Prospective, Randomized Comparison of Three Endoscopic Submucosal Dissection (ESD) Knives in an Acute Porcine Model: Results from a U.S. Center

Abstract

Purpose: Few studies have compared the efficacy and safety of ESD knives in prospective, randomized comparisons, none involving gastroenterologists early in their ESD experience. We present a randomized study of three mechanistically different ESD knives in the hands of a U.S. gastroenterologist in the early stages of the ESD learning curve. Methods: 12 Yorkshire pigs were used. Each of 3 knives, Dual (Olympus), IT2 (Olympus) or Hybrid T-type (ERBE) was used to resect 7 esophageal, 7 proximal gastric and 7 distal gastric simulated lesions (total of 63 lesions, 21 lesions per knife). Randomization was per pig and per lesion with randomized order of resections. Hetastarch was used for submucosal “lifting” for the Olympus knives and saline was used for the HKT (which is capable of injecting through the tip of the knife). At the end of each session the pig was euthanized and the esophagus and stomach harvested; a leak test was performed and resection craters along with their respective mucosal specimen were submitted for histology. The primary end point was total ESD operative time (OT). Secondary outcomes were effect of lesion location and learning curve on OT, and differences among the knives in episodes of significant hemorrhage requiring use of coag-grasper, number of accessory exchanges (injection needle, coag-grasper and knife), macroscopic perforation (by endoscopic assessment or leak test), degree of thermal injury to the deep aspect of the mucosal specimen and injuries and depth of injuries to the muscularis propria (MP) layer as assessed histologically by a pathologist blinded to the type of knife used. Results: Mean O.T. for all knives was 30 minutes (10-58). HKT and IT2 knives were faster than the dual knife (28 min, 28 min, 34 min respectively) but did not reach statistical significance. Mean O.T. for gastric ESD was shorter than for esophageal (25 min, 42 min, p=<0.001). ESD for proximal gastric lesions was faster compared to distal ones (20 min, 29 min p=0.002). A learning curve effect was demonstrated. For the proximal gastric lesions OT significantly decreased after the first 7 ESDs (all knives combined) from mean of 24 min for the first 7 ESDs to 17 minutes for the subsequent 14 (p=<0.05). There was no statistically significant difference in hemorrhage, thermal injury to mucosa and injuries to MP amongst the knives. One macroscopic perforation was seen only (Hybrid knife). Conclusion: Our results confirm prior studies indicating that there is a steep learning curve. Lesion location is an important determinant of OT (esophageal>distal gastric> proximal gastric) with ESD in the esophagus being more challenging and time consuming than gastric ESD.