SUMMARY The study's aim was to retrospectively evaluate the surveillance history of Barrett's esophagus (BE) patients with endoscopically treated early neoplasia. All BE patients endoscopically treated for early cancer (EC) or high-grade intraepithelial neoplasia (HGIN) in a lesion or mass between 1998 and 2005 were included. Endoscopy and histology records were reviewed. Ninety-four patients (78 males, mean age 67 years, 24 HGIN, 70 EC) were included. In 36 (38%) patients, HGIN/EC was diagnosed at (or within 6 months after) initial endoscopy. The remaining 58 (62%) patients had a surveillance history (median duration 7 years, mean 6.7 endoscopies). Seventy-nine percent of these had low-grade intraepithelial neoplasia (LGIN) diagnosed at least once during their surveillance period with a median of seven endoscopies and a median number of biopsies that was 50% of what should have been taken according to the Seattle protocol. Patients without any dysplasia during earlier surveillance (n = 12, 21%) had undergone significantly less endoscopies (median four endoscopies, P = 0.02) and had a median biopsy percentage that was 23% of the Seattle protocol (P < 0.001 versus 50% in LGIN). In this selected cohort of patients with early Barrett's neoplasia, 38% of patients were diagnosed at initial endoscopy. Of the patients with a surveillance history, 79% had shown LGIN prior to HGIN/EC diagnosis. Only 21% of patients had a surveillance history without any dysplasia, which in general encompassed endoscopies with an insufficient number of biopsies, suggesting sampling error. This underlines the importance of obtaining an adequate number of biopsies during surveillance endoscopies.