Gastrointestinal (GI) tract examination with CT and MRI is currently performed by slice- based visual inspection despite the volumetric nature of the anatomy and of the imaging modalities. Parcellation, visualization and quantification of major GI tract components (stomach, duodenum, small bowel, colon and rectum) is sought to improve diagnostic performance and intervention assessment. The entire abdomen may now be continuously scanned within a single breath-hold with spiral CT, which allows retrospective overlapping reconstruction for improved longitudinal resolution. The investigators seek to extract and unravel the GI tract keyed to anatomic landmarks: gastro-esophageal junction, pylorus, ligament of Treitz, iliocecal valve and anus. The GI tract lumen will be explicitly mapped onto an elongated planar display to analyze lumen and wall caliber, surface features and their locations relative to the anatomic landmarks. The goal is to develop a GI tract spiral CT image unraveling system and optimize the protocols to segment, unravel and quantify the GI tract and associated tumors/lesions.
Specific aims are: (1) to separate the contrast-filled GI tract from spiral CT abdominal image volumes while enhancing image resolution and reducing image noise, (2) to unravel the GI tract from the gastro-esophageal junction to the anus and map it into a calibrated linear display with an emphasis on unraveling the large intestine, (3) to quantify features of normal and abnormal bowel, especially to analyze focal mucosal lesions such as adenocarcinoma of the colon, and (4) to validate the mapping and quantification via numerical stimulation, phantom, animal and cadaver experiments, and correlation with pathologic descriptions of surgically obtained specimens and results of endoscopic and other GI imaging procedures performed in vivo. Since the GI tract is highly convoluted, the GI tract image unraveling system will eliminate visualization difficulties, enable quantitative analyses and provide a framework for computer-assisted diagnosis.
Showing the most recent 10 out of 13 publications
