Conventional endoscopic mucosal resection versus modified endoscopic mucosal resection for duodenal neuroendocrine tumor.

Abstract

As the incidence of duodenal neuroendocrine tumors (DNET) is steadily increasing, the role of endoscopic treatment for appropriate lesions is becoming more significant. We aimed to compare the outcomes according to lesion size and endoscopic mucosal resection (EMR) techniques for DNET treatment. Patients who underwent endoscopic treatment for DNET between June 2000 and December 2019 were included. The clinicopathologic features and treatment outcomes were investigated by reviewing medical records. Overall, 104 cases underwent endoscopic resection for nonampullary DNET, including conventional EMR (n = 57), cap-assisted EMR (EMR-C, n = 19), and precut EMR (EMR-P, n = 28). The en bloc resection rates (100% vs. 94.7% vs. 96.4%) and histologic complete resection rates (45.6% vs. 52.6% vs. 57.1%) were not significantly different between the EMR, EMR-C, and EMR-P groups. The histologic complete resection rates were significantly higher in lesions < 10mm than in lesions ≥ 10mm (69.8% vs. 38.9%, P = 0.013). In lesions < 10mm, perforation occurred more frequently in the modified EMR group than in the conventional EMR group (13.2% vs. 0.0%, P = 0.007). During the median follow-up period of 88.0months, the recurrence-free survival (92.2% vs. 94.4% vs. 92.1%) and overall survival (98.0% vs. 88.1% vs. 100.0%) rates did not show significant differences between the EMR, EMR-C, and EMR-P groups. Conventional EMR and modified EMR are feasible and effective for the treatment of nonampullary DNET sized < 10mm and limited to mucosal and submucosal layer. Additionally, endoscopists should be aware of the high risk of perforation in modified EMR.