Effectiveness of a novel ex vivo training model for gastric endoscopic submucosal dissection training: a prospective observational study conducted at a single center in Japan
Article information
Abstract
Background/Aims
The efficacy of endoscopic submucosal dissection (ESD) for early-stage gastric cancer is well established. However, its acquisition is challenging owing to its complexity. In Japan, G-Master is a novel ex vivo gastric ESD training model. The effectiveness of training using G-Master is unknown. This study evaluated the efficacy of gastric ESD training using the G-Master to evaluate trainees’ learning curves and performance.
Methods
Four trainees completed 30 ESD training sessions using the G-Master, and procedure time, resection area, resection completion, en-bloc resection requirement, and perforation occurrence were measured. Resection speed was the primary endpoint, and learning curves were evaluated using the Cumulative Sum (CUSUM) method.
Results
All trainees completed the resection and en-bloc resection of the lesion without any intraoperative perforations. The learning curves covered three phases: initial growth, plateau, and late growth. The transition from phase 1 to phase 2 required a median of 10 sessions. Each trainee completed 30 training sessions in approximately 4 months.
Conclusions
Gastric ESD training using the G-Master is a simple, fast, and effective method for pre-ESD training in clinical practice. It is recommended that at least 10 training sessions be conducted.
INTRODUCTION
The treatment options for gastric cancer include endoscopic therapy, surgery, and chemotherapy. Recently, advances in endoscopic diagnostics have enabled the early detection of lesions, and the number of cases that can be curatively treated with endoscopic therapy has increased.1 The results and safety of endoscopic treatment for early-stage gastric cancer have been established,2-4 and are now widely used in Japan. Approximately 60% of all early-stage gastric cancers are resected using endoscopic treatment,5 and the demand for this treatment is increasing annually. However, mastering endoscopic treatment techniques is challenging, and their acquisition is complex, especially for novice endoscopists. In endoscopic submucosal dissection (ESD), treatment experiences of approximately 30 cases are required to achieve a certain technical level.6-8 However, it takes considerable time for a trainee to experience 30 cases in clinical practice when the number of cases is limited. Unskilled techniques increase the risk of perforation and other treatment contingencies, such as hemorrhage. Because the occurrence of such contingencies is a significant disadvantage for patients, prior training in treatment techniques is essential.9-11
Conventional ESD training is performed using excised or live porcine stomach tissues. However, in recent years, from a hygienic standpoint, when porcine stomachs have been used, an endoscope must be explicitly prepared for training, and excised porcine stomachs must be disposed of as infectious waste. Moreover, living porcine stomachs require a dedicated facility for animal experiments, making it challenging to conduct ESD training conveniently.
The G-Master (Fig. 1) is an ex vivo gastric ESD training model developed in Japan.12 This plant-derived pseudomucosa model is hygienic, eliminates the need for a dedicated scope for training, and allows for general disposal after use. Moreover, G-Master can be used to set lesions in arbitrary positions, making it easy to perform training similar to actual gastric ESD. This is the first study to report the effectiveness of gastric ESD training using the G-Master.
The G-Master. (A) G-Master body. (B) Mucosa model.
METHODS
This prospective ex vivo study complied with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
Four trainees underwent ESD training using G-Master, with each completing 30 sessions. An instructor was present at each session and provided appropriate guidance on ESD techniques. The instructor’s role was to provide verbal guidance when trainees encountered difficulties with treatment techniques or procedures and avoid substitutions whenever possible. Additionally, during each training session, the instructor noted the strengths and areas for improvement and provided feedback on the training process. Training was limited to a maximum of one session per day. The training model was positioned on the anterior wall of the gastric antrum, and a 30 mm diameter area marked on the mucosal sheet, was excised by ESD (Fig. 2). We measured the treatment time, resection area, possibility of an en-bloc resection, and occurrence of intraoperative perforation at each session. The primary endpoint was the resection speed (mm2/min) as an objective measure of improvement in the ESD technique, and a learning curve was created. The treatment time was measured from the initial mucosal incision to resection completion, and the resected specimen area was calculated using the ImageJ image processing software (https://imagej.net/ij) (Fig. 3).
Endoscopic submucosal dissection (ESD) in the G-Master. (A) Pseudolesion with a diameter of 30 mm marked on a mucosal sheet. (B) Injection of glycerol with indigo carmine. (C, D) Pre-cut using a dual knife. (E–K) Mucosal incision and submucosal dissection using an IT-2 knife. (L) Ulceration after ESD.
Area measurement of the resected specimens using ImageJ.
Selection of trainees
All trainees were endoscopists in their 6th–9th year after graduation with limited clinical ESD experience; however, four trainees had experience of over 1,000 cases of routine esophagogastroduodenoscopy, 500 cases of colonoscopy, and 100 cases of emergency endoscopy and had observed ESD by experts more than 10 times. Only one of the four held a Board-Certified Fellowship from the Japan Gastroenterological Endoscopy Society (Table 1).
Background and endoscopic experience of trainee participants in the study
ESD procedure
A waterjet-furnished upper gastrointestinal endoscope (GIF-Q260J or GIF-H290T; Olympus) was used. A transparent-tip hood (D-201-11804; Olympus) was used to simplify the procedure. A 23 G, 4.0 mm injection needle (Olympus) was used for submucosal injection. A 2.0 mm dual knife (Olympus) was used for pre-cutting, and an IT knife 2 (Olympus) was used for cutting and dissection. The high-frequency generator used was the VIO3 (ERBE) with the end-cut I (effect 2.0) and swift coagulation (effect 7.0). Injection solutions were prepared using 10% glycerin with 0.9% saline, 5% fructose (Glycerol; Terumo Co.), and indigo carmine.
Learning curve and statistical analyses
The Cumulative Sum (CUSUM) method was used to evaluate the learning curves. The CUSUM method is a statistical quality control technique used to analyze learning curves in medical technology.13-16 Learning curve analysis is used to evaluate the progress of healthcare practitioners as they acquire proficiency in specific techniques or procedures. The CUSUM method monitors performance changes throughout the learning process and aids in assessing mastery and detecting anomalies. In learning curve analysis, it is expected that the performance indicators of the procedures performed by healthcare practitioners will change over time. The CUSUM method involves continuously monitoring the outcomes of each procedure, calculating the difference between the observed results and the target values, and cumulating the CUSUM statistics.
This quantifies the progress in performance. The target value of this study was the median resection speed for each trainee over 30 training sessions. A CUSUM curve was drawn by plotting the CUSUM after each training session against the number of procedures performed. In this setting, the CUSUM curve initially descends as negative values are added because each procedural speed may be slower than the target speed. By accumulating experience, procedural speed can be improved to match or exceed the target speed. Hence, the CUSUM curve reaches a horizontal value or increases when positive values are added. The inflection point on the curve represents the phase shift at which the endoscopist attained proficiency and mastery.15
Statistical analyses were performed using JMP Pro 17 (SAS Institute) and R software ver. 4.3.3 (The R Foundation for Statistical Computing). Continuous variables are expressed as means or medians with ranges. Continuous variables were compared using Wilcoxon rank-sum tests. Statistical significance was set at p<0.05.
Ethical statements
The Ethics Committee of Toho University Omori Medical Center approved the study (approval number: M23178), and the trainees provided written informed consent before the initiation of the study.
RESULTS
All four trainees achieved resection completion and en-bloc resection of lesions in all 30 training sessions, with no intraoperative perforation (Table 2). The resection speed increased with training for all four trainees (Fig. 4). The learning curve comprised three phases for all trainees (Fig. 5). Table 2 shows the duration of the training sessions and median resection speeds of all trainees in each phase.
Results of training for four trainees
Raw plot of resection speeds for the four trainees.
CUSUM plot of resection speed for the four trainees. CUSUM, Cumulative Sum.
The median number of training sessions, the transition point between phases for the four trainees, was 10 for the first to second phase and 19 for the second to third phase. All four trainees had significantly faster resection speeds in phase 3 than in phase 1 (p<0.05). The time required for each trainee to perform 30 training sessions was approximately 4 months.
DISCUSSION
This study is the first to demonstrate the usefulness of gastric ESD training using a novel ex vivo training model, G-Master. Although ESD is an established treatment option, its execution remains challenging. Although several reports have emphasized the need for training, the type of training that should be conducted before trainees engage in clinical practice remains unclear. One reason for this is that previous ESD training primarily utilized resected or live porcine stomachs, which poses hygiene and equipment challenges. Consequently, training was conducted individually and tailored to each facility’s condition. The key to ESD training is to establish an environment in which anyone can easily undergo the same training. The G-Master is a hygiene training tool that can be introduced into any facility, making it an excellent choice.
Previous reports have suggested that approximately 30 ESD procedures are required for skill acquisition. However, performing 30 ESD procedures at a single facility is time-consuming.17 At our institution, it takes at least one year for a trainee to perform 30 ESD procedures. In this study, it took approximately 4 months for a trainee to complete 30 training sessions, suggesting a substantial reduction in the training period.
All four trainees achieved complete and en-bloc resection in every training session, and no intraoperative perforations occurred. As the training progressed, the resection speed increased. This indicates increased speed and the acquisition of precise and careful techniques. The learning curves analyzed using the CUSUM method comprised three phases for all four trainees. The first phase was the initial growth period, which was considered the stage of understanding and adaptation to the technique. The resection speed was gradually increased during this phase. The second phase was the plateau period, in which the trainees gained understanding, adapted to the technique, and experimented with further improvements. During this phase, the resection rate remained constant. The third phase was the late growth period when the technique was refined and the resection speed increased again. The median number of sessions required to transition from the first to the second phase was 10, and from the second to the third phase was 19, with no significant variation observed among the four trainees. No final plateau was observed during the 30 training sessions. For trainees, understanding and adapting to the ESD technique are considered the most critical aspects of ex vivo training. In living-tissue ESD, the ability to respond to various conditions, including complications such as bleeding and perforation, patient movement, and respiratory fluctuations, is essential.18 However, mastering basic techniques and understanding the procedure may lead to smoother ESD performance, even if it does not encompass all living tissue procedures. Kim et al.19 also demonstrated the usefulness of ESD training for challenging locations using the EndoGel, suggesting that novel ex vivo ESD training models such as the G-Master and EndoGel may be suitable for ESD training in various situations.
However, this study has several limitations. First, the study was conducted at a single institution with only a few cases. Second, although G-Master allows multiple virtual lesion sites to be set, this study was limited to the anterior wall of the gastric antrum. Third, the four trainees had similar experience levels and may need to be tested in a setting with greater variation in experience. Finally, the contribution of G-Master training to clinical outcomes of ESD was not validated in this study. Further accumulation of cases and validation of the effects of the training in clinical practice are required. In conclusion, based on our results, trainees should practice at least 10 times with the G-Master before proceeding to the next step.
Notes
Conflicts of Interest
Dr. Takahisa Matsuda received lecture fees from Olympus Corporation and EA Pharma Corporation. The other authors have no potential conflicts of interest.
Funding
This research was supported by Project Research Grant No. 22–28 from the Toho University School of Medicine.
Author Contributions
Conceptualization: TT, RS, AF, TM; Data curation: TT, TI, AN, NS, AH; Formal analysis: TT, RS; Writing–original draft: TT, RS, AF, TM; Writing–review & editing: all authors.