Semi-Automation in Image-guided Robotic Brain Surgery

Abstract

Surgical robots have been widely used to assist surgeons to carry out dexterous surgical procedures via various ways. Robotically assisted surgery allows surgeons to overcome the limitations in the minimally invasive surgery and thus to enhance the surgical performance. Most of the robotically assisted surgery require surgeon’s operation directly or indirectly via a teleoperation station. A certain level of autonomy in robotic surgery could not only free the surgeon from some tedious and repetitive tasks, but also utilize the advantages of the robot: high dexterity and accuracy. This work considers the image-guided robotic neurosurgery, aiming to semi-automate a neurosurgical procedure for clean up the tumor residues after manual removal of the brain tumor using a surgical robot, where the manual operation is a tedious and time-consuming task for surgeons due to long time integration of the fluorescent signal. The entire medical procedure is modeled using behavior tree framework, and encoded and tested on RAVEN TM II surgical robotic platform using a simulated brain phantom. Furthermore, a path planner is developed that automatically synthesize different ablation path plans to cover the entire tumorous regions. At the current stage, the path planner is able to generate three types of path patterns for the surgeon to choose the most reliable plan before robotic execution. Along with the generated path plans, other computed metrics such as the total number of the via-points in each plan, estimated time for the robotic execution, will also be delivered to the surgeon for reference.