Further results on the evaluation of the performance of a digital breast tomosynthesis system in the clinical environment

Abstract

Introduction Digital breast tomosynthesis (DBT) is spreading throughout medical centres worldwide. However, there is a wide variability of technical characteristics among DBT systems that need to be assessed for quality control (QC) implementation. Purpose After evaluating the technical performance of three commercial DBT systems in a previous work, we pursue to research a fourth system with notorious differences with the previous ones. Materials and methods We evaluate a General Electric SenoClaire DBT system with angular range (AR) of 25°, 9 projections, step-and-shoot X-ray tube motion and iterative reconstruction. Tests for assessing the image quality of projections and reconstructed planes were: contrast-to-noise ratio (CNR); spatial resolution, via the in-plane modulation transfer function (MTF) and the in-depth point spread function (Z-PSF), and artefact spread function (ASF). All was implemented in custom-designed phantoms according to literature. Results In the reconstructed slices and slabs, CNR has a very weak dependency on PMMA thickness, as opposed to projections, possibly due to the iterative algorithm. The tube travel MTF is 30% larger respect to the measured average from continuous tube movement systems, due to the step-and-shoot setup, while the chest-wall nipple resolution remains similar. The Z-PSF is narrower than systems with larger AR due to the iterative reconstruction but the ASF is similar to systems with a smaller AR, because of a greater angular spacing between projections. Conclusion We observe in the clinic some theoretical results when comparing with previously studied DBT systems, and the results could be useful as a guide for future QC procedures. Disclosure None to disclose.