There is growing controversy over the cost-effectiveness of surveillance endoscopy for patients with Barrett's esophagus. A retrospective review was performed of 80 patients who underwent resection for Barrett's adenocarcinoma to assess the influence of endoscopic surveillance on long-term survival. Twelve patients initially were diagnosed with benign Barrett's esophagus and were followed with endoscopic surveillance. The remaining 68 patients had the diagnosis of Barrett's esophagus made at the time of their cancer diagnosis or resection. Patients in surveillance programs were younger (53 vs. 64 years; P = 0.008), had better performance status (8.9 vs. 8.2; P = 0.04), had less weight loss (0.3 vs. 5.5 kg; P < 0.001), had a similar incidence of gastroesophageal reflux disease symptoms (75% vs. 60%), and were less likely to undergo preoperative chemotherapy and/or radiation therapy (8% vs. 28%). Pathologic stage was 0 or I in 9 (75%) of 12 patients in the surveillance group compared to 12 (18%) of 68 of those in the no surveillance group (P < 0.001). Median survival for patients in the surveillance group was 107 months compared to 12 months for those in the no surveillance group (P < 0.001). Stratifying for stage, surveillance (hazard ratio = 3.05; confidence interval = 1.09 to 8.57; P = 0.034) was the only predictor of survival. Surveillance endoscopy permits early diagnosis of adenocarcinoma in patients with Barrett's esophagus and contributes substantially to long-term survival.