Comparison of the diagnostic accuracy of stationary digital breast tomosynthesis to digital mammography with respect to lesion characterization in breast tissue biopsy specimens: a preliminary study

Abstract

Current practice for imaging surgical breast specimens is a single 2D magnification view on a mammography system, but 2D imaging overlaps the tissue in different planes causing distortion of lesion margins. Digital breast tomosynthesis (DBT) could be used as an alternative imaging modality for imaging breast specimens. DBT systems acquire multiple low dose projection images, over a small angular span, which are then reconstructed into a partial 3D volume. The reconstructed images can be used to increase visualization of lesion margins and extent of microcalcifications (MCs). Current commercial DBT systems use a single rotating X-ray source, the movement of which produces motion blur. Motion blur reduces visualization of small objects such as MCs. MCs, depending on size and structure, can be implicative of breast cancer. We have developed a stationary DBT (s-DBT) system using a linearly distributed, CNT Xray source array. S-DBT allows for rapid acquisition of projection images with no image degradation from X-ray source motion. Full tomosynthesis datasets can be acquired, allowing visualize of both masses and microcalcifications. Here we report the preliminary results of a reader study comparing breast specimen images from a 2D commercial mammography system and an s-DBT system. Preliminary results show that s-DBT is capable of producing equivalent image quality to 2D mammography, and in some cases is superior.