Selection and Interpretation of Biomarker Tests for Alzheimer Disease

Alzheimer disease (AD) is defined by the abnormal accumulation of extracellular amyloid-β (Aβ) plaques and spread of intraneuronal hyperphosphorylated tau tangles in the brain. Historically, a definitive AD diagnosis required autopsy confirmation because clinical diagnoses were discordant with neuropathologic findings in ~15% to 30% of cases.^1,2 Recent advances in biomarker testing now enable accurate in vivo detection of amyloid and tau molecular pathology. These breakthroughs have deepened our understanding of the natural history and clinical heterogeneity of AD, informed new biomarker-based diagnostic criteria,^3,4 and fostered the first disease-modifying therapies targeting core molecular features of AD.^5,6

Biomarkers of AD pathology are now increasingly available and integrated into clinical practice to guide diagnosis and treatment decisions. Clinical AD biomarker testing is only recommended for individuals with objective evidence of cognitive impairment to determine whether AD pathology is present and therefore may be causing or contributing to symptoms.^7-9 Biomarker evidence of amyloid pathology is required for initiation of amyloid-targeting treatments (ATTs) in individuals with suspected early symptomatic AD.^10,11 As use of AD biomarkers increases, the risks of both false-positive and false-negative diagnoses underscore the growing importance of careful biomarker selection and interpretation in routine neurologic care. The Table provides an overview of biomarker testing modalities for AD.

Biomarkers Available in Clinical Practice
Amyloid positron emission tomography (PET), the first Food and Drug Administration (FDA)–approved biomarker test for AD (approval granted in 2012), remains the gold standard for clinical practice and for validation of other biomarkers. Amyloid PET scans are interpreted visually as positive or negative, reflecting high or low burden of amyloid neuritic plaques, with ≥88% sensitivity and specificity for AD neuropathology.^12-14 Strengths of amyloid PET relative to other biomarker tests include high diagnostic accuracy, robust validation against autopsy, and ability to monitor clearance of amyloid pathology by ATTs.⁵ Limitations include the high cost (often >$5000), although Medicare and many other insurances now reimburse amyloid PET testing when used in accordance with appropriate use criteria, especially to determine eligibility for ATTs in symptomatic individuals.⁷ PET also requires nuclear medicine infrastructure and specialists, and allows for testing for only one molecular pathology. 

Tau PET with ¹⁸F-flortaucipir was FDA-approved in 2020 to estimate the density and distribution of tau neurofibrillary tangles, with >85% accuracy.¹⁵ Strengths of tau PET are that it reflects the biologic stage of AD,³ is associated with the clinical symptoms and severity of AD,^16,17 and may stratify benefit from ATTs.⁵ Tau PET is rarely used in the clinic, largely due to high costs, limited coverage, and difficulty obtaining the tracer.

Several cerebrospinal fluid (CSF) biomarker tests for AD were FDA-approved starting in 2022. In AD, CSF concentrations of Aβ42 are reduced and phosphorylated tau at threonine 181 (p-tau181) and total tau (t-tau) are elevated. However, it is the ratio of analytes (Aβ42/Aβ40, p-tau181/Aβ42, or t-tau/Aβ42) rather than single analytes that are most strongly associated with amyloid pathology (~90% agreement) and constitute the FDA-approved tests.¹⁸ Strengths of CSF biomarkers include the option to test for other causes of symptoms (eg, inflammatory, neoplastic, prion, α-synuclein diseases), lower costs, and broad insurance coverage. Limitations include the need for lumbar puncture, which is an invasive procedure with potential risks that needs to be performed by trained personnel.⁸

The first blood-based test for AD was FDA-approved in May 2025, and multiple other laboratory-developed blood tests are clinically available. The accuracy of blood tests varies considerably and test results can be influenced by medical comorbidities, such as renal disease and heart failure, although biomarker ratios may reduce the effects of comorbidities.^19,20 Tests that include p-tau217 consistently demonstrate superior accuracy for amyloid and tau pathology compared with tests for Aβ42/40 or p-tau181.²¹ Some tests have a positive, negative, and intermediate zone; values in the intermediate zone warrant follow-up CSF or PET testing. Compared with other modalities, blood tests are the least invasive and most accessible and cost-effective (typically <$1000). Limitations include inconsistent reimbursement for the cost of the blood test, potential influence of medical comorbidities, and that private insurance typically will not approve ATTs based on blood test results. 

Approach to Selecting Biomarkers
The first step in biomarker selection is determining which modalities are practical and feasible. This includes assessing local availability and expected out-of-pocket costs. For blood tests, clinicians should confirm the accuracy of blood test options and strongly consider only using tests that include p-tau217 (see https://www.alzbiomarkerhub.org/performance-database-tool).

Next, it is important to identify biomarker modalities that align with the individual’s clinical presentation and care goals. For individuals being considered for ATTs, CSF or PET testing is often required for eligibility, although some centers and insurances may accept high-accuracy blood tests. CSF or PET testing is preferred for individuals with comorbidities, such as chronic kidney disease, that may confound blood test results. CSF testing is also preferred in individuals with a broad differential that may include prion disease, autoimmune or paraneoplastic encephalitis, normal-pressure hydrocephalus, or Lewy body disease. CSF testing may not be the optimal modality in some individuals who are on anticoagulation or who have lumbar spine disease or a history of surgical procedures. Blood tests may be ideal for individuals who desire a more certain diagnosis but are not candidates for ATTs. Notably, although positive results on blood tests with lower specificity should be followed by PET or CSF testing, a follow-up test is often not necessary if a high-specificity blood test is used.²²

The rationale for biomarker testing and the available options should be discussed with patients and caregivers, who may have strong preferences. For example, some individuals are unwilling to undergo lumbar puncture due to invasiveness. Some individuals are concerned about out-of-pocket expenses for tests. Clinicians should only proceed with testing if the patient and caregiver believe the test will be beneficial and have identified an option that aligns with their care goals and preferences. 

Interpretation of AD Biomarkers
Whereas biomarkers indicate the likely presence of AD pathology, clinical judgment is required to determine whether AD is the primary cause of cognitive impairment or whether an alternative etiology (eg, copathology, medication side effect) is largely responsible for symptoms. To diagnose symptomatic AD, clinicians must integrate biomarker results with a comprehensive assessment that includes a history, examination, cognitive testing, imaging, and routine laboratory studies. In some cases, the individual has an amnestic syndrome with little evidence of other contributing etiologies and clear-cut positive AD biomarkers, enabling a high-certainty diagnosis. However, many individuals have more complex clinical factors or biomarker results.

Although most amyloid PET scans typically only provide a positive or negative read, quantitative values on the Centiloid scale are becoming more common (a Centiloid score of 0 is typical of healthy individuals; 100 is typical of AD dementia; >25 reliably identifies a moderate to high burden of amyloid neuropathology).^23,24 Notably, many individuals with Centiloid values between 25 and 40 are cognitively unimpaired, and interpretation may be challenging if an individual with cognitive impairment has a Centiloid value in this range. In individuals with more advanced dementia symptoms, this scenario should prompt suspicion that an additional etiology may be present, and AD could be a copathology contributing to cognitive symptoms. 

A common issue with CSF biomarkers is discordance between separate analytes (eg, abnormal Aβ42 but normal p-tau181). It is essential to use ratios (p-tau181/Aβ42), which may adjust for non-AD–related factors that vary between individuals and may affect concentrations of single analytes.^18,25 If the ratio of CSF analytes is borderline (ie, near the test-specific cutoff for positivity), it is recommended to perform an amyloid PET scan to clarify the result.

A variety of blood tests are available, including for Aβ42/40, p-tau181, p-tau217, glial fibrillary acidic protein, and neurofilament light chain. Plasma p-tau217 has the highest associations with amyloid PET, tau PET, brain volumes, and cognitive symptoms; the other analytes have lower performance.²¹ In addition, glial fibrillary acidic protein and neurofilament light chain are not specific to AD, and are also affected by age. If results between analytes are discordant, clinicians should base their assessment of AD pathology primarily on p-tau217 or on tests that include p-tau217 (eg, p-tau217/Aβ42 or %p-tau217). If the p-tau217 value is borderline (ie, in the intermediate zone), CSF or PET testing may be indicated. 

In occasional cases, individuals may have clear-cut positive results by one modality (eg, CSF, blood) and clear-cut negative results by another modality (eg, amyloid PET). In these cases, clinical judgement is even more important to guide diagnosis and clinical care. If available, tau PET can provide helpful information, as it has higher specificity for AD pathology associated with cognitive impairment. 

Performing multiple biomarker tests is burdensome. If initial biomarker results are uncertain but biomarker status affects management (eg, initiation of ATTs), it is reasonable to quickly follow-up uncertain results with a second test to arrive at a more certain diagnosis. However, if biomarker testing is performed primarily for diagnosis without directly affecting management, it is reasonable to accept the indeterminate result and to consider retesting the individual later.

Conclusion
Biomarker testing for AD is likely to become increasingly common, especially because ATTs and more accessible blood tests are now available. The decision about which modality to use can be complex and depends on local availability and patient factors. Interpretation of fluid biomarkers should primarily be based on ratios for CSF tests and p-tau217 for blood tests. A second biomarker test should be performed if results are uncertain and biomarker status may affect clinical management.