Computer-aided pedicle screw placement using frameless stereotaxis.

Abstract

In vitro assessment of accuracy and reliability of frameless stereotaxis for insertion of pedicle screws in human cadaveric lumbar spine. To assess a new method of targeting and placing pedicle screws in a human cadaver study. Pedicle screw instrumentation is common. Complications may occur from improper placement of screws. Even when performed by experienced spinal surgeons, improper placement can occur in 5.2% of pedicles instrumented. Development of computer-guided methods of pedicle screw insertion may decrease this complication rate. The technique used preoperative computed tomography scans together with a commercial neurosurgical navigational computer system to assist in placing guidewires in the pedicles. A section of human cadaver spine was first scanned and the data transferred to the workstation. The image data set and physical specimen were then registered by using an instrumented articulated arm to identify selected points on the specimen and randomly sample surface points. Eight highly repeatable locations on each vertebral body were found to be suitable for registration, but better overall accuracy was obtained when surface matching was used in combination with these points. Under guidance of image on the computer, Kirschner wires were inserted into the pedicles of four vertebral bodies. The spine was rescanned, and the planned and resulting positions of the wires compared. The average distance between the planned and resulting wire entry point was 1.2 mm, with an average difference in planned and resulting trajectories of 6.0 degrees. Computer-aided pedicle screw instrumentation is feasible. Further technical points require clarification before widespread use is possible.