Aili K. Maki Bloomquist BEng, Presenter: Research collaboration, General Electric Company

James G. Mainprize PhD, Abstract Co-Author: Institutional research agreement, General Electric Company

Martin J. Yaffe PhD, Abstract Co-Author: Research collaboration, General Electric Company Founder, Matakina International Ltd Shareholder, Matakina International Ltd Co-founder, Mammographic Physics Inc

Gordon Mawdsley, Abstract Co-Author: Manager, Medical Physics Incorporated Research collaboration, General Electric Company

Sam Zhongmin Shen, Abstract Co-Author: Research agreement, General Electric Company

Digital breast tomosynthesis (DBT) is evolving steadily, with units from several vendors available in Europe and numerous clinical trials ongoing or recently published. A quality control (QC) program is being developed for use in the proposed ACRIN screening trial, TMIST. Harmonized QC, with tests and phantoms applicable across different platforms, will help ensure consistent image quality.

A draft manual has been written for a DBT quality control program. Phantoms and automated analysis software have been developed. The daily QC phantom includes a sphere of contrasting material for measurement of signal difference to noise ratio, an aluminum BB for assessment of the artefact spread function (ASF) in x, y and z directions and groups of microcalcifications at different heights for evaluation of geometric reconstruction accuracy (location within the volume). A uniform region within the phantom is used to calculate an in-slice 2D normalized NPS. Technologist's tests include imaging the QC phantom so that these parameters can be tracked with the software, as well as assessment of the review workstation for lag. Physicist's tests include evaluation of collimation, NEQ, and performance of the automatic exposure control as well as estimation of mean glandular dose.

All routine technologists tests were performed on a GE Senoclaire system (not yet FDA approved), and results tracked over time. Radially averaged NNPS at 1, 2 and 4 mm-1 and SDNR values were found to be stable over time (COVs <=5%). The FWHM of the ASF in the x, y and z directions was found to be less stable, warranting further investigation. Some of the tests were performed on both a GE Senoclaire and a Hologic Selenia Dimensions. FWHM values of the ASF in the x and y directions were comparable (0.5-0.7 mm). FWHM in the z direction was 3.3 mm for the Senoclaire and 7.9 mm for the Selenia.

A QC program for DBT has been designed, including phantoms and analysis software. The tests characterize key parameters of the quality of the reconstructed volume. Further work is required to set suitable pass/fail thresholds that capture clinically relevant changes in image quality. The appropriate frequencies for performance of the different tests must also be determined.

A platform-independent QC program is fundamental to ensuring consistent clinical image quality for DBT.

Maki Bloomquist, A, Mainprize, J, Yaffe, M, Mawdsley, G, Shen, S, Development of a Quality Control Program for Digital Breast Tomosynthesis.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14046029.html