OBJECTIVES: I. Determine whether oral eflornithine (DFMO) given in this study will cause significant reduction of the Ki67 labelling index in subjects with intestinal type Barrett's esophagus and low grade dysplastic Barrett's esophagus. II. Determine whether oral DFMO will alter the pathology and morphology of Barrett's esophagus. III. Determine whether there is a difference in cellular DNA ploidy and/or nuclear or nucleolar morphometry in patients with dysplastic Barrett's esophagus and nondysplastic intestinal type Barrett's esophagus compared to normal gastric fundic mucosa. Determine whether DFMO modulates changes in these surrogate endpoint biomarkers towards normal mucosal values. IV. Determine whether cells demonstrating nuclear p53 protein accumulation are either lost or undergo a change in cellular distribution, following treatment of patients with dysplastic Barrett's mucosa with DFMO. V. Determine whether DFMO modulates changes in growth factor or oncogene expression in dysplastic Barrett's esophagus and nondysplastic intestinal type Barrett's esophagus. VI. Determine whether pathologic or biologic surrogate modulation occurring during 6 months of DFMO treatment reverts 6 months after treatment is discontinued.

OUTLINE: This is a randomized, placebo controlled, double blind prevention study. Patients are initially stratified by dysplasia status at baseline (metaplastic vs low grade dysplastic) and treatment group (placebo vs eflornithine). Patients are randomized to receive daily doses of eflornithine (DFMO) or placebo for 26 weeks. At 0, 4, 8, 12, 16, 20, and 26 weeks there are toxicity and adherence evaluations and at weeks 26 and 52 patients have follow-up endoscopies.

PROJECTED ACCRUAL: A total of a 152 evaluable patients will be accrued in this study.