Biomarkers have fundamentally transformed the diagnosis and management of Alzheimer's Disease (AD), shifting the field from a purely clinical syndromic classification to a biological definition of the disease. The 2018 NIA-AA Research Framework formalized the AT(N) biomarker system, which defines AD by the presence of amyloid-beta pathology (A), tau] pathology (T), and neurodegeneration (N), independent of clinical symptoms Jack et al., 2018. This framework was further updated in 2024 by the Alzheimer's Association Workgroup, which expanded the biomarker categories to include inflammation/astrocyte reactivity markers and established Core 1 and Core 2 biomarker tiers to guide clinical and research use Jack et al., 2024. These advances have been accelerated by the approval of anti-amyloid therapies such as lecanemab and donanemab, which require biomarker-confirmed amyloid pathology for treatment initiation [2].
The AT(N) framework classifies individuals along three pathological axes:
The framework enables biological staging that tracks disease progression independent of symptoms:
This staging enables enrollment of biomarker-defined populations in prevention trials and guides treatment decisions for anti-amyloid therapies [3].
Cerebrospinal fluid biomarkers remain the gold standard for biochemical detection of AD pathology. CSF is in direct contact with brain interstitial fluid, providing a window into central nervous system biochemistry [4].
| CSF Biomarker | What It Measures | Typical AD Change | Diagnostic AUC | Key Clinical Utility |
|---|---|---|---|---|
| Abeta42 | Fibrillar amyloid plaque burden (inversely) | Decreased (~50% of normal) | 0.85-0.90 | Earliest biochemical change; detectable 15-20 years before symptoms |
| Abeta42/40 ratio | Amyloid pathology (normalized for total Aβ production) | Decreased (cutoff ~0.062) | 0.90-0.95 | Corrects for interindividual variation; superior to Abeta42 alone |
| p-tau181 | Tau phosphorylation at threonine 181 | Increased (2-3 fold) | 0.90-0.93 | Well-validated; widely available; tracks with Braak staging |
| p-tau217 | Tau phosphorylation at threonine 217 | Increased (4-8 fold) | 0.94-0.97 | Highest diagnostic accuracy; best correlation with tau PET; distinguishes AD from other tauopathies |
| p-tau231 | Tau phosphorylation at threonine 231 | Increased | 0.88-0.92 | May detect very early changes before p-tau181 elevation |
| t-tau | Total tau (neuronal injury marker) | Increased (3-fold) | 0.85-0.90 | Less specific; elevated in stroke, CJD, and other acute injuries |
| NfL | Axonal degeneration (neurofilament light chain) | Increased | 0.75-0.85 | Non-specific neurodegeneration marker; prognostic value; tracks treatment response |
CSF Abeta42 begins to decline approximately 15-20 years before symptom onset, making it the earliest detectable biochemical change. CSF p-tau rises approximately 10-15 years before symptoms, followed by t-tau and NfL closer to clinical onset Palmqvist et al., 2020 [5].
amyloid PET imaging provides direct in vivo visualization of fibrillar amyloid-beta plaque burden. Three F-18-labeled tracers are FDA-approved for clinical use:
All three tracers have been standardized to the Centiloid scale, enabling cross-tracer comparisons. A Centiloid value above 20-25 typically indicates amyloid positivity. Notably, 20-30% of cognitively normal older adults are amyloid PET-positive, reflecting preclinical AD Palmqvist et al., 2020 [6].
Tau PET tracers enable in vivo visualization and quantification of neurofibrillary tangle pathology (Jack et al., 2018):
| Feature | CSF Biomarkers | Blood Biomarkers | amyloid PET | Tau PET |
|---|---|---|---|---|
| Invasiveness | Lumbar puncture | Blood draw | IV injection + scan | IV injection + scan |
| Cost | $500-1,500 | $200-800 | $3,000-6,000 | $3,000-6,000 |
| Accessibility | Moderate (requires LP expertise) | High (any clinic) | Low (specialized centers) | Low (specialized centers) |
| Earliest detection | Abeta42 changes 15-20 yrs pre-symptom | p-tau217 changes ~10 yrs pre-symptom | Detects plaques ~10-15 yrs pre-symptom | Detects tangles ~5-10 yrs pre-symptom |
| Quantitative precision | High | Moderate-High | High (Centiloid scale) | High (SUVr) |
| Spatial information | None | None | Regional plaque distribution | Regional tangle distribution (Braak staging) |
| Monitoring utility | Good for treatment response | Good for screening and serial monitoring | Good for amyloid clearance assessment | Best for tracking clinical progression |
| Regulatory status | Established clinical use | Emerging (first FDA clearance 2025) | FDA-approved (3 tracers) | FDA-approved (flortaucipir) |
Blood-based biomarkers represent the most transformative recent advance in AD diagnostics, offering scalable, minimally invasive screening that can be deployed in primary care settings Teunissen et al., 2024 [7].
Plasma p-tau217 has emerged as the single best-performing blood biomarker for AD. In the largest prospective study to date (n = 1,767 across 5 European cohorts), the Lumipulse G plasma pTau217/Abeta1-42 ratio achieved concordance rates of 91.7% (positive) and 97.3% (negative) with amyloid PET and CSF biomarkers Palmqvist et al., 2024. On May 16, 2025, the FDA granted its first clearance for a blood-based AD biomarker test: the Lumipulse G pTau217/Abeta1-42 Plasma Ratio, approved for symptomatic patients aged 55 and older [8].
The PrecivityAD2 test (C2N Diagnostics) combines the percent p-tau217 (%p-tau217) ratio with plasma Abeta42/40, achieving an AUC of 0.95 for amyloid PET positivity. This test has received regulatory approval in the UK and is under FDA review in the US [9].
The plasma Abeta42/40 ratio shows moderate accuracy for amyloid detection alone (AUC 0.75-0.80) but improves substantially when combined with p-tau markers. In Japan, the HISCL Abeta42/40 assay has been approved for clinical use [10].
Glial fibrillary acidic protein (GFAP is an astrocytic marker elevated in AD plasma that reflects reactive astrocytosis. Plasma GFAP rises approximately 10 years before symptom onset, potentially earlier than p-tau changes, and correlates with amyloid plaque burden. The 2024 AA revised criteria include GFAP as an inflammation/astrocyte reactivity ("I") category biomarker Pelkmans et al., 2024. GFAP may mediate the relationship between amyloid deposition and downstream tau pathology, as significant correlations between brain amyloid burden and tau are observed exclusively in individuals with abnormal plasma GFAP levels [11].
Plasma NfL (neurofilament light chain) correlates strongly with CSF NfL (r > 0.80) and is elevated in AD, though it is not disease-specific. Serial plasma NfL measurements track disease progression and may detect treatment effects earlier than clinical measures, making it valuable for clinical trial monitoring Mattsson et al., 2017 [12].
The approval of anti-amyloid monoclonal antibodies has made biomarker testing a clinical necessity:
A two-cutoff strategy for blood biomarkers is emerging: patients with clearly positive or clearly negative results are triaged directly, while those in the intermediate zone proceed to confirmatory CSF or PET testing. This approach reduces the need for expensive imaging by approximately 50% while maintaining diagnostic accuracy above 90% Palmqvist et al., 2024 [13].
Several emerging biomarkers are expanding the AT(N) framework:
The rapid translation of blood biomarkers into clinical practice marks one of the most significant advances in neurology. By enabling earlier and more accurate diagnosis, these biomarkers support timely intervention with disease-modifying therapies at a stage when preservation of neuronal function is still achievable [14].
The study of Biomarkers Of Alzheimer's Disease has evolved significantly over the past decades. Research in this area has revealed important insights into the underlying mechanisms of neurodegeneration and continues to drive therapeutic development [1].
Historical context and key discoveries in this field have shaped our current understanding and will continue to guide future research directions [2].
Recent evidence reinforces plasma p-tau217-centered workflows as practical gatekeepers for preclinical detection, prognosis, and differential diagnosis in Alzheimer's Disease.
🟡 Moderate Confidence
| Dimension | Score |
|---|---|
| Supporting Studies | 20 references |
| Replication | 33% |
| Effect Sizes | 75% |
| Contradicting Evidence | 0% |
| Mechanistic Completeness | 50% |
Overall Confidence: 56%