P2‐252: AN AMYLOID LIGAND‐FREE OPTICAL RETINAL IMAGING METHOD TO PREDICT CEREBRAL AMYLOID PET STATUS
Bibliographic record
Abstract
A simple, low-cost approach to identify amyloid positive subjects at or before the earliest stages of cognitive impairment could dramatically impact clinical trials evaluating disease modifying treatments for Alzheimer's disease (AD) by greatly reducing cerebral PET amyloid imaging-related expenses, and could also be clinically useful for screening purposes. In this pilot study, a non-invasive retina (an extension of the central nervous system) imaging approach with the Metabolic Hyperspectral Retinal Camera requiring no amyloid-labeling agent is evaluated as a mean to identify biomarkers correlating with the cerebral load of amyloid plaques determined with PET imaging. The cohort (n=40) included probable AD (n=15) and age-matched controls (53 to 85 years) with no concomitant retinal diseases nor significant ocular media opacity. Hyperspectral retinal measurements were obtained at 450-900 nm. Image analysis based on texture of the spatial/spectral dimensions in segmented retinal vascular areas allowed extraction of 16 different statistical measures. A classifier was trained using 102 datasets (1-3 per subject) to establish the predictive value of those texture features, based on the cerebral amyloid status determined from binary reads by an expert rater on 18F-Florbetaben PET studies. A leave-one-out approach determined the sensitivity and specificity values of the method. Other vascular metrics are also evaluated in the retinal images for possible correlation with the cerebral amyloid status. At least one good quality hyperspectral dataset was acquired in the vast majority (n=40) of the participants enrolled in the study (n=42). Excellent retinal scanning correspondence with PET amyloid status was achieved, independently of cognition, when texture features extracted from the principal retinal vessels were used, with estimated sensitivity and specificity values of 85% and 93% respectively. The developed machine learning approach, based on a non-invasive hyperspectral retinal imaging technique which does not require amyloid labeling, shows promise in predicting cerebral amyloid PET status and could serve as a screening tool to identify subjects in the early stages of the AD continuum, for instance in a drug development context. The study is ongoing to test this approach and other retinal measures in more subjects.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 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; a candidate call from one teacher head, not a consensus.
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".