WE-G-103-03: BEST IN PHYSICS (IMAGING) - The Feasibility of An X-Ray Differential Phase Contrast Tomosynthesis System Adapted From a Clinical Digital Breast Tomosynthesis System

Abstract

Purpose: This work aims to investigate the feasibility of a grating‐based x‐ray differential phase contrast (DPC) tomosynthesis imaging system based on an absorption‐contrast DBT system currently available in the clinical practice. The main significance of this study is the potential improvement in lesion/calcification detection performance when the DPC mechanism is combined with the tomosynthesis imaging method. Methods: First, a framework was developed to simultaneously reconstruct differential phase contrast (edge information), phase contrast (electron density information), and absorption contrast images. Second, a physical phantom and a benchtop DPC imaging system (40 kVp, 80 micron pixel pitch) were used to evaluate the reconstruction framework as well as the contrast improvement with the DPC mechanism. Third, a DPC system was designed based on the current hardware of a clinical absorption DBT system (Hologic Selenia Dimensions), and the task‐based model observer detectability indices of the DPC tomosynthesis system were evaluated using a theoretical framework that quantitatively relates the noise properties of DPC‐DBT with absorption DBT. Results: Reconstructions of physical phantoms show improved signal difference to noise ratio (SDNR) compared with absorption images acquired under the same exposure (SDNR_PMMA = 5.9 and 0.6 for DPC and absorption, respectively). Equivalent spatial resolution for the two contrast mechanisms was observed from the line profiles and artifact spread functions of a bead. Design parameters of the DPC‐DBT system are fully compatible with the current clinical system. The accuracy of the framework that predicts detectability in DPC‐DBT was validated experimentally, and it suggests that the DPC mechanism will Result in improved detectabilities of both small objects (e.g. calcification) and irregular‐shaped objects (e.g. spiculated lesions). Conclusion: It is feasible to build a DPC tomosynthesis system using the system setup of an existing clinical DBT system. The system shows promise in improving lesion and calcification detectability, and therefore merits further investigation.