This article provides a framework for clinicians who are attempting the difficult task of interpreting the Barrett's biomarker literature with the goal of improving care for their patients. Although many articles. including more that 60 proposed biomarkers, have been published on this subject, only a few describe phase 3 and 4 studies that are of interest to the clinical gastroenterologist (Table 1). For year, dysplasia grade has been the sole means of risk stratification for patients with BE, and it likely will continue to be used in the foreseeable future. The current authors believe that dysplasia classification can be valuable using the team management approach and quality controls described previously. Significant problems, however, have emerged in phase 2 through 4 studies of dysplasia that make it imperative for the Barrett's field to incorporate additional biomarkers as they are validated. These problems include poor reproducibility of dysplasia interpretations, poor predictive value for negative, indefinite, and low-grade dysplasia, and inconsistent results for HGD in different centers, all of which makes it virtually impossible to develop national guidelines for surveillance. Some studies have even suggested that endoscopic biopsy surveillance using dysplasia may not be worthwhile. Currently, flow cytometric tetraploidy and aneuploidy have progressed furthest in biomarker validation (see Table 1). With proper handling, endoscopic biopsy specimens can be shipped to reference laboratories that have the instruments, computer analytic methods, and expertise to reproducibly detect tetraploidy and aneuploidy. The results of phase 4 studies indicate that flow cytometry appears to be useful in detecting a subset of patients who do not have HGD and yet have an increased risk of progression to cancer that cannot be identified by dysplasia grade. For many reasons, the authors anticipate that the number of validated biomarkers will increase substantially in the future. Biopsy repositories are now readily available for phase 3 studies that can evaluate and compare biomarkers. There are initiatives for multi-institutional Barrett's Centers of Excellence that could provide rapid progress in biomarker evaluation. In addition to new candidate biomarkers, the human genome project has provided high-throughput methodologies and methods for computer analysis of data, which can provide the volume and quality control required for clinically useful biomarkers. Currently, 17p (p53) LOH has progressed the furthest among molecular biomarkers. The authors do not recommend its routine clinical use at the present time, however. Finally, it is likely that clinicians will want to follow the results of clinical treatment-response studies and epidemiologic studies that evaluate relationship between clinical interventions or environmental risk and protective factors and surrogate endpoints, especially if the endpoints are progessing well along the phases of biomarker validation. These studies are likely to be of clinical interest because they may becoming the basis for randomized clinical trials to prevent cancer in BE.