9 March 2017 Denoised ordered subset statistically penalized algebraic reconstruction technique (DOS-SPART) in digital breast tomosynthesis
Author Affiliations +
Abstract
Digital breast tomosynthesis (DBT) is a three dimensional (3D) breast imaging modality in which projections are acquired over a limited angular span around the compressed breast and reconstructed into image slices parallel to the detector. DBT has been shown to help alleviate the breast tissue overlapping issues of two dimensional (2D) mammography. Since the overlapping tissues may simulate cancer masses or obscure true cancers, this improvement is critically important for improved breast cancer screening and diagnosis. In this work, a model-based image reconstruction method is presented to show that spatial resolution in DBT volumes can be maintained while dose is reduced using the presented method when compared to that of a state-of-the-art commercial reconstruction technique. Spatial resolution was measured in phantom images and subjectively in a clinical dataset. Noise characteristics were explored in a cadaver study. In both the quantitative and subjective results the image sharpness was maintained and overall image quality was maintained at reduced doses when the model-based iterative reconstruction was used to reconstruct the volumes.
© (2017) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
John Garrett, John Garrett, Yinsheng Li, Yinsheng Li, Ke Li, Ke Li, Guang-Hong Chen, Guang-Hong Chen, } "Denoised ordered subset statistically penalized algebraic reconstruction technique (DOS-SPART) in digital breast tomosynthesis", Proc. SPIE 10132, Medical Imaging 2017: Physics of Medical Imaging, 101324F (9 March 2017); doi: 10.1117/12.2255584; https://doi.org/10.1117/12.2255584
PROCEEDINGS
7 PAGES


SHARE
Back to Top