Accuracy of single-time, multilevel registration in image-guided spinal surgery

Computerized frameless stereotactic image-guidance has been used in recent years to improve the accuracy and safety of pedicle screw placement during spine surgery. Because the possibility of intervertebral motion exists, and because the patient is usually in a different position when preoperative imaging is performed compared with the operative position, it has been suggested that the imaging model of the complete lumbar spine and the surgically exposed lumbar spine may be significantly discordant. Consequently, current protocols suggest registering each spinal level (single-level registration) separately before pedicle screw placement at that level, a time-consuming process.

To assess the accuracy of single-time multilevel registration for multilevel pedicle screw placement during image-guided, computer-assisted spine surgery, in the setting of degenerative disorders of the lumbar spine.

This is a prospective clinical and radiological study of 45 patients with degenerative disorders of the lumbar spine who underwent instrumented fusion with the use of single-time multilevel registration computer-assisted, image-guided tomography. The accuracy of the pedicle screws placement was confirmed on the basis of a protocol that included intraoperative spontaneous electromyographic (EMG) recordings, direct pedicle visualization, and computer tomography (CT) scans when clinically indicated during the follow-up period.

Forty-five consecutive patients who fulfilled the criteria of computer-assisted, image-guided tomography pedicle screw placement for degenerative lumbar spine disease without overt instability.

The principal outcome measure was the accuracy of pedicle screw placement with single-time multilevel registration for multilevel pedicle screw placement during image-guided, computer-assisted spine surgery; postoperative CT performed for clinical indications during the follow-up course was used for the assessment of pedicle screw placement.

Patients were assessed clinically before and after the operation. Data from 45 consecutive cases of image-guided, computer-assisted lumbar spinal fusion were statistically analyzed to determine the relationship between the number of levels registered during single-time registry and the mean registration error (MRE). Intraoperative spontaneous EMG, direct visualization, and postoperative CT scans were used to assess the accuracy of pedicle screw insertion.

None of the patients involved in this study experienced clinical sequelae of improper pedicle screw placement. MREs after surface mapping and after point merge were small (less than 1.00 mm and less than 3.00 mm, respectively). During the intraoperative assessment of the pedicle screws placement, no significant spontaneous EMG activity was recorded and the pedicular walls were found intact in direct visualization. The postoperative CT scans showed in 10 patients accurate placement in 55 of the 57 pedicle screws with expansion of the medial wall in two screws.

Single-time, multilevel registration may decrease operative time relative to repeated, single-level registrations, without compromising the increased accuracy of pedicle screw placement afforded by this technique in the setting of degenerative disorders of the lumbar spine. Despite the advantages in computer-guided image surgery, cautious application in the individual patient is recommended until more comprehensive data can be gathered in specific degenerative pathology with overt instability; thus the knowledge of the anatomy remains crucial.