Kenny Heekon Cha MSc, Presenter: Nothing to Disclose

Lubomir M. Hadjiiski PhD, Abstract Co-Author: Nothing to Disclose

Heang-Ping Chan PhD, Abstract Co-Author: Institutional research collaboration, General Electric Company

Richard H. Cohan MD, Abstract Co-Author: Consultant, General Electric Company Consultant, Medscape, LLC

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

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

To develop a computer-aided diagnosis system for bladder lesion detection in CT urography, which potentially can assist radiologists in detecting bladder cancer.

Initially, the bladder was automatically segmented by our previously developed Conjoint Level set Analysis and Segmentation System (CLASS). In this preliminary study, we focused on detecting mass lesions within the contrast-enhanced (C) region of the bladder as a prescreening step. The C region was delineated from the segmented bladders using a method based on maximum intensity projection. The bladder wall of the C region was extracted by using adaptive thresholding to remove the contrast material, and transformed into a profile of wall thickness normal to the wall surface. The morphology and voxel intensity along the profile were analyzed and suspicious locations were labeled as lesion candidates. With IRB approval, a data set of 70 patients with 102 biopsy-proven bladder lesions within the C region was collected. All lesions were marked by experienced radiologists in the CTU volumes as reference standard and rated by their conspicuity. The cases were split evenly into independent training and test sets. The training set contained 30 subjects having 37 malignant and 9 benign lesions with average size of 20.1 mm (range: 4.2–61.7 mm). The test set contained 33 subjects having 47 malignant and 9 benign lesions with average size of 18.8 mm (range: 1.4–61.1 mm). The average lesion conspicuity rating in both sets was 2.2 (scale 1 to 5, 5 very subtle).

Our system detected 78% (36/46) of the bladder lesions with 3.5 (123/35) false positives per patient in the training set, and 77% (43/56) of bladder lesions with 4.4 (155/35) false positives per patient in the test set. The false negatives were mainly caused by the non-uniformity of the contrast material, camouflaging the lesions as a part of the bladder wall.

Our study demonstrates the feasibility of our method for detection of bladder lesions within the contrast-enhanced region of the CTU for lesions of a variety of shapes and sizes. Further work is underway to increase the sensitivity and reduce the false positives, and to detect lesions in the entire bladder.

Early detection of bladder cancer is crucial for improved patient survival. This study shows a CAD system useful for automatic bladder cancer detection within the contrast-enhanced region of CTU.

Cha, K, Hadjiiski, L, Chan, H, Cohan, R, Caoili, E, Wei, J, Automatic Detection of Bladder Mass Lesions within Contrast Enhanced Region in CTU.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14045671.html