Clinical outcome of endoscopic resection for nonampullary duodenal tumors.

BACKGROUND AND STUDY AIMS Compared with any other location in the gastrointestinal tract, the duodenum presents the most challenging site for endoscopic resection. The aim of this study was to analyze the clinical outcomes of duodenal endoscopic resection and to assess the feasibility of the technique as a therapeutic procedure. PATIENTS AND METHODS A total of 113 consecutive patients with 121 nonampullary duodenal tumors underwent endoscopic resection by endoscopic submucosal dissection (ESD), endoscopic mucosal resection (EMR), or polypectomy between January 2000 and September 2013. Long-term outcomes were investigated in patients with more than 1 year follow-up. RESULTS The median tumor size was 12 mm (range 3 - 50 mm). Lesions consisted of 63 adenocarcinomas/high-grade intraepithelial neoplasias (53 %) and 57 adenomas/low-grade intraepithelial neoplasias (48 %). Endoscopic resection included 106 EMRs (87 %), 8 ESDs (7 %), and 7 polypectomies (6 %). En bloc resection was achieved in 77 lesions (64 %), and 43 lesions (35 %) underwent piecemeal resection; one procedure was discontinued due to perforation. There were 14 cases of delayed bleeding after EMR (12 %), 1 perforation (1 %) during ESD, and 1 delayed perforation (1 %) after ESD, which required emergency surgery. Of the 76 patients who were followed for more than 1 year, none of the patients died from a primary duodenal neoplasm, and there were no local recurrences during the 51-month median follow-up period (range 12 - 163 months). CONCLUSIONS Duodenal endoscopic resection was feasible as a therapeutic procedure, but it should only be performed by highly skilled endoscopists because of its technical difficulty. Piecemeal resection by EMR is acceptable for small lesions, based on these excellent long-term outcomes.