BACKGROUND: Although endoscopic submucosal dissection (ESD) represents a significant progress in therapeutic endoscopy, little is known about how trainees attain its technical proficiency. The aims of this study were to evaluate the learning curve and clinical outcomes of gastric ESD performed by trainee endoscopists. METHODS: Between April 2009 and March 2014, 334 gastric ESD procedures were performed by seven trainees (42-50 consecutive cases per trainee). The learning curve for gastric ESD was evaluated using the cumulative sum (CUSUM) method. Long-term outcomes were also analyzed. RESULTS: The calculated CUSUM learning curve comprised three phases: Phase I (the initial 15 cases), Phase II (the middle 15 cases), and Phase III (the final 20 cases). The resection speed was 5.7 ± 3.2 min/cm2 in Phase I, 4.7 ± 2.7 min/cm2 in Phase II, and 4.3 ± 2.2 min/cm2 in Phase III. The resection speed was significantly faster in Phase II than in Phase I (p = 0.019). The rate of self-completion reached more than 95% in Phases II and III. The overall complete resection rate was 95%, showing no significant differences between the three phases. Local recurrence was diagnosed in none of the patients during a median follow-up of 38 months. CONCLUSIONS: For the trainees, 30 cases were required to attain a higher technical level of competence in gastric ESD. After the learning curve phase, trainee operators could be expected to complete the procedure successfully with favorable long-term outcomes.
BACKGROUND: Although endoscopic submucosal dissection (ESD) represents a significant progress in therapeutic endoscopy, little is known about how trainees attain its technical proficiency. The aims of this study were to evaluate the learning curve and clinical outcomes of gastric ESD performed by trainee endoscopists. METHODS: Between April 2009 and March 2014, 334 gastric ESD procedures were performed by seven trainees (42-50 consecutive cases per trainee). The learning curve for gastric ESD was evaluated using the cumulative sum (CUSUM) method. Long-term outcomes were also analyzed. RESULTS: The calculated CUSUM learning curve comprised three phases: Phase I (the initial 15 cases), Phase II (the middle 15 cases), and Phase III (the final 20 cases). The resection speed was 5.7 ± 3.2 min/cm2 in Phase I, 4.7 ± 2.7 min/cm2 in Phase II, and 4.3 ± 2.2 min/cm2 in Phase III. The resection speed was significantly faster in Phase II than in Phase I (p = 0.019). The rate of self-completion reached more than 95% in Phases II and III. The overall complete resection rate was 95%, showing no significant differences between the three phases. Local recurrence was diagnosed in none of the patients during a median follow-up of 38 months. CONCLUSIONS: For the trainees, 30 cases were required to attain a higher technical level of competence in gastric ESD. After the learning curve phase, trainee operators could be expected to complete the procedure successfully with favorable long-term outcomes.