Breast density scales: the metric matters
Bibliographic record
Abstract
OBJECTIVE: Measures of percent mammographic density (PMD) are often categorized using various density scales. The purpose of this study was to examine information loss associated with the use of categorical density scales. METHODS: Baseline PMD was assessed at 1% precision for 2,374 females. The data were used to create 21-category, 4-category and 2-category density scales. R-squared and root mean square error were used to evaluate the effect of categorizing PMD. The area under the receiver operator characteristic curves were compared between cancer risk models employing solely categorical PMD scales and solely baseline PMD for a subset of females (424 cases, 848 controls). RESULTS: R-squared value decreased from 1.00 (1% PMD) to 0.56 (2-category scale), while root mean square error increased from 0.00 (1% PMD) to 10.83 (2-category scale). The area under the receiver operator characteristic curve decreased from 0.64 for a cancer risk model using 1% PMD to 0.58 for a risk model using a 21-category density scale (p < 0.0001), 0.55 for a 4-category Breast Imaging, Reporting and Data System-like scale (p < 0.0001) and 0.50 for a 2-category Breast Imaging, Reporting and Data System-like scale (high vs low) (p < 0.0001). CONCLUSION: Categorizing PMD measures into categorical density scales leads to a significant loss of information. Indeed, a simple high versus low split of PMD using a 50% cut point yields a cancer risk model with no discriminatory power. Advances in knowledge: Use of categorical mammographic density scales rather than continuous percent mammographic density measures leads to significant loss of information. Breast cancer risk models using categorical mammographic density scales perform more poorly than models using continuous PMD measures.
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 | 0.001 |
| 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".