Investigation of Different PV Distributions in Digital Breast Tomosynthesis (DBT) Mammography

Abstract

Digital Breast Tomosynthesis (DBT) Mammography is an emerging technique that has the potential to improve breast cancer detection. In DBT, low-dose mammograms are acquired at a number of projection angles over a limited range and the 3D breast volume is reconstructed. In this study, we investigated the effect of different distributions of projection-view (PV) images that included different angular range and angular spacing on the reconstruction image quality. A GE prototype DBT system was used to acquire a total of 21 PVs in 3º increments over a ±30º range, from which multiple subsets containing the same number of 11 PV images were selected. A custom-built breast phantom and a selected patient case were used to evaluate the image quality. For breast phantom study, the contrast-to-noise ratio (CNR), the normalized line profiles of test objects, and an artifact spread function (ASF) were used as performance measures to compare the results for the different subsets and for the full set. The simultaneous algebraic reconstruction technique (SART) was used to reconstruct the DBT under all conditions. Our results demonstrated that large DBT angular range gave superior CNR and ASF for masses and less interplane blurring for high-density objects. Narrow angular range favors in-plane edge sharpness for high-density objects.