Substituting Blood-Based Biomarkers for Imaging Measures in Alzheimer’s Disease Studies: Implications for Sample Size and Bias
Bibliographic record
Abstract
Background: Blood-based biomarkers for Alzheimer's disease (AD) pathology are appealing options in large population-based studies due to their low cost, minimal invasiveness, and feasibility of collection in non-clinical settings. Despite these benefits, blood-based biomarkers have lower test-retest reliability than neuroimaging measures like amyloid positron emission tomography (amyloid-PET) Centiloids; trade-offs in power and bias remain unexplored. Methods: We use data from Alzheimer's Disease Neuroimaging Initiative (ADNI) and the Anti-Amyloid Treatment in Asymptomatic Alzheimer's Disease (A4) studies, which include both amyloid-PET and blood-based measures, to assess differences in statistical power, required sample size, and bias when replacing a neuroimaging measure with a blood-based measure. We use simulations parameterized based on these studies to show potential implications of using plasma p-tau181 or p-tau217, blood-based AD biomarkers, in place of Centiloids from amyloid-PET, when the biomarker is either the exposure or the outcome in an analysis of interest. Results: We demonstrated that substituting amyloid-PET Centiloids with a blood-based measure of p-tau can substantially reduce power, requiring 3 to 6 times the sample size to achieve 80% power compared to amyloid-PET. In addition, using a blood-based biomarker as the exposure can introduce significant regression dilution bias, attenuating estimated associations. Conclusions: Due to their lower cost and ease of collection compared with neuroimaging, blood-based biomarkers facilitate AD pathology measures on larger, more diverse samples with longitudinal follow-up. Consideration of the sample sizes they necessitate and their potential for bias is critical for the design and interpretation of studies employing these biomarkers.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.238 | 0.470 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.003 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.001 | 0.005 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.003 | 0.003 |
| Research integrity | 0.003 | 0.003 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".