Lubomir M. Hadjiiski PhD, Presenter: Nothing to Disclose

Heang-Ping Chan PhD, Abstract Co-Author: Nothing to Disclose

Elaine M. Caoili MD, Abstract Co-Author: Nothing to Disclose

Richard H. Cohan MD, Abstract Co-Author: Consultant, General Electric Company Research funded, General Electric Company

Hyun-Chong Cho, Abstract Co-Author: Nothing to Disclose

Chuan Zhou PhD, Abstract Co-Author: Nothing to Disclose

Jun Wei PhD, Abstract Co-Author: Nothing to Disclose

To develop a computerized method for segmentation of ureters in CT Urography (CTU) scans for a computer-aided detection and characterization system for ureter cancer.

IRB approval was obtained. Our system tracks the contrast-filled lumen of ureters in CTU. It consists of three stages: (1) adaptive thresholding and region growing, (2) propagation and tracking, and (3) edge profile extraction. In the first stage, from a starting point, adaptive thresholding and region growing are applied to the local region to extract the contrast-filled structure based on the CT numbers and voxel connectivity. In the second stage, a prediction of the location of the next search region is performed and the segmentation process is propagated. In the third stage, an edge profile is extracted based on gradient enhancement and analysis in the local region. The segmentation performance was evaluated on a data set including 114 ureters on 74 CTU scans from 74 patients collected retrospectively. The 114 ureters were filled with IV contrast material, with moderate to good level of opacification. On average the ureter occupied 286 CT slices (range:164 to399, median: 301). More than half of the ureters contained malignant or benign lesions and some had malignant ureter wall thickening. The lesions were marked by experienced radiologists using graphic user interface (GUI). In this preliminary study, an initial starting point for each of the 114 ureters was selected with the GUI, which served as an input to the automated tracking and segmentation system. The segmentation performance was quantitatively assessed by estimating the percentage of the length that was successfully tracked and segmented for each ureter.

Of the 114 ureteres, 61 (54%) were segmented completely (100%), 80 (70%) were segmented through at least 70% of its length, and 96 (84%) were segmented at least 50%. The system was able to track across the ureter lesions and wall thickening and revealed the narrowing of the lumen.

Our preliminary study demonstrates the feasibility of using automated method for segmenting ureters in CTU scans. Further study is underway to improve the segmentation performance with a larger data set.

Ureter segmentation is a crucial step in CAD systems for detection and characterization of ureter cancer and ureter wall thickening. This study demonstrated a useful method for ureter segmentation.

Hadjiiski, L, Chan, H, Caoili, E, Cohan, R, Cho, H, Zhou, C, Wei, J, Automatic Segmentation of Ureters on Multidetector Row CT Urography.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11034590.html