Real-world patterns of Alzheimer’s biomarker testing: Insights from a large-scale clinical dataset

Authors: David Alfego, PhD & Joseph Volpe

Alzheimer’s disease evaluation has evolved rapidly with the adoption of blood-based biomarkers. Biomarker-based frameworks include ATN profiling, tau markers, amyloid ratios, and genetic risk indicators such as APOE ε4. Previously, biomarkers were evaluated in cerebrospinal fluid (CSF), which requires a lumbar puncture/spinal tap to acquire a sample. The advent of blood biomarkers allows for a much less invasive testing process. Eventual CSF may be needed for confirmation, but if blood-based testing is negative, it helps reduce the need for follow-on CSF testing. Despite these diagnostic advances, much of the existing evidence on biomarker use is derived from clinical studies or narrowly defined cohorts.

Real-world data offers a complementary perspective by capturing how clinicians apply biomarker testing in everyday practice—across patient populations, clinical contexts and care pathways. Understanding these real-world patterns is critical for:

• Interpreting biomarker utilization at scale
• Informing evidence generation strategies
• Supporting regulatory, payer and scientific decision-making
• Developing and updating diagnostic guidelines

Data source and methods

This exploratory data analysis (EDA) utilized Labcorp clinical testing data through December 31, 2025. The dataset included patients with at least one Alzheimer’s-related biomarker test.

Order	N, patients (%)*
p-Tau217	98,616 (42.1%)
ATN Profile	83,948 (35.8%)
Beta Amyloid 42/40 Ratio  (Plasma/Serum)	65,831 (28.1%)
APOE AD Risk	56,411 (24.1%)
p-Tau181	30,240 (12.9%)
Beta Amyloid 42/40 Ratio  (CSF)	4,416 (1.9%)

Table 1. Data source overview and study scope

* As of end of 2025, testing not mutually exclusive. Excluding non-AD-specific biomarkers, like Glial Fibrillary Acid Protein (GFAP) and direct ordering of Neurofilament Light Chain (NfL)                              

Analyses focused on

• Initial biomarker testing patterns
• Sequencing and flow of subsequent tests
• Co-testing behavior within defined time windows
• Longitudinal biomarker changes among patients with repeat testing
• Alignment with guideline-relevant diagnostic practices

Cohort overview

• 234,450 patients had at least one Alzheimer’s biomarker test performed
• Biomarker tests included the ATN Profile, p-Tau217, p-Tau181, beta-amyloid 42/40 ratio, NfL, GFAP and APOE testing
• Testing volume reflects broad Labcorp market penetration rather than curated research enrollment

Figure 1 - Overall Alzheimer’s biomarker testing volume

	ATN Profile	P-Tau217 (Direct)	Beta Amyloid 42/40 Ratio, plasma/serum	p-Tau181
N	83,948	98,616	70,015	30,240
Age, mean std	73.2 ± 10.5	73.4 ± 10.4	72.5 ± 10.7	72.7 ± 10.6
N, female (%)	49,052 (58.4%)	56,704 (57.5%)	40,917 (58.4%)	17,530 (58.0%)
Result, mean ± std	Beta Amy.: 0.11 ± 0.04 p-tau181: 1.44 ± 1.34 NfL: 4.9 ± 11.3	0.41 ± 0.7	0.11 ± 0.02	1.42 ± 1.30
N, Result (%)
Normal	Beta Amy.: 53,728 (64.1%) p-tau181: 32,489 (38.8%) NfL: 72,093 (85.9%)	39,386 (40.3%)	42,405 (65.8%)	11,863 (39.4%)
Abnormal	Beta Amy.: 29,825 (35.6%) p-tau181: 51,059 (60.9%) NfL: 11,569 (13.8%)	58,432 (59.7%)	21,963 (34.1%)	18,216 (60.6%)
Incomplete	Beta Amy.: 219 (0.3%) p-tau181: 227 (0.3%) NfL: 286 (0.3%)	23 (0.0%)	289 (0.1%)	3 (0.0%)
N, with at least a second  biomarker (any time during patient journey) (%)	18,627 (22.2%)	20,657 (20.9%)	19,176 (27.4%)	4,406 (14.6%)

Figure 2. Summary statistics by AD biomarker order

This scale enables robust analysis of real-world diagnostic behavior across heterogeneous patient populations

Real-world biomarker testing pathways

Initial testing patterns

Analysis of first biomarker orders in Figure 3 revealed heterogeneity in starting points, with certain biomarkers—such as the ATN Profile and p-Tau217—frequently serving as entry points into the diagnostic journey.

These findings underscore that real-world diagnostic pathways are non-linear and can vary substantially from structured research protocols. In clinical practice, p-Tau217 is typically ordered by neurologists when symptomology is clear. By contrast, ATN is a framework that, by design, serves as a comprehensive assessment that can indicate various forms of dementia, including but not limited to Alzheimer’s. ATN is more often ordered in the primary care setting; it is likely that primary care providers are more familiar with ATN and utilize it to determine diagnostic direction when symptomology is less conclusive. Another contributing factor to the ATN Profile being a popular entry point is its earlier commercial introduction to the market. In addition to symptomology and specialty area, the usage of biomarkers continues to evolve based on the rapid discovery, development and clinical data to support the use of these biomarkers.

Figure 3. Common starting points for AD biomarker testing

Testing flow after initial biomarker assessment

Patients initiating testing with ATN Profiles demonstrated distinct downstream testing patterns compared to p-Tau217-first testing pathways.

These flows illustrate how clinicians adapt diagnostic strategies based on early results, reinforcing the need for longitudinal RWD to understand care decisions over time.

Figure 4. Testing flow after ATN Profile as first biomarker order

Figure 5. Testing flow after p-Tau217 Profile as first biomarker order

Longitudinal biomarker dynamics

Among patients with repeat testing, changes in biomarker values over time were observed, enabling exploration of:

• Disease monitoring practices
• Diagnostic refinement
• Real-world application of emerging biomarker guidance

Such longitudinal visibility is rarely available in traditional study designs.

	N, at least two tests	1st Result, Mean ± Std	Mean ± Std	Median [Q3, Q4]	Reference Range	Improving (>)  or  worsening (<)
pTau-217	2,715	0.42 ± 0.47	0.0091 ± 0.23	0.010 [-0.040,0.060]	0.00 – 0.18	<
pTau-181	4,504	1.46 ± 1.48	-0.032 ± 1.17	-0.010 [-0.18, 0.15]	0.00 – 0.97	>
Amyloid Beta 42/40 Ratio	5,324	0.11 ± 0.04	0.0022 ± 0.042	0.0020 [-0.0040, 0.010]	>0.102	>
NfL	3,316	5.1 ± 25.0	-0.083 ± 9.91	-0.070 [-0.54, 0.36]	0.00-3.65	>
GFAP	1,654	71.6 ± 177.7	0.67 ± 127.9	3.5 [-5.0, 12.8]	0.00 – 65.80	<

Figure 6. Changes in Alzheimer’s biomarkers among patients with repeat testing

Co-testing and diagnostic context

ATN Profile co-testing

An analysis of ATN Profile orders revealed frequent co-testing within 90 days, including metabolic, inflammatory and endocrine markers (e.g., A1c, TSH, Vitamin B12, Vitamin D).

This suggests that clinicians often evaluate Alzheimer’s biomarkers within a broader clinical context, reflecting real-world diagnostic complexity rather than isolated biomarker use. For example, clinicians often perform endocrine marker testing prior to investigating Alzheimer’s as rule out testing; it has been established that B12 deficiency or imbalanced thyroid function can present with dementia-like symptoms.

APOE genetic testing

Only 27.3% of patients with ATN testing also received APOE genetic testing, highlighting variability in genetic risk assessment and potential gaps between emerging research and routine clinical practice. Of note, out of the patients who had ATN and APOE testing, 34.9% were heterozygous for ε4 and 5.8% were homozygous for ε4, providing additional insight into varying genomic profiles. APOE testing trends may also be explained by highly specific payer reimbursement policies.

Implications for Alzheimer’s research and evidence generation

These findings demonstrate that large-scale real-world data can provide insights not readily observable in trials, including:

• How Alzheimer’s biomarkers are used and sequenced
• Variability in diagnostic approaches across patient populations
• Opportunities to study guideline adherence and real-world implementation
• Foundations for regulatory-grade RWE and HEOR analyses