A virtual trial evaluation of patient motion in arc and linear system designs in body digital tomosynthesis

Background. Patient movement is an ever-present reality in any clinical imaging scenario. In the context of digital tomosynthesis, thoracic motion can profoundly impact the resulting image quality. Design of the system, based on either linear or multi-directional source positions, can influence this impact. The purpose of this study was to assess this influence through a virtual imaging trial.Methods. We utilized computational human models from the 4D extended cardiac-torso (XCAT) series, incorporating specific pathologies in the chest and abdomen, as well as four types of patient motions: cardiac, respiratory, bowel gas, and incidental movements. The models were imaged using representative simulators of digital tomosynthesis (DTS), with both a conventional linear motion system and a multi-source dome system (also known as an arc) featuring multi-directional source positions. The images were analyzed quantitatively and visually through an observer study to assess the relative impact of motions in the resulting images.Results. The multi-source arc system demonstrated an overall advantage over the conventional linear tomosynthesis system in terms of reduced susceptibility to motion artifacts. Observer scores indicated better image quality (by 35%) and lesion conspicuity (by 54%) with the arc system. Quantitative metrics indicated up to a two-fold increase in contrast, CNR, and SSIM for the arc system.Conclusions. The findings suggest that the arc geometry may offer improved robustness against patient motion compared to linear tomosynthesis. This advantage is attributed to the arc system's ability to capture images from multiple angles in rapid sequence, thereby minimizing the impact of motion artifacts and enhancing overall image quality.