Reliability of Postoperative Multislice Computed Tomography in Assessment of Pedicle Screw Placement in Thoracic and Lumbar Spinal Fixation

Abstract

Objective This study aims to evaluate the reliability of three-dimensional (3D) volume rendering computed tomography (CT) in postoperative assessment of pedicle screw placement. Methods A total of 98 patients with previous spinal fixation surgery were included. All inserted screws were evaluated postoperatively within the first week using CT. Degree of misplacement and difference between CT images were recorded. Results Overall, 76, 68, and 39 misplaced screws were detected in 3D reformatted image, coronal reconstruction, and axial CT images, respectively. Dorsal spine had higher rate of violation 46 (9.95%) with the highest rate at T4 (12.8%) comparing it with lumber 32 (6.9%). Statistically significant differences were found between lateral and medial violation ( p = 0.03), between findings of CT 3D reformatted and axial image ( p = 0.04), and also in detecting end plate perforation and anterior vertebral encroachment in different CT images ( p = 0.013). Sensitivity for 3D reformatted image and axial image compared with surgical finding in six revised screws was 100 and 95.8% and specificity 100 and 88.7%, respectively. Highly momentous agreement reported with Kappa coefficient = 0.95 Â ± < 0.001. Conclusion The vertebral body and the pedicle are 3D structures. Therefore, we have to study them as they are. The postoperative evaluation of pedicle screw using CT 3D reconstruction was a reliable method and unveiled the deficiencies in conventional axial CT.