Accuracy of pedicle screw placement in lumbar fusions by plain radiographs and computed tomography.

STUDY DESIGN

Postoperative radiographs and computed tomography scans were used to evaluate 74 pedicle screws in 16 consecutive patients who underwent lumbar spine fusion with pedicle screw fixation.

OBJECTIVE

To evaluate pedicle screw placement using plain radiographs versus computed tomographic scans.

SUMMARY OF BACKGROUND DATA

Plain radiographs are the primary means of assessing pedicle screw placement. Comparison of plain radiographs and computed tomography has not been done.

METHODS

Screws were graded as IN, OUT, or QUESTIONABLE; the direction of misplacement was noted. All evaluations were performed independently by three observers.

RESULTS

Fewer screws were clearly within the pedicle on computed tomography when compared with plain radiographs. Computed tomography showed 10 times as many screws violating the medial cortex as did radiographs. Interobserver differences were not statistically significant. Intraobserver differences approached statistical significance when the two tests were compared. No recognized neurologic complications resulted from pedicle screw placement.

CONCLUSIONS

Plain radiographs alone may not accurately reveal pedicle screw placement. Plain radiographs and thin section computed tomographic scans should be used to evaluate postoperative neurologic deficits in patients undergoing instrumented lumbar spine fusion with pedicle screws.