Image Quality Improvement of Breast Specimen Imaging on Cone Beam CT Using Iterative Reconstruction: A Phantom Study

Breast specimen imaging from cone-beam computed tomography (CBCT) becomes attractive due to fast acquisition time. However, as well known, X-ray scattering effect and noise in CBCT are still the main factors which affect image quality, especially the contrast in reconstructed images. In this study, we propose image quality improvement for CBCT imaging of a breast specimen. Here, X-ray scattering signals embedded in raw projection data from CBCT are eliminated by using a deconvolution technique. Then the corrected projection data are reconstructed by the proposed iterative reconstruction technique to reduce noisy images. In the experiment, the simulated microcalcification phantom and the contrast phantom are used for evaluating image quality and compared with cross-section images from filtered back-projection reconstruction. The image quality results from the proposed technique show that cupping artifacts and noise are reduced and the contrast is significantly improved. Thus, this work has the potential to improve image quality of breast specimen imaging on CBCT.