IC‐P‐057: Automatic Anatomical Structure Segmentation Predicts Conversion from MCI to Alzheimer's Disease
Bibliographic record
Abstract
Volumetry of certain structures has demonstrated discriminative power to predict conversion from MCI to AD. We have previously developed tools to automatically segment hippocampus (HC) [Collins, ICAD09], lateral ventricles [Fonov, ISMRM2010] and other anatomical structures from MRI data [Collins, Human Brain Mapping 1995;3:190-208]. We applied these tools to a subset of baseline MCI data from the Alzheimer's Disease Neuroimaging Initiative (ADNI). T1w 3D MPRAGE MRI data for 200 MCI subjects were randomly selected from the baseline scans of the ADNI database. By 12 months, follow-up data was available for 163 of the 200 subjects, where 36 of these (22% of followup) converted to AD (the remained stable or reverted to normal). By 24 months, follow-up data was available for 136 subjects of the 200 original MCI subjects, where 58 subjects (42.3% of followup) had converted to AD and 69 remained stable or reverted to normal. Our goal was to predict which MCI subjects would convert to AD using only the baseline MRI data. MRI data were corrected for image intensity non-uniformity [Sled, Trans Med Imag 1998;17:87-97], stereotaxically transformed [Collins, J Comput Assist Tomog 1994;18:192-205] and resampled onto a 1mm3 grid. ANIMAL+fusion was used to segment HC on the left and right sides and to segment lateral ventricle. Our ANIMAL model-based, automatic structure segmentation procedure was used to segment 36 other anatomical structures. All volumes were entered as features into a linear discriminant analysis (LDA). Using 10 optimally selected variables yields 73.9% accuracy, 69% sensitivity and 74% specificity for prediction of conversion at 12m. Important discriminatory variables included hippocampus, lateral ventricle, caudate, frontal white matter (WM) and parietal WM. (Noise in brain masking reduced discriminatory power of cortical structures). Prediction accuracy at 24m dropped to 65%. Automatic structure segmentation is robust and accurate and does not suffer from inter-rater variability. The structure volumes can be combine to create a linear discriminant function that predicts progression with high accuracy. Such a tool will enable enrichment of clinical trials and could be used for patient selection for early treatment.
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 | 0.000 |
| 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) | 0.003 | 0.001 |
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".