Evaluating the Effect of Tomosynthesis Acquisition Parameters on Image Texture: A Study Based on an Anthropomorphic Breast Tissue Software Model

Mammographic texture features have been shown to correlate with the risk of developing breast cancer. Digital breast tomosynthesis (DBT) is an emerging 3D x-ray breast imaging modality with superior tissue visualization compared to mammography, having the potential to provide more accurate estimation of parenchymal texture features. In this paper, we investigate the effect of DBT acquisition parameters on computer-extracted texture features. DBT images were simulated using an anthropomorphic breast tissue software model allowing for variations in DBT acquisition geometry. Our results show that DBT acquisition geometry appears to have an impact on the computed texture features; angular range appears to have a greater effect than the selected number of source projections. We attribute this effect to the differences in image quality resulting from the different reconstruction geometries. Our ultimate goal is to determine the DBT acquisition geometry that provides the optimal image quality to estimate parenchymal texture.