Clinical Accuracy of Cervicothoracic Pedicle Screw Placement

Abstract

Posterior transpedicular fixation at the cervicothoracic junction (CTJ) is increasing in popularity. However, the clinical accuracy of pedicle screw placement at the CTJ has not been specifically assessed. Between January 2000 and July 2004, 60 consecutive patients underwent a variety of posterior spinal procedures necessitating pedicle screw placement at C7, T1, and T2. Thirty-two patients had cervicothoracic screws (3.5 to 4.5 mm) placed by an "open" technique (laminectomies or lamino-foraminotomies) and 28 patients with either a closed (before any decompression) 2-dimensional (n=19, fluoroscopy) or 3-dimensional (n=9, CT) computer-assisted technique. Screws were independently assessed for pedicle breach on postoperative CT and scored using a points-based classification system. The total number of screws placed was 86, 63 and 45 in the open, closed-2-dimensional and closed-3-dimensional groups, respectively. Overall, 61(70.9%), 51(81%), and 40(89%) screws were completely within the pedicle. In the open group, the majority of pedicle breaches were more than 2 mm [n=3 (<2 mm), n=20 (2-4 mm), n=2 (>4 mm)]. Screw violation occurred laterally 11/25(44%), medially 3/25(12%), inferiorly 7/25(28%), and superiorly 4/25(16%). In the closed technique, all breaches were lateral. Seventeen screws (n=11-2-dimensional, n=5-3-dimensional) breached the pedicle by a margin of less than 2 mm and 1 screw (2-dimensional) by 2 to 4 mm. Pedicle screw accuracy was significantly improved with computer-assisted techniques. However, there was no significant difference between the 2-dimensional and 3-dimensional techniques. For all patients, there were no clinically significant screw misplacements, nor any need for screw revision. Computer-assisted surgery allows for more accurate placement of pedicle screws at the CTJ. Although a higher proportion of major pedicular breaches occurred in the "open lamina/lamino-foraminotomy" group, no screws required revision in either group.