Su1283 OUTCOMES AND RISK FACTORS OF COMPLICATIONS OF ENDOSCOPIC RESECTION FOR SUPERFICIAL DUODENAL EPITHELIAL NEOPLASIA: A RETROSPECTIVE OBSERVATIONAL STUDY OF 321 CONSECUTIVE CASES
Motohiko Kato,Yasutoshi Ochiai,Tadateru Maehata,Motoki Sasaki,Yoshiyuki Kiguchi,Teppei Akimoto,Atsushi Nakayama,Ai Fujimoto,Osamu Goto,Takanori Kanai,Naohisa Yahagi
DOI: https://doi.org/10.1016/j.gie.2018.04.1670
IF: 10.396
2018-01-01
Gastrointestinal Endoscopy
Abstract:Endoscopic resection (ER) could be an alternative therapy for superficial duodenal epithelial tumor (SDET), considering the invasiveness of pancreaticoduodenectomy (Whipple’s procedure). However, detailed outcomes of ER have remained unknown, especially the differences according to the procedure type (EMR or ESD) and risk factors for complications, due to the rarity of SDET. The aim of this study was to elucidate the outcomes of ER. This was a retrospective observational study. From June 2010 to June 2017, a total of 321 cases of endoscopically resected SDET (146 EMR and 175 ESD) were included in this study. We analysed the proportions of resection in a single piece, R0 resection, perforation and bleeding as outcomes of ER and compared the outcomes between EMR and ESD groups. Next, we collected data about the features and clinical course of cases with complications. Associations among outcomes of ER and lesion features, endoscopic treatment type (ESD or EMR) were analysed using a multivariate logistic regression model. The proportion of resection in a single piece and R0 resection among all cases was 96.6% and 83.4%, respectively. ESD achieved resection in a single piece in over 95% of cases regardless of lesion size. The incidence of perforation and bleeding was 8.8% and 3.4%, respectively. Most incidences of perforation were successfully managed by conservative treatment and mortality rate was 0%, and all patients were discharged with a median hospital stay of 8.5 days (range 4-52 days). When the hospital stay duration was compared according to the lesion circumference, it was significantly longer for lesion present on the medial wall than for the other lesions (median 41 vs 7 days, p=0.0331, Wilcoxon’s rank-sum test). Multivariate analysis revealed, lesion size (OR 1.43 for every increase of 10 mm, 95% CI 1.13-1.83, p<0.01) and ESD (OR 13.6, 95% CI 1.70-108, p<0.01) was independently associated with the risk for perforation. For bleeding, being located on the medial wall was the only independent predictor for delayed bleeding (OR 4.56, 95% CI 1.27-16.4, p<0.01), whereas ESD was not associated with the risk of bleeding (OR 2.19, 95% CI 0.38-12.7, p=0.36). The present study revealed that ER achieved secure resection in a single piece, by choosing ESD or EMR according to the size of lesion. Medial wall was associated with worse outcomes such as bleeding or prolonged hospital stay after perforation.