Abstract PO3-07-08: Breast density estimation with a microwave-frequency imaging system
Bibliographic record
Abstract
Abstract Background: Microwave imaging has been proposed as an alternative method of breast imaging that is low-cost and comfortable for women as it avoids excessive compression. Microwave properties of tissues relate to water content and behavior (Gabriel et al, 1996); specifically, fatty tissues have lower properties and glandular tissues have greater properties (Lazebnik et al, 2007). These differences in microwave signatures of fatty and glandular tissues provide the opportunity to map the composition of the breast and create a density score without a mammogram. This density score may find utility in risk stratification or monitoring interventions aimed at decreasing breast density (Salazar et a, 2020). Purpose: We examine the feasibility of developing a density score based on microwave images that correlates to mammographic breast density (VOLPARA score) in a pilot study with healthy volunteers. Imaging System: We have developed a microwave imaging system that facilitates scanning of large groups of patients, as well as comparison to x-ray mammography (Mojabi et al, 2023). The system consists of two plates which are placed in contact with the breast. Microwave transmitters and receivers are embedded in the plates; signals transmitted through the breast are used to estimate microwave frequency properties of tissues, and maps of these estimates form a 2D image. Methods: 50 patients provided informed consent (study approved by Health Ethics Research Board of Alberta CC-21-0082). Both breasts of each volunteer were scanned. Previously performed mammograms were available for 21 of the volunteers. The number of volunteers with VOLPARA scores A, B, C, and D is 2, 8, 6, and 5, respectively. The percent density reported with the VOLPARA score is also available for these volunteers. Microwave images were formed for each scan and analyzed to predict density with three approaches: (1) average permittivity, (2) segmented regions, and (3) pixel-based intensities. The results demonstrate that the average permittivity of the breast typically increases from VOLPARA A to D, with some overlap in average values observed between the density categories. A correlation between average permittivity and percent breast density was observed. Regions representing glandular tissues are segmented from microwave images; the average values of these regions clearly differentiate between VOLPARA A and D, however do not show consistent ranges for VOLPARA scores B and C. Finally, the microwave breast density estimated using pixel-based intensities shows good correlation with the percentage density calculations from mammograms. Conclusions: Microwave images contain features related to the glandular tissues embedded in fat. By analyzing the composition of the breast in these images, density scores are created. While average permittivity appears to correlate to percent density calculated from mammograms, area or pixel-based approaches appear to have greater potential for categorizing into density classes. Expanding the number of participants, identifying biomarkers, and exploring deep learning techniques for density prediction are considered for future work Citation Format: Elise Fear, Jeremie Bourqui, Pedram Mojabi, Bobbie-Jo Docktor, Anita Garland, Danielle Deutscher, Zahra Lasemiimeni, Kathleen McMahon, Brendon Besler, Roger Tsang. Breast density estimation with a microwave-frequency imaging system [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO3-07-08.
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 | 0.001 | 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 | 0.001 |
| 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".