Dose reduction in breast imaging: A phantom study

This paper proposed a method for does reduction in breast imaging. While mammogram is a gold standard for early detection of breast cancer, it still limits to dense breast women. So, an additional investigation using a Digital Breast Tomosynthesis (DBT) is then required. This makes the total dose of radiation that patients received is approximately double. We propose a method to reduce the radiation does by generating a synthesized 2D mammographic image from DBT data using Super-Resolution (SR) technique. In this paper, the motion estimation algorithm is performed within a region of interest (ROI) and then the estimated motion parameters are used to reconstruct a SR image. The synthesized 2D mammographic image is the result of SR reconstruction step. We evaluated the method using the Peak signal to noise ratio (PSNR) and mean structure similarity (MSSIM). The results are 17.2383 and 0.9976 respectively. In additions, we also evaluated the result using the phantom passing score. Our synthesized 2D mammographic image got 5, 3, 4 for fibers, group of microcalcifications and masses, respectively.