Development of a novel iterative image reconstruction algorithm for low-dose chest tomosynthesis

Abstract

Computed Tomography (CT) which provides 3D anatomical image by x-ray attenuation is developed in the direction to reduce dose to patients. Digital Tomosynthesis (DTS) is similar to CT but projects the patients only in the limited angle. Instead of reducing the dose, it makes incomplete anatomical image. DTS is mostly used for diagnosing the breast cancer, and also for obtaining the image of chest region. DTS image reconstructed by analytical Filtered Backprojection (FBP) reconstruction algorithm is contaminated by severe artifacts and noise if sampling is not enough. By solving this problem, if the image quality is maintained although the amount of data is reduced, there is an advantage to reduce dose to patients. For that, we developed the iterative reconstruction method based on Expectation Maximization (EM) algorithm and compared with FBP.