Accuracy of Transpedicular Thoracic Screws in Patients With and Without Coronal-Plane Spinal Deformities **in vivo

Walter Reed Army Medical Center, Washington, DC, USA

PURPOSE:
To examine the in vivo placement accuracy of transpedicular thoracic screws in patients with (WCD) and without (WOCD) coronal-plane spinal deformities.

METHODS:
399 thoracic pedicle screws in 47 patients were studied using preoperative radiographs and postoperative CT. Screws were inserted under fluoroscopic guidance using anatomic landmarks and the so-called "all in" or the "in-out-in" technique. Screws were regionally grouped for analysis (T1-T4: 47 screws, T5-T8: 132 screws, T9-T12: 220 screws). Pedicle wall perforations were graded in 2-mm increments. Curve magnitude and segmental vertebral rotation (Nash-Moe and Perdriolle methods) were determined from preoperative radiographs. Curve magnitude greater than 20° was designated as WCD.

RESULTS:
There were 302 screws in 28 patients WCD and 97 screws in 19 patients WOCD. In patients WCD, the incidence of full containment within the pedicle was statistically lower (42%) than in patients WOCD (62%) (P=0.001). The incidence of screws positioned with > 2mm of canal intrusion however, was statistically similar in patients WCD (7 screws, 2.3%) and WOCD (1 screw, 1.0%) (P=0.69). Overall, the percentage of "acceptably" positioned screws (fully contained screws or with either < 2 mm of medial wall perforation or < 6 mm of lateral wall perforation) was 97.7% in patients WCD and 99% in patients WOCD. Penetration of the anterior vertebral cortex occurred with greater frequency in patients WCD (8.0% vs. 1.0%) and was associated with a statistically smaller mean transverse screw angle (TSA) (P<0.0005) and with lateral perforation of the pedicle wall (P=0.001). For patients WCD, the incidence of full containment within the pedicle did not correlate with preoperative curve magnitude or with segmental rotation. There were no neurologic or vascular complications. Two screws (one in each group), in close proximity to the aorta, were revised to prevent any long-term sequelae.

CONCLUSIONS:
The incidence of canal intrusion > 2mm during the in vivo placement of transpedicular thoracic screws was statistically similar in patients WCD and WOCD. The overall percentage of "acceptably" positioned screws was 97.7% in patients WCD and 99% in patients WOCD. Anterior cortical penetration occurred more often in patients WCD and was associated with a smaller TSA and perforation of the lateral pedicle wall.