Cui Meng, Ren Wenqing, Cui Tengfei, Chen Ruifeng, Shan Yi, Ma Xiaodong
Department of Emergency Medicine, The Sixth Medical Center, Chinese PLA General Hospital, Beijing, 100048, China.
Department of Neurosurgery, The First Medical Center, Chinese PLA General Hospital, Beijing, 100853, China.
Heliyon. 2024 Nov 13;10(23):e40364. doi: 10.1016/j.heliyon.2024.e40364. eCollection 2024 Dec 15.
To develop an intelligent human-machine collaborative control robot-assisted craniotomy system, and test its efficacy by experiments.
The system integrated a UR5 robotic arm (Universal Robots, Denmark), a host computer, a double six-degree-of-freedom force sensor(Nanjing Yuli Instrument Co., Ltd.), a medical drill(AESCULAP®, Germany), a Polaris Optical navigation system(NDI,Canada), with a self-designed navigation procedure and a visual graphical user interface(GUI). According to a preoperative CT and resection plan, the motion of robotic arm can be restricted in a precise and safe area. Through experiments of the 3D-printed skull models and animals (Bama mini pig), we tested the accuracy, efficiency and safety of the robot system.
After successfully developed the robot-assisted craniotomy system, we tested the collaborative controlling fluency of robotic arm with the average response time less than 1 s, as well as feedback sensitivity of force sensor with an average result of 60 N and 50 N when drilling on skull models and mini pigs respectively. In addition, compared with "surgeon" group, "robot" group had less average positioning error (1.87 ± 0.66 mm VS 3.14 ± 0.73 mm, P < 0.001) and time spent (6.64 ± 1.15min VS 8.06 ± 1.10min, P = 0.001) in skull model experiments. Also, in mini pig experiments, "robot" group had less average positioning error (3.26 ± 0.51 mm VS 4.39 ± 0.75 mm, P = 0.008) and time spent (11.83 ± 0.92min VS 26.10 ± 1.62min, P < 0.001) compared with "surgeon" group. No matter in skull model experiments or in mini pig experiments, the durations of robot startup and navigation process were not different between the experimental group and control group (3.44 ± 0.98 VS 3.75 ± 1.00min, P = 0.39 [skull model experiments]; 6.42 ± 0.65 VS 7.10 ± 1.12min, P = 0.11 [mini pig experiments]). Because of limited samples, we compared the incidence of tissue injury between "robot" and "surgeon" group jointly (3.8 % [1/26] VS 19.2 % [5/26], P = 0.193).
Successfully developed, the human-machine collaborative robot-assisted craniotomy system achieved craniotomy procedure fluently providing a sensitive force feedback to surgeon and did better than manual work by surgeon in accuracy, efficiency and safety. Further experimental research needs to be performed to testify its applicability in neurosurgery in future.
研发一种智能人机协作控制的机器人辅助开颅系统，并通过实验测试其有效性。
该系统集成了一台UR5机器人手臂（丹麦优傲机器人公司）、一台主机、一个双六自由度力传感器（南京煜立仪器有限公司）、一台医用钻（德国蛇牌）、一个北极星光学导航系统（加拿大NDI公司），以及自行设计的导航程序和可视化图形用户界面（GUI）。根据术前CT和切除计划，机器人手臂的运动可被限制在精确且安全的区域内。通过3D打印颅骨模型和动物（巴马小型猪）实验，我们测试了该机器人系统的准确性、效率和安全性。
成功研发出机器人辅助开颅系统后，我们测试了机器人手臂的协作控制流畅性，平均响应时间小于1秒，以及力传感器的反馈灵敏度，在颅骨模型和小型猪上钻孔时，平均结果分别为60N和50N。此外，在颅骨模型实验中，与“外科医生”组相比，“机器人”组的平均定位误差更小（1.87±0.66mm对3.14±0.73mm，P<0.001），花费时间更短（6.64±1.15分钟对8.06±1.10分钟，P = 0.001）。同样，在小型猪实验中，与“外科医生”组相比，“机器人”组的平均定位误差更小（3.26±0.51mm对4.3,9±0.75mm，P = 0.008），花费时间更短（11.83±0.92分钟对26.10±1.62分钟，P<0.001）。无论在颅骨模型实验还是小型猪实验中，实验组和对照组的机器人启动和导航过程持续时间均无差异（3.44±0.98对3.75±1.00分钟，P = 0.39[颅骨模型实验]；6.42±0.65对7.10±1.12分钟，P = 0.11[小型猪实验]）。由于样本有限，我们联合比较了“机器人”组和“外科医生”组的组织损伤发生率（3.8%[1/26]对19.2%[5/26]，P = 0.193）。
成功研发的人机协作机器人辅助开颅系统能够流畅地完成开颅手术，为外科医生提供灵敏的力反馈，在准确性、效率和安全性方面比外科医生手工操作表现更好。未来需要进行进一步的实验研究以验证其在神经外科手术中的适用性。
