Creating a semantic lesion database for computer-aided MR mammography
Bibliographic record
Abstract
This work presents the creation of a semantic lesion database which will support research into computer-aided lesion detection (CAD) in breast screening MRI. As an adjunct to conventional X-ray mammography, MR-mammography has become a popular screening tool for women with a high risk of breast cancer because of its high sensitivity in detecting malignancy. To address the needs of research and development into CAD for breast MRI an integrated tool has been designed to collect all lesion related information, conduct quantitative analysis, and then present crucial data to clinicians and researchers. A lesion database is an essential component of this system as it provides a link between the DICOM database of MR images and the meta-information contained in the Electronic Patient Record. The patient history, radiology reports from MRI screening visits and pathology reports are all collected, dissected, and stored in a hierarchical structure in the database. Moreover, internal links between pathology specimens and the location of the corresponding lesion in the image are established allowing diagnostic information to be displayed alongside the relevant images. If "ground truth" for an imaging visit can be established either by biopsy or by 2-year follow-up, then the case is labeled as suitable for use in training and testing CAD algorithms. At present a total of 1882 lesions (benign/malignant), 200 pathology specimens over 405 subjects and 1794 screening (455 CAD studies) are included in the database. As well as providing an excellent resource for CAD development this also has potential applications in resident radiologists' training and education.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.005 | 0.003 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.003 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.007 | 0.005 |
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 source (direct Gemma or distilled Codex), 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".