Endoscopic Submucosal Dissection In A Western Country: Learning Curve From An Animal Training Workshop

Su1579 Endoscopic Submucosal Dissection in a Western Country: Learning Curve From an Animal Training Workshop Joaquin De La Pena*, Alvaro Teran, Jose C. Marin-Gabriel, Jose C. Manuel-Palazuelos, Fernando Mujica, Gloria Fernandez-Esparrach, Sarbelio Rodriguez-Munoz Digestive diseases, HU Marques de Valdecilla, Santander, Spain; H Virtual Valdecilla, Santander, Spain; Gastroenterology Service, H U 12 Octubre, Madrid, Spain; Gastroenterology, H Donostia, San Sebastian, Spain; Gastroenterology, H Clinic, Barcelona, Spain Introduction: Endoscopic submucosal dissection (ESD) is a difficult and relatively novel procedure in western countries with a non-well-established learning curve. Japanese experts recommend beginning in ex vivo stomachs, continuing with living pigs finally switching to patients only after reaching an en bloc resection (R0) level O 90% with a! 5% perforation rate. Aim Our aim was to assess ESD results and learning curve differences comparing two different expertise levels of endoscopists and two different kinds of knives. Methods: Sixteen endoscopists participate in an ESD training workshop on a living swine gastric model. They alternate two different ESD techniques: classic knives (DC-D2618 and DP-D2518, Pentax), C-ESD, and water-jet assisted hybrid knives (I-type or T-type, ERBE), H-ESD. Two-centimeter lesions were simulated by making 10 marks on the gastric mucosa. Endoscopists were classified in two levels according to their previous experience: level 1 (% 20 prior ESDs), L1, or level 2 (O 20 prior ESDs), L2. Perforation, muscular layer damage, need for hemostasis and R0 rates, specimen size and damage, procedural time and speed were analyzed. The study was performed in October 2012 and November 2014. Local Animal Ethics Committee approval was obtained. Results: Ninety nine ESDs were performed, 6(rank 4-7) procedures/endoscopist, with an H-ESD:C-ESD ratio of 2:1. There were no statistical differences on mean specimen size for L1 and L2 (687 470 vs 735.7 397.6 mm) or ESD techniques, C-ESD and H-ESD (702 402 vs 709 461 mm). Procedural speed was higher for H-ESD compared to C-ESD (16.2 9.4 vs 11.7 8.6 mm/min, p: 0.03) and for L2 compared to L1 (17.9 9.9 vs 12.5 8.4 mm/min, p: 0.07). Global R0 rate was poor (70.8%) while the highest R0 rate was found for L2 and C-ESD (86.7%), but no statistical differences between groups were found. L2 group had lower perforation rate than L1, while incomplete muscular layer damage, specimen cuts and need for hemostasis were similar for both L1 and 2 (table 1). Procedural speed was the only parameter which improved after first cases but R0 or perforation rate did not. Conclusions: Hybrid-ESD devices improve dissection speed while it does not ensure a more precise or safer technique. ESD learning curve in animal model could take up to 40 cases to reach acceptable R0 and perforation rates.