In my frequent journeys to the radiology suite, I am struck by the wonders of technology and marvel at how we as gastroenterologists as well as our patients have been the beneficiaries. One such tangible example is the advancements over the last decade in CT and magnetic resonance (MR) imaging. These modalities have unquestionably made us better diagnosticians. As endoscopy has made parallel strides for intraluminal imaging, these radiological studies have, to date, been most helpful for extraluminal imaging. CT is a superb modality for evaluation of the liver and pancreas as well as for detection of intraabdominal lymphadenopathy and inflammatory diseases of the bowel. MR, although not used as frequently for abdominal imaging, has recently been touted for its ability to noninvasively image the pancreaticobiliary tree. Its role for routine abdominal imaging, however, remains limited. As pointed out in the two articles by Horton and Fishman and by Anderson, CT and MR have also made great progress in the evaluation of the vascular system. Heretofore, quality imaging of the intraabdominal vasculature could only be done by conventional angiography. CT angiography (CTA) is performed by administering a large intravenous bolus of contrast, followed immediately by CT scanning during the early arterial phase. MR angiography (MRA) also uses contrast agents to highlight the vascular tree. With both techniques, image reconstruction using a variety of software-enhanced techniques can further delineate the vascular tree. The radiographic images provided in these two reviews are persuasive and illustrate the great potential for noninvasive imaging of the abdominal vasculature. As clinical and radiological practice evolves, I foresee that these studies will become the first tests ordered to evaluate patients with suspected intestinal ischemia or other abdominal vascular abnormalities. Already, CTA has supplanted conventional angiography as the preferred initial imaging modality in many institutions. As technological advancements continue, resulting in further improvements in resolution of the vascular tree, could they play a role in the diagnosis of the difficult patient with GI bleeding? Because the vast majority of patients with upper and lower GI bleeding can be diagnosed and treated effectively by the endoscopic approach, we must ask the question whether these imaging techniques may fit into our diagnostic armamentarium, and, if so, in what capacity? Currently, I see little role for CTA or MRA in the overwhelming majority of patients with upper GI bleeding. In the rare case in which bleeding is significant but the diagnosis remains in doubt after endoscopy, it is possible that CTA or MRA may be able to demonstrate vascular malformations (e.g., Dieulafoy’s lesion) or occult varices that were not appreciated endoscopically, thereby supplanting the need for diagnostic angiography. The problem GI bleeders are those who present with recurrent acute lower GI bleeding in whom colonoscopy (and upper endoscopy) detects no lesion, as well as those patients in whom persistent blood obscures colonoscopic visualization, prohibiting a definitive diagnosis, but the nuclear tagged red cell scan also is negative or nondiagnostic. A rapid purge, as advocated by Jensen and coworkers,1 can help define the cause of colonic bleeding and direct colonoscopic hemostatic therapy. This technique has not been universally embraced, although I have found it helpful. If the endoscopist cannot find the bleeding site, an accurate, easily performed imaging technique is needed that will find the bleeding lesion, allowing directed therapy by the surgeon or angiographer. Another problem patient is the occult/obscure bleeder, i.e., the patient with either iron deficiency anemia requiring continued transfusions or recurrent but self-limited melena or hematochezia in whom no etiology can be established despite upper endoscopy, colonoscopy, enteroscopy, red cell scanning, and angiography. Small bowel bleeding is often suspected in such patients, but is exceedingly difficult to prove, in part due to the low rate of blood loss.