Intraoperative anteroposterior and oblique fluoroscopic views for detection of mediolateral pedicle screw misplacement in the lumbar spine: a randomized cadaveric study.

Pedicle screws are commonly used for posterior fixation of the lumbar spine. Inaccuracy of screw placement can lead to disastrous complications. As fluoroscopic assisted pedicle screw instrumentation is the most frequently used technique, the aim of this study was to assess the specificity, sensitivity and accuracy of intraoperative fluoroscopy to detect mediolateral screw malpositioning. We also analyzed whether the addition of an oblique view could improve these parameters. On 12 human cadavers, 138 pedicle screws were placed intentionally either with 0 to 2 mm (75 screws), with 2 to 4 mm (six medial and 12 lateral screws) and with >4 mm (22 medial and 23 lateral screws) breach of the pedicle from Th12 to L5. Three experienced spine surgeons evaluated the screw positioning in fluoroscopic AP views and 4 weeks later in AP views and additional oblique views. The surgeons' interpretation was compared with the effective screw position on postoperative CT scans. Pedicle breaches greater than 2 mm were detected in 68% with AP views and in 67% with additional oblique views (p=.742). The specificity of AP views was 0.86 and 0.93 with additional oblique views (p=<.01). The accuracy was 0.78 with AP views and 0.81 with AP + oblique views (p=.114). There was a substantial inter-reader agreement (Fleiss's kappa: 0.632). Fluoroscopic screening of pedicle screw misplacement has a limited sensitivity. Adding an oblique view improves specificity but not sensitivity and accuracy in detecting screw malpositions. When in doubt of a screw malpositioning, other modalities than a fluoroscopic assisted pedicle screw instrumentation such as intraoperative CT imaging or an intraoperative exploration of the screw trajectory must be evaluated. [ABSTRACT FROM AUTHOR]