Integration of 3D scale-based pseudo-enhancement correction and
partial volume image segmentation for improving electronic colon cleansing in
CT colonograpy
Affiliations: Department of Radiology, State University of New York
at Stony Brook, Stony Brook, NY, USA | Department of Biomedical Engineering, State University
of New York at Stony Brook, Stony Brook, NY, USA | Department of Engineering Science and Physics, City
University of New York at College of Staten Island, Staten Island, NY,
USA
Note: [] Corresponding author: Zhengrong Liang, Department of Radiology,
State University of New York at Stony Brook, NY 11794, USA. E-mail:
jerome.liang@sunysb.edu
Abstract: Orally administered tagging agents are usually used in CT
colonography (CTC) to differentiate residual bowel content from native colonic
structures. However, the high-density contrast agents tend to introduce
pseudo-enhancement (PE) effect on neighboring soft tissues and elevate their
observed CT attenuation value toward that of the tagged materials (TMs), which
may result in an excessive electronic colon cleansing (ECC) since the
pseudo-enhanced soft tissues are incorrectly identified as TMs. To address this
issue, we integrated a 3D scale-based PE correction into our previous ECC
pipeline based on the maximum a posteriori expectation-maximization partial
volume (PV) segmentation. The newly proposed ECC scheme takes into account both
the PE and PV effects that commonly appear in CTC images. We evaluated the new
scheme on 40 patient CTC scans, both qualitatively through display of
segmentation results, and quantitatively through radiologists' blind scoring
(human observer) and computer-aided detection (CAD) of colon polyps (computer
observer). Performance of the presented algorithm has shown consistent
improvements over our previous ECC pipeline, especially for the detection of
small polyps submerged in the contrast agents. The CAD results of polyp
detection showed that 4 more submerged polyps were detected for our new ECC
scheme over the previous one.
