(Tuesday, 7/14/2020) 11:30 AM - 12:30 PM [Eastern Time (GMT-4)]
Room: Track 1
Purpose: To assess the image quality of tomosynthesis mammography systems used at sites participating in a large, ongoing screening trial, to ensure consistent quality control (QC) and understand typical machine performance for each model.
Methods: In addition to regular weekly testing, participating sites run extended tests yearly using a modular phantom. The phantom consists of a series of semi-circular (breast shaped) polymethyl methacrylate (PMMA) slabs that can be configured to different thicknesses, along with various inserts and polystyrene spacers. The inserts include a layer with an embedded a polyethylene sphere for signal difference to noise ratio (SDNR) measurement, and a grid of 16 aluminum BBs (0.03125 mm diameter) used to assess the artefact spread function (ASF). The SDNR insert is used with different total thicknesses of PMMA and the polystyrene spacers to assess the automatic exposure control. The ASF insert is positioned at different heights within the phantom to assess the geometric fidelity of the reconstructed volume. For this report, basic statistics (mean and standard deviation) are calculated for each model of equipment in the study.
Results: Four models of equipment from three vendors are being used. To date, there are 173 thickness tracking tests, with 680 SDNR measurements on 141 units. The COVs for the SDNR measurements for each thickness configuration and model of unit range between 5% and 16%. Average doses for the 4.5 cm configuration range between 1.1 and 2.0 mGy.
There are 1353 ASF volumes from 139 units, generating 21648 ASF measurements in x, y and z directions. The full-width half maximum COVs range between 3-7% for x, 2-9% for y and 4-8% for z directions.
Conclusion: The collected data is promising and will form the basis for establishing pass-fail thresholds for the SDNR and ASF tests.
Funding Support, Disclosures, and Conflict of Interest: The authors of this work have a research collaboration agreement with GE Healthcare in the area of tomosynthesis mammography. TMIST is publicly funded by the National Cancer Institute (NCI), part of the National Institues of Health, thorugh the NCI Community Oncology Research Program.
Keywords
Quality Control, Tomosynthesis, Clinical Trials
Taxonomy
IM- Breast X-Ray Imaging: Quality Control and Image Quality Assessment
Contact Email
