-derived, amyloid-enriched subset. Discussion: Amyloid regional extent and tau PET were associated with different features of dyadic CFI data: reporter balance and CFI burden across reporters, respectively. Joint interpretation of CFI-PT and CFI-SP may support biomarker-informed interpretation of the instrument, although the cross-sectional findings require replication and longitudinal validation.

INTRODUCTIONThe growing number of assay platforms measuring blood-based biomarkers (BBMs) for Alzheimers disease (AD) has introduced challenges in interpretability and comparability across assays. Differences across studies also limit comparability of data. To address these challenges, a systematic evaluation of harmonization methods is needed to support BBM data integration within or across studies. METHODSTwo multisite studies, Alzheimers Disease Neuroimaging Initiative (ADNI, n = 219) and Human Connectome Project (HCP, n = 111), were used to evaluate harmonization methods for mitigating assay and cohort effects in plasma p-tau217 measurements. Methods includes various normalization, regression, and standardization approaches, including the recently developed CentiMarker. Assay effects were evaluated using repeated-measures data across assay platforms within each cohort, whereas cohort effects were assessed using pooled ADNI and HCP data. Harmonization performance was evaluated using distributional statistics and downstream modeling of p-tau217. RESULTSQuantile normalization and quantile mapping methods were most effective for mitigating assay effects, whereas conditional quantile mapping performed best for pooled multi-cohort data. These methods also preserved biological variability. In contrast, simple means adjustment and reference-based z-score standardization were least effective for mitigating assay effects, while simple means adjustment, z-score standardization, and quantile normalization were least effective for mitigating cohort effects. CentiMarker had minimal impact on assay or cohort effects. DISCUSSIONBased on our evaluation, we recommend (conditional) quantile mapping for p-tau217 studies integrating data across multiple assays or cohorts. In contrast, we caution against using CentiMarker and z-score-based methods, as they limit comparability and do not effectively mitigate technical variability.

INTRODUCTIONIn people at risk of Alzheimers disease (AD), hippocampal hyperactivity relates to worse hippocampus-dependent task performance and promotes AD pathology. Reducing hippocampal hyperactivity promises to improve hippocampus-dependent memory performance. METHODS78 participants (69.29{+/-}5.87 years old, 60% female) used real-time fMRI neurofeedback twice to downregulate the hippocampus or a control region. We assessed hippocampal activity and memory performance before and after neurofeedback. We classified AD pathology risk using a blood-based biomarker, an established risk-score, or clinical characterisation. RESULTSDuring hippocampus, but not during control-region neurofeedback, individuals significantly downregulated hippocampal activity with the effect still detectable one week later. In those at high risk of AD pathology, with all three classifications, stronger hippocampal downregulation resulted in stronger memory improvement, independent of baseline hippocampal hyperactivity. DISCUSSIONDownregulating hippocampal activity using fMRI neurofeedback may be a complementary treatment for people at risk for AD, well before there is hippocampal hyperactivity.

INTRODUCTION: Brain resilience to tau has been mainly studied in relation to grey matter, while its role in white matter remains unclear in Alzheimer's disease (AD). Sex may moderate associations between brain resilience and cognition. METHODS: We analyzed medial temporal lobe tau PET SUVR, entorhinal cortical thickness, cingulum-hippocampal mean diffusivity, and cognition in 205 amyloid-positive individuals from ADNI. Associations between grey- and white-matter resilience to tau and cognitive performance or decline were examined using linear and mixed-effects models, including sex interactions and stratified analyses. RESULTS: Higher grey-matter resilience to tau related to better cross-sectional memory and language performance (p<0.005), whereas higher white-matter resilience related to slower decline in these domains (p<0.03). These associations were seen in men but not women. DISCUSSION: White-matter resilience may better capture early cognitive changes in AD. Sex differences suggest greater sensitivity of cognition to brain resilience in men.

Importance: Broadening access to biomarker-informed risk stratification in mild cognitive impairment (MCI) has become even more critical to early assessment in Alzheimer disease given recent developments in regulatory approvals of disease-modifying therapies and advancements of blood-based biomarkers. This requires accessible approaches that can be deployed at scale to better differentiate the disease biology from the clinical progression risk prediction. While entorhinal tau positron emission tomography (PET) can refine near-term prognostic assessment, the cost and logistic burden of imaging limit broad clinical use. Objective: Evaluate whether a brief informant-reported screen derived from the Functional Activities Questionnaire (FAQ) could better stratify scalable biologically anchored prognostic information for 3-year progression from MCI to Alzheimer disease dementia. The primary study was designed around FAQ-derived screens performance relative to entorhinal tau PET standardized uptake value ratio (SUVR), plasma phosphorylated tau 217 (p-tau217) and Mini-Mental State Examination (MMSE) score. Secondary analyses evaluated the stable FAQ-derived screen selected for clinical risk separation, tau and amyloid PET biological context, additional plasma biomarkers, resource-use scenarios and sensitivity analyses around subgroups, calibration, decision-curve, survival, timing, early-progressor exclusions and endpoint-ascertainment IPW. Design, Setting, and Participants: This retrospective secondary progression risk prediction study analyzed 350 Alzheimer's Disease Neuroimaging Initiative (ADNI) participants with a baseline clinical diagnosis of MCI at the tau PET anchor visit. All studies were conducted in cohorts with 3-year progression status known. The first primary benchmarking included 157 participants (including 32 progressors) for FAQ with entorhinal tau PET SUVR comparisons and 153 participants (including 31 progressors) for FAQ, entorhinal tau PET SUVR and MMSE comparisons. The second primary benchmarking was derived from a smaller UPENN plasma p-tau217 subset of 66 participants (including 13 progressors). Exposures: The FAQ-derived candidate screens were evaluated by leakage-controlled repeated nested cross-validation. The stable 3-item FAQ-derived screen selected was defined as any informant-reported difficulty in at least one of the three activities comprising finances/checkbook, shopping and games/hobbies ("Locked FAQ Trio"). The Locked FAQ Trio was compared against both biological and cognitive comparators: entorhinal tau PET SUVR, plasma p-tau217 and MMSE score. Amyloid PET status and Centiloid burden as well as plasma biomarkers paired per same-file plasma such as A{beta}42/40 ratio, glial fibrillary acidic protein (GFAP), neurofilament light chain (NfL) and a directionally adjusted 4- marker plasma composite were used for biology or exploratory context and not for defining the clinical endpoint. Main Outcomes and Measures: The primary binary endpoint was progression from baseline MCI at the tau PET anchor visit to Alzheimer disease dementia within 3 years. Model performance used the cross-validated area under the receiver operating characteristic curve (AUC), the difference in AUC ({Delta}AUC) was bootstrap 95% confidence intervals (CI) at the participant level with P values adjusted using the Benjamini-Hochberg (BH) procedure. Other measures included Brier scores, calibration summaries, survival discrimination and operating characteristics such as sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and screen-positivity prevalence, while decision-curve analyses and resource-use scenarios remained exploratory. Results: A leakage-controlled nested cross-validation selection repeatedly identified a 3-item screen defined as any difficulty in at least one of the three following activities comprising finances/checkbook, shopping and games/hobbies (Locked FAQ Trio). In an independent 3-year progression benchmark analysis of base-covariate models, the Locked FAQ Trio showed higher numerical, directional but not statistically significant, discrimination than entorhinal tau PET among 157 participants including 32 progressors (AUC, 0.787 vs 0.780; {Delta}AUC, +0.007; 95% CI, -0.099 to 0.113; BH-adjusted P = 0.926) and was statistically significantly higher than MMSE score (AUC, 0.796 vs 0.637; {Delta}AUC, +0.159; 95% CI, 0.045 to 0.276; BH-adjusted P = 0.029). The Locked FAQ Trio was positive in 37.6% of participants and captured 27 of 32 progressors, showing sensitivity of 84.4%, specificity of 74.4%, PPV of 45.8%, and NPV of 94.9%. Progression within 3 years occurred in 45.8% of screen-positive participants versus 5.1% of screen-negative participants and the corresponding adjusted hazard ratio over full follow-up was 7.46. The screen was also associated with higher entorhinal tau burden and remained consistent across survival, timing-sensitive, amyloid and missingness analyses. A different 3-item FAQ-derived companion screen ("Companion FAQ Trio") was evaluated for sensitivity, it was defined as any impairment in at least one of the three activities comprising forms/papers, shopping and remembering appointments/medications/holidays. The Companion FAQ Trio was positive in 54.1% participants and captured 96.9% of progressors, with 36.5% of screen-positive progressing to dementia versus 1.4% of screen-negative. In a second primary benchmark analysis of a smaller matched plasma subset of 66 participants including 13 progressors, plasma p-tau217 showed the highest discrimination (AUC, 0.890) across all single predictors in a base-covariates model, compared with the Locked FAQ Trio (AUC, 0.749) and entorhinal tau PET SUVR (AUC, 0.798). A stratification study of the Locked FAQ Trio combined with p-tau217 showed separation of observed risk, differentiating lower and higher risk of progression per strata. Notably, none (0 of 31) of the participants in the lower risk cohort progressed and 64.3% (9 of 14) of participants in the higher risk cohort progressed. Nevertheless, 37.5% (3 of 8) of participants in the Locked FAQ Trio-negative/p-tau 217-high cohort progressed. This emphasizes that patients should not be excluded from further biomarker testing when clinical concern remains. Conclusion: A brief 3-item stable FAQ-derived screen was identified as a compelling front-end additional layer to prognostic triage in MCI patients. This Locked FAQ Trio screen demonstrated a higher numerical discrimination than entorhinal tau PET SUVR in 3-year base-covariates prediction risk models. Plasma p-tau217 remained the strongest scalable predictor of progression to dementia in a smaller plasma subset. These findings reinforce that adding a brief functional screen to the staged prognosis assessment triage pathway can help prioritize and contextualize biomarker escalation, offering a scalable, deployable, and low burden solution to expand screening to a broader patient population.

IntroductionPhosphorylated tau-217 (p-tau 217) is widely used as a plasma-based biomarker for Alzheimers Disease (AD) detection, demonstrating superior accuracy for detecting brain amyloid pathology. However, 30-50% of patients fall within an intermediate diagnostic "gray zone" where biomarker results are indeterminate, often decreasing physician confidence and requiring subsequent diagnostic workup. To address this, we developed a two-stage machine learning algorithm GRAD: Gatekeeper & Reflex for Alzheimers Disease to increase clinical confidence and reduce the AD health economic burden. MethodsWe initially analyzed 320 participants from the Alzheimers Disease Neuroimaging Initiative (ADNI) with plasma biomarkers and amyloid PET. We then built a two-stage machine learning classifier mimicking real clinical workflow where the stage 1 Gatekeeper used the gold-standard marker: p-tau217 with respective 25%/75% probability thresholds. The stage 2 Reflex step applied Random Forest multi-marker classification (p-tau 217, AB42/40, NFL, GFAP) for difficult-to-diagnose gray zone cases. To ensure statistical robustness, leave-one-out cross-validation with bootstrap confidence intervals was used. We externally validated the GRAD architecture on 1,644 A4 Study participants, with MRI enhancement analysis in 1,044 gray zone cases. To measure cost-effectiveness we compared our GRAD-staged testing to universal PET. ResultsThe models Gatekeeper resolved 55.6% of ADNI cases with 88.8% accuracy (NPV 91.8%, PPV 85.0%). The complete pipeline achieved AUC 0.867 (95% CI: 0.825-0.904), with 80.6% sensitivity, 80.0% specificity, LR+ 4.03, LR-0.24. For the difficult-to-diagnose gray zone cases, the Reflex machine learning model achieved AUC 0.755. In our A4 validation, the predictions correlated strongly with centiloid (r= 0.693). Expanding beyond plasma biomarkers, MRI integration improved gray zone classification from AUC 0.829 to 0.853 (p=0.014). The cost modeling analysis projected a 67% reduction in spending versus the current standard of universal PET. DiscussionOur clinically-staged diagnostic algorithm, GRAD, provides actionable classifications for the majority of patients while routing uncertain cases for additional workup. The GRAD framework offers a practical, cost-effective approach for implementing plasma biomarkers in clinical practice. Future iterations of this framework, with integration of novel biomarkers like MTBR-tau243 present a significant opportunity to alleviate the AD health-economic burden and eliminate expensive but unnecessary diagnostic measures. HighlightsO_LIGRAD: Two-stage "Gatekeeper + Reflex for Alzheimers Disease" algorithm resolves indeterminate plasma p-tau217 or gray zone patients with AUC of 0.755. C_LIO_LIOverall AUC of 0.867 (95% CI: 0.825-0.904) validated via leave-one-out cross-validation C_LIO_LIExternal validation in A4 Study demonstrates strong correlation with amyloid burden (r=0.693) C_LIO_LIMRI volumetric integration provides significant incremental value ({Delta}AUC=+0.025, p=0.014) C_LIO_LIProjected 67-71% cost reduction compared to universal PET screening C_LI Research in ContextO_ST_ABSLiterature ReviewC_ST_ABSWe searched PubMed, Google Scholar, and medRxiv databases for studies up to December 2025 that examined plasma p-tau217 diagnostic accuracy as well as "gray zone" management of patients. While several studies demonstrate area-under-the-curve (AUC) of >0.90, these studies largely compare cognitively normal individuals to those with established AD dementia with maximal biomarker separation [1-6]. The gray zone problem, affecting 30-50% of tested individuals, remains unaddressed in the vast majority of clinical implementation frameworks [7,8]. More recent work has established probability-based interpretation [9], but more cohesive algorithms for gray zone resolution through multi-marker integration remain rare if present. Furthermore, the health economic impacts of such resolution have not been fully established. InterpretationOur two-stage algorithm provides a workflow with clinical implementation potential, analogous to established laboratory medicine (i.e TSH with reflex free T4 testing). By first identifying high-confidence cases through univariate p-tau217 (55.6% resolution at 88.8% accuracy), and then applying multi-marker classification only to uncertain cases, we are able to achieve optimal resource utilization while simultaneously maintaining diagnostic accuracy. The finding that MRI usage provides statistically significant improvement ({Delta}AUC=+0.025) has practical implications given the fact that there is a reasonable level of MRI availability in clinical settings. Future DirectionsWhile this work accomplishes several key priorities, future work is required to validate them in diverse clinical populations. In addition, integration of other plasma markers (ex. MTBR-tau243), development of clinical decision support tools, reimbursement mechanisms, and longitudinal validation for treatment monitoring will be necessary to ensure the appropriate infrastructure exists to support providers and patients. Preliminary evidence suggests that %p-tau217 (the ratio of phosphorylated to total tau-217) and MTBR-tau243, a mass spectrometry-based marker of tau tangle pathology, may substantially improve gray zone classification by capturing complementary aspects of tau biology not reflected in absolute p-tau217 concentrations alone, which is a direction that future technical work should examine further.

INTRODUCTIONWe evaluated the performance of plasma brain-derived (BD)- as well as total-p-Tau181, p-Tau217 and p-Tau231 in detecting beta-amyloid positivity (A{beta}+) and cognitive decline in a Singapore-based cohort of older people with cerebrovascular disease. METHODSBrain amyloid status (A{beta}- [n = 139] vs A{beta}+ [n = 74]) was determined by positron emission tomography (PET) scans. Plasma BD and total p-Tau were measured using NUcleic acid Linked Immuno-Sandwich Assay multiplexing platform (NULISAseq). RESULTSBD-p-Tau217 (area under the curve [AUC] = 0.965) outperformed other BD and total-p-Tau species in detecting PET A{beta}+ (AUC = 0.823-0.937; all p [&le;] 0.008). Using three-range or binary references, BD-p-Tau217 demonstrated high sensitivity (>90%), specificity (>90%), positive (>85%) and negative (>95%) predictive values. BD-p-Tau217-derived High-risk group exhibited faster cognitive decline than the Low-risk group. DISCUSSIONRisk stratification for PET A{beta}+ based on plasma BD-p-Tau217 suggests superior diagnostic and prognostic utility, warranting further assessment.

BackgroundEarly detection of Alzheimers disease (AD) is crucial; however, standard neuropsychological tests often lack sensitivity. Process scores, such as proper name (PN) recall from Logical Memory (LM), may improve the detection of AD-related biomarker positivity. We examined whether baseline PN recall predicted future cerebrospinal fluid (CSF) amyloid (A{beta}42/A{beta}40) and tau (pTau181) status, and whether biomarker status predicted PN recall trajectories. MethodsWe analyzed 271 cognitively unimpaired BIOCARD participants (mean age = 57.3, 60.3% female, mean follow-up = 15.5) using logistic regression and mixed-effects models to examine the associations between PN recall and CSF biomarkers. ResultsHigher baseline PN recall predicted lower amyloid positivity (odds ratio [OR] = 0.72, p = 0.015). Amyloid and tau positivity have been linked to a faster decline in PN. Biomarker-positive participants in the biomarker-negative group lacked practice effects. ConclusionsPN recall predicts future AD biomarker positivity and may enhance early detection of AD-related cognitive decline.

INTRODUCTIONTo address an urgent need for a scalable, accurate blood test for brain amyloid pathology that provides a conclusive result for the greatest number of patients, we developed a multi-analyte algorithmic test combining p-Tau 217 with four other biomarkers. METHODSMultiplexed digital immunoassays measured p-Tau 217, A{beta}42/40, GFAP, and NfL in 730 individuals to establish an algorithm with cutoffs, then in 1,082 individuals with cognitive symptoms from three independent cohorts to identify brain amyloid pathology. RESULTSThe tests algorithmic risk score AUC was 0.92, yielding 90% agreement with amyloid PET and CSF. Positive predictive value was 92% at 55% amyloid prevalence. The multi-marker algorithm reduced the intermediate zone 3-fold to 12% vs. p-Tau 217 alone. Diagnostic performance was similar by race, ethnicity, sex, age, and apoE4 status. DISCUSSIONThe LucentAD Complete multi-analyte blood test demonstrated high clinical validity for brain amyloid pathology detection while substantially reducing inconclusive intermediate results.

INTRODUCTIONAlpha-synuclein (Syn) is the most common co-pathology in Alzheimers disease (AD), yet its role within the amyloid-tau-neurodegeneration (ATN) cascade is unknown. METHODSWe analyzed 636 ADNI participants with CSF Syn seed amplification assay, amyloid PET, regional tau PET (Braak I-VI), structural MRI, and cognitive composites. Interaction models tested whether Syn modifies the amyloid-tau and tau-cognition associations. RESULTSSyn positivity (19.0%) amplified the amyloid-tau association across all Braak stages (meta-temporal interaction {beta} = 0.258, 95% CI 0.104-0.411, p = 0.001), with strongest effects in Braak III-IV. Syn did not modify tau-cognition associations in any domain (all interaction p > 0.18). DISCUSSIONSyn co-pathology selectively amplifies amyloid-driven tau propagation without modifying downstream tau-cognition relationships, identifying a node-specific effect within the ATN cascade with implications for patient stratification. Research in ContextO_ST_ABSSystematic reviewC_ST_ABSWe searched PubMed for studies combining -synuclein seed amplification assays with amyloid and tau PET in Alzheimers disease. One recent study (Franzmeier et al., 2025) demonstrated that -synuclein co-pathology accelerates amyloid-driven tau accumulation. No study has examined whether -synuclein modifies the downstream tau-cognition relationship or assessed regional tau specificity across all Braak stages. InterpretationIn 636 ADNI participants, -synuclein co-pathology amplified the amyloid-tau association across all Braak stages but did not modify tau-cognition relationships. This dissociation identifies -synuclein as a node-specific modifier of the ATN cascade, acting at the amyloid-to-tau transition. Future directionsLongitudinal studies with serial tau PET and -synuclein SAA are needed to establish temporality. Clinical trials should evaluate whether -synuclein stratification improves prediction of anti-amyloid treatment response.

Structured AbstractO_ST_ABSINTRODUCTIONC_ST_ABSAPOE-{varepsilon}4 is the strongest common genetic risk factor for Alzheimers disease (AD), yet many carriers remain cognitively unimpaired into late life. We tested whether a protected-{varepsilon}4-first proteomic approach could identify plasma proteins associated with delayed clinical onset among {varepsilon}4 carriers. METHODSWe analyzed harmonized plasma proteomics from the Global Neurodegeneration Proteomics Consortium. Protected {varepsilon}4 carriers ({varepsilon}3/{varepsilon}4 aged [&ge;]75 years; {varepsilon}4/{varepsilon}4 aged [&ge;]65 years; CDR=0; n=456) were compared with {varepsilon}4 carriers with AD (n=1,096). Protein-wise linear models adjusted for age, sex, {varepsilon}4 dosage, and plasma proteomic principal components. Top signals were integrated with high-confidence loss-of-function burden testing and plasma/CSF Mendelian randomization. RESULTS{varepsilon}4 protected was associated with 721 protein levels. Integrated analyses prioritized proteins linked to {varepsilon}4-modified disease biology, including LILRA5, DBI, BPNT1, PTEN, EPHA1, and PCDH10, and proteins aligned with broader AD-related change, including OMG, SELENOW, VAT1, and TPPP3. TREM2 and ACE were also identified, providing internal biological validation of the approach. DISCUSSIONA protected-{varepsilon}4-first plasma proteomic strategy highlights immune, synaptic, metabolic-stress, and myelin/axonal pathways that may delay AD onset and helps prioritize candidate {varepsilon}4-specific modifiers for prevention-focused therapeutics.

INTRODUCTION: MethylCog is a 29-CpG blood DNA methylation (DNAm) proxy for general cognitive ability (g). Its incremental association with blood biomarkers of Alzheimer's disease and related dementias (ADRD) and prospective cognitive ability remains unclear. METHODS: In the held-out test set from the original MethylCog study, we tested whether MethylCog explained baseline g beyond four ADRD blood biomarkers, and whether it predicted six-year follow-up g beyond baseline g and biomarkers. RESULTS: MethylCog showed a stronger age-adjusted association with baseline g than individual biomarkers (r=.368 vs absolute r=.083-.162). MethylCog added 10.0% variance beyond all four biomarkers cross-sectionally (p<.001) and predicted six-year follow-up g in the biomarker-adjusted model (beta=.108, p=.002). No individual ADRD biomarker independently predicted follow-up g. DISCUSSION: MethylCog may provide cognition-related DNAm information complementary to blood-based ADRD biomarkers.

INTRODUCTIONAccurate antemortem differentiation among dementia etiologies remains challenging, particularly for atypical or mixed clinical presentations. Multiplexed plasma proteomics paired with supervised machine-learning offers a minimally invasive and accessible approach for differential diagnosis. METHODSPlasma from 194 participants was analyzed using the NULISA CNS 120+ plasma biomarker panel. Differentially abundant protein patterns associated with AD, frontotemporal lobar degeneration, Lewy body disease, and vascular disease were identified. These features were used to train supervised XGBoost classifier models. Models were then applied to participants with mild cognitive impairment to generate data-driven predictions of etiology. RESULTSNULISA plasma biomarkers revealed disease-specific protein patterns. XGBoost classifiers differentiated disease etiologies with high specificity. Application of the models to participants with mild cognitive impairment yielded robust etiologic predictions. DISCUSSIONThese results support the feasibility of using multiplexed NULISA plasma proteomics, combined with machine learning, for differential diagnosis of complex neurodegenerative dementia etiologies. HighlightsO_LIMultiplex plasma proteomics revealed distinct protein markers of dementia subtypes C_LIO_LISupervised XGBoost classifiers accurately distinguished each dementia etiology C_LIO_LIModel application to unknown etiologies produced interpretable probability profiles C_LIO_LIThe combined NULISA-machine learning framework demonstrates diagnostic feasibility C_LI Research in ContextO_LISystematic review: The authors reviewed the literature using traditional sources (e.g. PubMed), meeting abstracts and presentations. NULISA technology has been utilized to analyze blood biomarkers in people with neurological diseases and differential protein expression based on clinical diagnosis and presumed etiologies has been observed. These citations are appropriately cited. C_LIO_LIInterpretation: Our findings indicate that machine learning can be used in combination with NULISA technology to improve etiology prediction in people with dementia. C_LIO_LIFuture directions: Future studies using larger, more well-balanced participant cohorts will enable better understanding of the plasma biomarkers and demographic factors that best discriminate between dementia etiologies. C_LI

Background and ObjectivesAlthough amyloid positron emission tomography (PET) and Cerebrospinal fluid (CSF) biomarkers remain the standard for confirming Alzheimers disease (AD) pathology, their use is impractical for screening or routine prognostic assessment. Plasma phosphorylated tau 217 (p-tau217), subjective cognitive concerns, and computerized cognitive testing are non-invasive, scalable, and feasible to implement in large populations. We tested whether these measures independently predict the onset of cognitive impairment and whether combining them improves prognostic accuracy. MethodsWe analyzed 1,071 cognitively unimpaired adults aged 65-85 years from the Anti-Amyloid Treatment in Asymptomatic Alzheimers Disease (A4) trial (amyloid-positive; solanezumab or placebo arms) and the parallel Longitudinal Evaluation of Amyloid Risk and Neurodegeneration (LEARN) cohort (amyloid-negative). At baseline, participants completed plasma p-tau217 measurement, the Cognitive Function Index (CFI), and the Cogstate Computerized Battery (CCB). Over 240 weeks of follow-up, incident impairment was defined as conversion from a Global Clinical Dementia Rating Score (CDR-GS) of 0 to 0.5 or higher. The predictive value of each measure for subsequent decline was examined after adjustment for demographic and genetic covariates. ResultsDuring the follow-up, 365 of 1,071 participants (34.1%) developed cognitive impairment. Higher plasma p-tau217 (per-standard-deviation increase) was associated with higher odds of converting to CDR-GS>0 across all cohorts: A4-Placebo (HR=1.56; 95% CI, 1.37-1.78), A4-Solanezumab (HR=1.46; 95% CI, 1.29-1.65), LEARN (HR=1.25; 95% CI, 1.05-1.48). Similarly, higher CFI predicted incident impairment: A4-Placebo (HR=1.59; 95% CI, 1.42-1.79), A4-Solanezumab (HR=1.67; 95% CI, 1.47-1.91), LEARN (HR=1.37; 95% CI, 1.12-1.68). Lower CCB also conferred higher risk: A4-Placebo (HR=0.76; 95% CI, 0.65-0.91), A4-Solanezumab (HR=0.73; 95% CI, 0.62-0.87), LEARN (HR=0.68; 95% CI, 0.53-0.87). In models including all three predictors, each remained independently associated with progression. ConclusionPlasma p-tau217, subjective cognitive concerns, and computerized cognitive testing each independently predicted progression to cognitive impairment in cognitively unimpaired older adults. Together, these non-invasive and scalable measures provide practical tools for risk stratification years before clinical diagnosis. Combining biological, subjective, and digital markers may support earlier detection in clinical care and enhance efficiency in prevention trial enrollment.