Robot-assisted pedicle screw placement by using optical navigation

Pedicle screw placement is the gold standard for spinal pathologies that require surgical stabilization. Accurate screw placement is important to avoid severe complications, but ensuring accuracy remains a challenge. This work proposes an optical camera-based robot-assisted drilling system for spine surgery. Preoperative CT scans are used to plan screw trajectories in both synthetic phantoms and ex-vivo ovine spines. After robotic drilling, postoperative scans are registered to the preoperative plan to quantify drilling accuracy. In total, 48 screws are drilled into ex-vivo ovine spines, producing a median entry point distance error of 1.51 mm (IQR 0.91mm, 2.10 mm) and an angle error of $2.30°$ (IQR $1.28°$ , $3.13°$ ), where 95.83% of the placements were clinically successful. Optical camera-based robot-assisted drilling system for spine surgery may increase accuracy by providing precise spatial registration between the patient anatomy and the robotic system, enabling accurate alignment and execution of the drilling task.