Submillimeter-level Extremely High Precision Planning and Operation: An Example of Robotic Cochlear Implant Microsurgery

Robots have become essential tools in modern clinical surgeries, successfully applied in laparoscopic, orthopedic, percutaneous, and natural orifice transluminal endoscopic surgeries. Microsurgery demands higher precision, and the current millimeter-level accuracy does not meet the requirements, necessitating the development of more precise surgical robots. Take Cochlear Implant (CI), one of the most representative microsurgeries, as an example. Faced with narrow surgical fields, restricted visibility, and high operational difficulty in complex and delicate CI surgeries, the paper proposes a robotic in-situ surgical system that integrates operation and imaging to achieve sub-millimeter precision. At the planning level, we developed an adaptive scale segmentation network (ASS-Net) to precisely identify key anatomical structures and created an automated, geometry-constrained algorithm to plan the optimal direct cochlear access (DCA). At the execution level, we propose an adaptive drilling strategy based on bone-density awareness, allowing the robot to segmentally adjust drilling speed and force according to different tissue types to minimize the risk of tissue damage. Based on the innovative “operation and image integration” paradigm, we performed simulation experiments using our self-developed platform and conducted actual CI experiments in collaboration with hospitals. The proposed technology has the potential to enable higher-precision and faster robotic surgeries in fields such as neurosurgery and orthopedics.