Automated abdominal aortic calcification scoring from vertebral fracture assessment images and fall-associated hospitalisations: the Manitoba Bone Mineral Density Registry
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abdominal aortic calcification (AAC), a subclinical measure of cardiovascular disease (CVD) that can be assessed on vertebral fracture assessment (VFA) images during osteoporosis screening, is reported to be a falls risk factor. A limitation to incorporating AAC clinically is that its scoring requires trained experts and is time-consuming. We examined if our machine learning (ML) algorithm for AAC (ML-AAC24) is associated with a higher fall-associated hospitalisation risk in the Manitoba Bone Mineral Density (BMD) Registry. A total of 8565 individuals (94.0% female, age 75.7 ± 6.8 years) who had a BMD and VFA image from DXA between February 2010 and December 2017 were included. ML-AAC24 was categorised based on established categories (ML-AAC24 = low < 2; moderate 2 to < 6; high ≥ 6). Cox proportional hazards models assessed the relationship between ML-AAC24 categories and incident fall-associated hospitalisations obtained from linked health records (mean ± SD follow-up, 3.9 ± 2.2 years). Individuals with moderate (9.6%) and high ML-AAC24 (11.7%) had a greater proportion of fall-associated hospitalisations, compared to those with low ML-AAC24 (6.0%). In age and sex-adjusted models, compared to low ML-AAC24, moderate (HR 1.49, 95% CI 1.24-1.79) and high ML-AAC24 (HR 1.89, 95% CI 1.56-2.28) were associated with greater hazards for a fall-associated hospitalisation. Results were comparable (HR 1.37, 95% CI 1.13-1.65 and HR 1.60, 95% CI 1.31-1.95, respectively) after multivariable adjustment, including prior falls and CVD, as well as medication use. Integrating ML-AAC24 into bone density machine software to identify high risk individuals would opportunistically provide important information on fall and cardiovascular disease risk to clinicians for evaluation and intervention.
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.
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) | 0.000 | 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 it