[A robot-assisted system YunSRobot for soft endoscopy: the first trial of upper gastrointestinal endoscopy on human...

Objective: To evaluate robot-assisted system YunSRobot to perform oesophagogastroduodenoscopy(EGD) in simulation model and human volunteers. Methods: The YunSRobot was invented for soft endoscopy by our team. In this trial, gastroenterologists operated the double handles of YunSRobot to perform EGD endoscopy with Olympus GIF-H 260 in simulation model Lm103 of Koken Japan and human volunteers. The operating time, lumenal anatomic exposure, man-machine interaction and other parameters were recorded and compared with manual endoscopy. Results: In the endoscopy on model, each of four doctors performed 5 procedures of EGD by YunSRobot and traditional manual endoscopy, respectively. The average time of one procedure was (626.4±120.7) seconds in the robot group and (241.5±24.7) seconds in the manual group, the operating time in robot group was more than that in manual endoscopy group. (t=14.0, P<0.001). However, the robotic manipulation time of the four endoscopists was significantly shortened from the first case to the fifth case with ((783.5±154.8)s, (667.75±85.1)s, (582.0±74.7)s, (555.0±28.9)s, (543.8±29.7)s, F=15.353, P=0.03). In the endoscopy on human, three physicians operated the EGD endoscopy in 21 volunteers, each physician performed seven volunteers by YunSRobot and traditional manual endoscopy respectively. The average time of each procedure in the robot group and the manual endoscopy group was (875.6±179.8) vs (378.8±80.4)s, the operating time in robot group was also more than that in manual endoscopy group (t=10.278, P<0.001). Like the endoscopy on model, the robotic manipulation time by the three operators was significantly reduced from the first case to the seventh cases ((954.7±62.1)s, (936.7±116.9)s, (968.7±227.1)s, (1 008.0±229.4)s, (876.7±110.5)s, (735.0±149.2)s, (649.3±81.0)s, F=3.79, P=0.024). All endoscopic anatomy of EGD were clearly and skillfully observed including esophgus, cardia, fundus, body, gastric angle, antrum, pylorus, first and second parts of duodenum, and papilla in both groups. In all procedures, there were no complications. Conclusions: The soft endoscopy robot YunSRobot is competent to conduct endoscopy of EGD on human. Presently the initial learning time of YunSRobot for EGD is longer than that of experienced manual endoscopy. The learning curve shows that the time of robot manipulation reduces rapidly in the initial procedures. It indicates that YunSRobot system has user-friendly functions and is easy to master.