Placement of pedicle screws in the human cadaveric cervical spine: comparative accuracy of three techniques

This investigation was conducted in two parts. In the first part, a morphometric analysis of critical cervical pedicle dimensions were measured to create guidelines for cervical pedicle screw fixation based on posterior cervical topography. In the second part of the study, a human cadaver model was used to assess the accuracy and safety of transpedicular screw placement in the subaxial spine using three different surgical techniques: 1) using surface landmarks established in the first part of the study, 2) using supplemental visual and tactile cues provided by performing laminoforaminotomies, and 3) using a computer-assisted surgical guidance system.

To assess the accuracy of transpedicular screw placement in the cervical spine using three surgical techniques.

A three-column fixation device implanted to secure an unstable cervical spine can be a valuable tool with a biomechanical advantage in the spine surgeon's armamentarium. Despite this advantage, concerns over surgical neurovascular complications have surfaced. Cadaver-based morphometric measurements used to guide the surgeon in the placement of a pedicle screw show significant variability, raising legitimate concerns as to whether transpedicular fixation can be applied safely.

Precise measurements of 14 human cadaveric cervical spines were made by two independent examiners of pedicle dimensions, angulation, and offset relative to the lateral mass boundaries. On the basis of this analysis, guidelines for pedicle screw placement relative to posterior cervical topography were derived. In the second part of the study, 12 human cadaveric cervical spines were instrumented with 3.5-mm screws placed in the pedicles C3-C7 according to one of three techniques. Cortical integrity and neurovascular injury were then assessed by obtaining postoperative computed tomography scans (1-mm cuts) of each specimen. Cortical breaches were classified into critical or noncritical breaches.

Linear measurements of pedicle dimensions had a wide range of values with only fair interobservercorrelation. Angular measurements showed similarangulation in the transverse plane (40 degrees ) at each level. With respect to the sagittal plane, both C3 and C4 pedicles were oriented superiorly relative to the axis of the lateral mass, whereas the C6 and C7 pedicles were oriented inferiorly. The dorsal entry point of the pedicle on the lateral mass defined by transverse and sagittal offset had similar mean values with wide ranges, although there often was excellent correlation between observers. There were no significant interlevel, right/left, or male/female differences noted with respect to offset. Using one of three techniques, 120 pedicles were instrumented. In group 1 (morphometric data): 12.5% of the screws were placed entirely within the pedicle; 21.9% had a noncritical breach; and 65. 5% had a critical breach. In group 2 (laminoforaminotomy), 45% of the screws were within the pedicle; 15.4% had a noncritical breach; and 39.6% had a critical breach. In group 3 (computer-assisted surgical guidance system), 76% of the screws were entirely within the pedicle; 13.4% had a noncritical breach; and 10.6% had a critical breach. Regardless of the technique used, the vertebral artery was the structure most likely to be injured.

On the basis of the morphometric data, guidelines for cervical spine pedicle screw placement at each subaxial level were derived. Although a statistical analysis of cadaveric morphometric data obtained from the cervical spine could provide guidelines for transpedicular screw placement based on topographic landmarks, sufficient variation exists to preclude safe instrumentation without additional anatomic data. Insufficient correlation between different surgeons' assessments of surface landmarks attests to the inadequacy of screw insertion techniques in the cervical spine based on such specific topographic guide