The concept of screening is based on the notion that regular examination can reduce the mortality from gastrointestinal cancer, but there is as yet little evidence that this is the case in either Barrett's esophagus or ulcerative colitis. Screening is carried out by regular endoscopy with multiple biopsies in an attempt to detect dysplasia, which has traditionally been regarded as the end point of surveillance and the time when resection should be considered. Considerable departure from these guidelines has occurred. Attempts are increasingly being made to select patients in both of these groups who might be at particular high risk of developing dysplasia because of the presence of nuclear aneuploidy, the assumption being that these patients are at particular risk of developing dysplasia and carcinoma. However, data to support stratification at this level are unavailable. In addition, there is not good agreement as to whether surveillance is indicated at all in some groups of patients such as those with less than total colitis or short-segment Barrett's esophagus. The end point of surveillance has also been questioned, with reluctance to recommend colectomy in ulcerative colitis with the finding of low-grade dysplasia, possibly because the sampling problem is so great that it may be impossible to confirm that diagnosis on repeated endoscopy and biopsy. While better markers than dysplasia may be required to predict patients at highest risk, there are currently no marker other than the development of invasive carcinoma on which the decision to resect can be based.(ABSTRACT TRUNCATED AT 250 WORDS)