Abstract
Glenoid replacement is a manual bone removal procedure that can be difficult for surgeons to perform. Surgical robotics have been utilized successfully in hip and knee orthopaedic procedures but there are no systems currently available in the shoulder. These robots tend to have low adoption rates by surgeons due to high costs, disruption of surgical workflow and added complexity. As well, these systems typically use optical tracking which needs a constant line-of-sight which is not conducive to a crowded operating room. The purpose of this work was developing and testing a surgical robotic system for glenoid replacement.
The new surgical system utilizes flexible components that tether a Stewart Platform robot to the patient through a patient specific 3D printed mount. As the robot moves relative to the bone, reaction loads from the flexible components bending are measured by a load cell allowing the robot to “feel” its way around. As well, a small bone burring tool was attached to the robot to facilitate the necessary bone removal. The surgical system was tested against a fellowship-trained surgeon performing standard surgical techniques. Both the robot and the surgeon performed glenoid replacement on two different scapula analogs: standard anatomy and posterior glenoid edge wear referred to as a Walch B2. Six of each scapula model was tested by the robot and the surgeon. The surgeon created a pre-operative plan for both scapula analogs as a target for both methodologies. CT scans of the post-operative cemented implants were compared to the pre-operative target and implant position and orientation errors were measured.
For the standard shoulder analogs the net implant position and orientation errors were 1.47 ± 0.48 mm and 2.57 ± 2.30° for the robot and 1.61 ± 0.29 mm and 5.04 ± 1.92° for the surgeon respectively. For the B2 shoulders, the net implant position and orientation errors were 2.16 ± 0.36 mm and 2.89 ± 0.88° for the robot and 3.01 ± 0.42 mm and 4.54 ± 1.49° for the surgeon respectively.
The new tracking system was shown to be able to match or outperform the surgeon in most metrics. The surgeon tended to have difficulty gauging the depth needed as well as the face rotation of the implant. This was not surprising as the reaming tool used by the surgeon obscures the view of the anatomy and the spherical cutter hinders the ability to index the tool. The robot utilized only one surgical tool, the bone burr, precluding the need for multiple instruments used by the surgeon to prepare the glenoid bone bed. The force-space navigation method can be generalized to other joints, however, further work is needed to validate the system using cadaveric specimens.
