This study was conducted among trainee endoscopists who participated in ESD hands-on seminars held in August 2018 and September 2019 at the Kitakyushu Municipal Medical Center. A total of 22 trainees from 13 institutions performed ESD on porcine models; 20 trainees from 11 institutions in Japan and 2 trainees from 2 institutions in China. Before the ESD training, all trainees reviewed video lectures on ESD procedures with scissor-type and needle-type endo-knives. Thereafter, each trainee performed two ESDs under expert supervision, one using the scissor-type knife (ESD-S) and the other using the needle-type knife (ESD-N). Owing to differences in experience among the trainees, they were classified into two groups according to their experience in upper or lower gastrointestinal (GI) endoscopy. One group included 14 trainees who had experienced less than 1,000 upper or lower GI endoscopies (junior-trainee: J-Trainee), while the other group included eight trainees who had experienced more than 1,000 endoscopies (senior trainee: S-Trainee). Within each group, the trainees were further randomly assigned to two additional groups: ESD-S and ESD-N; after performing ESD with one knife, they were reassigned to the other group in which they performed with the other knife (Fig. 1). All ESDs were performed under expert assistance. One expert and one trainee were assigned to each endoscopy room, where the experts supervised and advised the trainees during the ESD procedure. After the completion of all procedures, the outcomes were recorded in a case report form.
The study protocol was approved by the Institutional Review Board of the Kitakyushu Municipal Medical Center (No. 201901076). The requirement for written informed consent was waived because this was an ex vivo porcine model study; all trainees were orally informed about the study and consented to participate. This study was registered as a retrospective study with the University Hospital Medical Information Network (UMIN) Clinical Trials Registry (www.umin.ac.jp/ctr/) (UMIN00041827). Although this was an ex vivo animal model study, it was conducted in accordance with the guidelines of the Animal Research Reporting In Vivo Experiments (ARRIVE) as much as possible.
Trainees and experts
Trainees who had performed more than 100 upper gastrointestinal (GI) endoscopies but no or less than five ESDs (performed independently without self-completion), were included in this study. All experts were endoscopy specialists certified by the Japan Gastroenterological Endoscopy Society and had performed more than 100 ESDs. Experts familiar with the scissor-type knife supervised the ESD-S procedure, while those familiar with needle-type knife supervised the ESD-N procedure.
Harvested pig stomachs and esophagi were used as the ex vivo porcine models for ESD. These were frozen for transport and thawed prior to the seminar (shown in Fig. 2). The gastric lumen was cleaned by water lavage, and the stomach was fixed on a plastic tray. An overtube (TOP Co., Tokyo, Japan) was attached to the esophagus, and the pylorus of the stomach was tied. Upper GI endoscopes (EG-450-RD5; Fujifilm, Tokyo, Japan, GIF-Q260J, Olympus, Tokyo, Japan) fitted with disposable straight sift distal attachments [D-201-11804 (Olympus) or Elastic touch (TOP Co.)] were used in the ESD procedure. VIO3 (ERBE, Tubingen, Germany) was used as an electrosurgical power source. The marking was performed in the Forced Coag 9.0 mode, while mucosal incision and submucosal dissection were performed in the endo-cut I, endo-cut 2, or Forced Coag 9.0 modes. Two types of endoscopy rooms were prepared: one for ESD-S and the other for ESD-N. The Clutch Cutter® (DP2618DT, Fujifilm) and Flash Knife (Fujifilm) were used as the scissor-type knife and the needle-type knife, respectively (shown in Fig. 3a, 3b). A mixture of hyaluronic acid and normal saline with a small amount of indigo carmine was used as the injection solution.
Endoscopic submucosal dissection procedure
The ESD procedure consists of the following steps: 1.) marking around the lesion, 2.) lifting the lesion by submucosal injection, 3.) making a circumferential mucosal incision outside the marking, and 4.) dissecting the submucosal layer. The same procedure was utilized in this study. Prior to ESD, the experts created mock lesions measuring 20 mm in diameter using eight marking dots. All subsequent procedures were performed by the trainees. A traction method using dental floss and a hemoclip (DFC) was used in all cases. After circumferential mucosal incision, the DFC was attached to the edge of the mucosal flap, and the floss was pulled. After confirming good visualization of the submucosal layer, dissection was initiated. The time required to attach the DFC was excluded from the procedure time. The resected specimen was fixed on a plate with proper tension, and the diameters of the mock lesion and the resected specimen were measured using a ruler.
The total procedure time was defined as the time from creating the mucosal incision to completion of submucosal dissection. The lesion location was classified into the upper, middle, and lower thirds of the stomach. Furthermore, the lesion position was categorized on the basis of its location in the anterior wall, posterior wall, lesser curvature, and greater curvature . En bloc resection was defined as the resection of the specimen intact with all marking dots. Self-completion was defined as the completion of en bloc resection by the trainee without expert intervention. The operator would be changed if the ESD could not be completed within the preset procedure time of 30 min, the supervisors considered the procedure difficult for the trainees to complete by themselves, or intraoperative perforation occurred. Perforation was defined as a wall defect directly visible on the outside of the stomach. Muscular injury was defined as the exposure and damage of the muscle layer. The difficulty of ESD for trainees was defined as non-self-completion, non-en-bloc resection, perforation, muscular injury, or long procedure time exceeding 20 min.
The primary outcome was the procedure time for ESD. The secondary outcomes included the mucosal incision time, submucosal dissection time, self-completion rate, en bloc resection rate, and rates of complications (muscular injury and intraoperative perforation). These outcomes were compared between the ESD-S and ESD-N groups.
In subgroup analysis, the predictive factors associated with the difficulty of ESD for trainees were investigated. Tumor location (upper/middle third in the stomach or lower), tumor position (greater curvature of the stomach wall or others), endo-knife (ESD-S or ESD-N), and experience of trainees (J-Trainee or S-Trainee) were included among the investigated factors.
All statistical analyses were performed using EZR ver. 1.27 (Saitama Medical Center, Jichi Medical University, Japan). Continuous valuables were expressed as medians with interquartile ranges (IQR) and analyzed by the student’s t-test after logarithmic transformation. Categorical valuables were expressed as numbers with percentages (%) and analyzed by the Chi-square test or the Fisher’s U test. Logistic regression analysis was used to calculate univariate and multivariate-adjusted odds ratios (ORs) and the 95% confidence intervals (95% CIs). In multivariate analysis, OR was adjusted for potentially confounding factors including tumor location, tumor position, endo-knife type, and experience of trainees. Statistical significance was considered as p < 0.05.