Evaluation of a retinal deep phenotyping platform to detect the likely cerebral amyloid PET status in humans
Bibliographic record
Abstract
Abstract Background Translating recent advances in the diagnosis of Alzheimer’s disease (AD) and other dementias to clinical practice using biomarker‐based approaches requires accessible, affordable biomarkers. This protocol tested the performance of a Retinal Deep Phenotyping platform for the detection of likely positron‐emission tomography (PET) amyloid status (negative or positive) in human subjects. This platform generates data rich retinal reflectance images captured non‐invasively with a Metabolic Hyperspectral Retinal Camera (MHRC) yielding specific spatial‐spectral features which are analyzed by a machine learning algorithm relatively to a recognized gold standard biomarker (in this case PET amyloid imaging). Methods 134 subjects (50 years and older) from 3 clinical sites were included, 94 with normal, 40 with abnormal cognition, imaged with the MHRC and amyloid PET (unanimous visual reads from a panel of 3 expert reviewers; 42 amyloid positive, 92 negative). Over 600 spatial‐spectral features were extracted from hyperspectral retinal images obtained at 450‐900 nm. After a feature selection based on F‐score ranking, a classifier was trained on the datasets from 87 patients (2/3 of the cohort) using the most discriminating features. The classification performance was then evaluated with the datasets from the remaining 47 patients. Results Resubstituting the training set of 87 patients into the classifier yielded sensitivity and specificity values of 88% and 91%, respectively. The training misclassification rate was predicted using the k ‐fold loss metric, with a predicted error rate of 19%. To verify this performance, the test set of 47 patients was classified by the model resulting in sensitivity and specificity values of 80% and 81%, respectively. Conclusions The Retinal Deep Phenotyping platform shows promise for detecting the likely cerebral amyloid PET status in human subjects and could serve as a screening tool to identify subjects in the AD continuum, for instance in a clinical or drug development context. The platform could also be used for the detection of other biomarkers involved in cognitive decline.
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) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| 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) | 0.000 | 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".