Model Experimental Study of Man-Machine Interactive Robot-Assisted Craniotomy.

Abstract

To evaluate the feasibility, safety, and accuracy of the new man-machine interactive robotic system in model experiment. The implantation of the 8 to 10 bone screws over the skull model obtained from real patient's digital imaging and communications in medicine (DICOM), three-dimensional spiral computed tomography (CT) scans were taken. The end of the robotic arm was replaced with standard parts (including marker ball) for cone beam computed tomography (CBCT) scanning. The marker ball and marker pin were segmented and marked and exported via txt format. The position of the robotic end and model was obtained through the conversion of spacious position of standard parts and the executive end was replaced eventually. The water balloon was placed inside the skull model to imitate the dura mater and the destruction was documented for the system's safety. The system accuracy was evaluated by the error between the actual drilling position and the virtual plan and selection of 14 points of the skull window of milling pathway, monitored intraoperatively via Micron Tracker system and the overall skull window overlapped percentage via Mimics. Five model experiments were successfully performed with the average registration time of 3 minutes without destruction of balloon. The error of the outer table was 0.85 ± 0.45 mm, the inner table was 0.78 ± 0.49 mm, the line segment error of milling cutter was 0.93 ± 0.50 mm and the overall skull window overlapped percentage was 97.37% ± 0.78%. The system shows safety, accuracy, and reliability which can be an optional assistant method for craniotomy in the future.