Abstract: Objective The efficacy of the two teaching methods for EBUS-TBNA was evaluated by comparing the learning curve, diagnostic accuracy, and specimen acquisition for genotyping by operators and so to find a better one for training pulmonologists. Methods The first fifty patients of each operator learning EBUS-TBNA between January 2015 and June 2019 in a single center were retrospective collected and grouped by the operators training methods. The diagnosis, genotyping and clinical characters were reviewed. Results A total of 442 patients were collected and reviewed. There were two training groups: a conventional training group (232 cases) and a simulator assistant training group (210 cases). There was no statistical difference in the diagnostic yield between the groups of trainees. However, specimen from lung cancer patients in simulator training group were more adequate for gene testing than in conventional training group (86.8% vs 71.2%, P＜0.05). By comparing the learning curve among the groups, it was found that the trainees in stimulator assistant training group needed less minimum number of operations to achieved proficiency than in conventional training group. Conclusions Simulation-based training is demonstrated to be more efficient in improving the bronchoscopists'skill in terms of getting sufficient specimen for gene testing. Simulation-based training may also shorten the time needed for bronchoscopists'training and can be applied in the EBUS-TBNA training course.

Key words: simulation-based/simulator training, EBUS-TBNA, learning curve, specimen acquisition