Colorectal cancer (CRC) is the second leading cause of cancer death in the US, and early detection and removal of polyps can reduce mortality. Computed tomography colonography (CTC) showed promising early detection sensitivities of up to 90%, but poorer results in screening patient populations. In addition, reading times for CTC remain long, requiring up to 30 minutes to read the approximately 400 to 800 images. Our goals are to provide improved computer graphics and computer aided diagnosis technology and to establish interpretation practices that will maximize polyp detection accuracy while minimizing inter-observer variability and interpretation time. We believe that optimized CTC can (1) detect colon polyps > 5 mm in diameter with a sensitivity equivalent to fiberoptic colonoscopy and (2) achieve this in 5 minutes or less of a radiologist's interpretation time.
Our Specific Aims are as follows:
Aim 1 : Develop and Validate Improved 3D Visualization Methods. We will enable rapid, accurate viewing of the entire colon surface on a single, flattened 3D image and develop new methods to renders tagging material (oral contrast) transparent.
Aim 2 : Develop and Validate Advanced CAD Algorithms. We will develop CAD algorithms that operate on flattened colon data and in the presence of tagging material. We will further advance CAD by developing methods to detect normal haustral folds to discriminate these from polyps, and that enable accurate supine-prone data registration.
Aim 3 : Optimize Radiologists' Interactions with new visualization methods and advanced CAD. We will measure radiologist viewing accuracy, efficiency, and inter-observer variability for the new visualization methods in the absence and presence of our advanced CAD techniques. The results of these efforts will enable the widespread deployment of CTC, which has the potential to improve patient screening, to increase detection of potential cancers, and to reduce morbidity and death from colorectal carcinoma.
Showing the most recent 10 out of 24 publications