More than density: validating a mammographic masking prediction model in Dutch breast cancer screening
Bibliographic record
Abstract
OBJECTIVES: To validate a lesion masking prediction model, Mammatus, previously developed on a North American cohort, on a larger retrospective breast cancer screening cohort from a single center in the Netherlands. MATERIALS AND METHODS: Mammatus was applied to all digital mammography screening examinations with a unilateral invasive breast cancer that was either diagnosed at screening or within 24 months after a negative screening, called interval cancers. All mammograms were retrospectively evaluated for the visibility of malignant masses using all available imaging and clinical information. The area under the receiver operator characteristic (ROC) curve (AUC) when using Mammatus to distinguish examinations with screen-detected cancers (assumed low masking risk) from interval cancers (assumed high masking risk) was computed. The AUC was compared to that of the original cohort and to that obtained using volumetric breast density (VBD) as a predictor. A second tghree-category ROC analysis was performed, with interval cancers that were retrospectively visible classified as intermediate lesion masking. RESULTS: Mammatus achieved an AUC of 0.69 (95% CI: 0.66-0.73) for distinguishing between screen-detected-cancer exams (n = 635) and interval-cancer exams (n = 304). This performance did not differ from the original study (AUC = 0.75 (95% CI: 0.68-0.82), p = 0.15), and outperformed VBD (AUC = 0.66 (95% CI: 0.63-0.70, p = 0.019). Mammatus was better at identifying mammograms at low risk of lesion masking (AUC = 0.73 (95% CI: 0.70-0.76)) compared to those with high risk (AUC = 0.69 (95% CI: 0.64-0.74)). CONCLUSION: Mammatus performed well in predicting breast cancer-masking risk in a Dutch screening cohort. This suggests that adding information other than density facilitates the prediction of lesion masking. KEY POINTS: Question Mammographic lesion masking prediction models, such as Mammatus, require external validation in other screening programs before clinical application is possible. Findings Mammatus maintained similar performance in predicting lesion masking in a Dutch screening cohort and showed added benefit compared to VBD. Clinical relevance An externally validated lesion masking prediction model for digital mammography could potentially be used to identify screened women who could benefit from supplemental or alternative screening, with better accuracy than VBD alone.
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 | 0.001 | 0.001 |
| 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".