IC‐P‐159: Voxel‐Wise Logistic Regression Improves Prediction Accuracy for Developing AD
Bibliographic record
Abstract
Predicting clinical trajectories in populations at risk, permit enrich populations for disease modifying clinical trials. Recent developments in machine learning techniques have enabled us to achieve highly accurate predictions in multiple clinical applications. Here we demonstrate a novel data driven method to improve the accuracy of a Random Forest based classifier to predict the development of dementia. [18F]Florbetapir images were acquired for 275 MCI individuals from the ADNI cohort and the images were processed using an established PET image processing pipeline. Regional SUVr values were extracted from brain regions such as the Angular Gyrus, Supramarginal Gyrus, Posterior Cingulate Cortex and Precuneus which were identified to have a significant amyloid accumulation based on existing literature. 70%(192) and 30% of the population were labeled as the training set and the testing set, respectively. Data driven method included a Voxel-wise logistic regression analysis using the training population to identify anatomically significant brain regions with highest odds-ratios (ORs) to develop dementia. Two Random Forest based predictors were trained with the regional SUVr values based on literature and the data driven method respectively. Their performances were measured against the testing population. Voxel-wise logistic regression analysis indicated that brain regions including Orbitofrontal Cortex, Mid Frontal Sulci, Mid Temporal Sulci, Temporal Occipital junction, PCC, Angular Gyrus, Precuneus, Putamen and the Nucleus Accumbens have the highest OR values for a unit SD increase of [18F]Florbetapir [Figure 1]. The Random Forest predictor trained using regions based on literature achieved 79% and 78% as validation and testing accuracy (0.89 AUC) while the predictor based on the data driven method achieved 84% for both validation and testing accuracy (0.91 AUC). The Temporal Occipital junction, Mid Temporal Sulci and Mid Frontal Sulci regions indicated the highest contribution in predicting the development of dementia [Figure 2]. The data driven method using Voxel-wise logistic regression analysis have increased the accuracy of the Random Forest predictor and have outperforms the methods developed in previously published literature and can be a utilized as a valuable clinical framework. Regions with the highest OR values. Orbitofrontal Cortex, Mid Frontal Sulci, Mid Temporal Sulci, Temporal Occipital junction, PCC, Angular Gyrus, Precuneus, Putamen and the Nucleus Accumbens are indicated. ROC curves and features importance for the two predictors. a) predictor with the regions based on existing literature b) predictor based on Voxel-wise logistic regression.
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 | 0.000 | 0.001 |
| 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".