To determine whether or not flow-cytometric evidence of aneuploidy and increased G2/tetraploid fractions predispose to neoplastic progression in Barrett's esophagus, 62 patients with Barrett's esophagus were evaluated prospectively for a mean interval of 34 months. Nine of 13 patients who showed aneuploid or increased G2/tetraploid populations in their initial flow-cytometric analysis developed high-grade dysplasia or adenocarcinoma during follow-up; none of the 49 patients without these abnormalities progressed to high-grade dysplasia or cancer (P less than 0.0001). Neoplastic progression was characterized by progressive flow-cytometric and histological abnormalities. Patients who progressed to high-grade dysplasia and carcinoma frequently developed multiple aneuploid populations of cells that were detectable flow-cytometrically. Similarly, patients appeared to progress through a phenotypic sequence that could be recognized histologically by the successive appearance of Barrett's metaplasia negative for dysplasia, abnormalities in the indefinite/low-grade dysplasia range, high-grade dysplasia, and eventually adenocarcinoma. These and prior results suggest that neoplastic progression in Barrett's esophagus occurs in a subset of patients who have an acquired genomic instability that generates abnormal clones of cells, some of which have aneuploid or increased G2/tetraploid DNA contents. With continued genomic instability, multiple aneuploid subclones may evolve, one of which may ultimately acquire the ability to invade and become an early carcinoma. The combination of histology and flow cytometry can be used to identify a subset of patients with Barrett's esophagus who merit more frequent endoscopic surveillance for the early detection of high-grade dysplasia or carcinoma.