Does intraoperative CT navigation increase the accuracy of pedicle screw placement in pediatric spinal deformity surgery? A systematic review and meta-analysis

Purpose Although pediatric spinal deformity correction using pedicle screws has a very low rate of complications, the long-term consequences of screw malposition is unknown. CT navigation has been proposed to improve screw accuracy. The aim of this study was to determine whether intraoperative navigation during pedicle screw placement in pediatric scoliosis makes screw placement more accurate. We also examined radiation exposure, operative time blood loss and complications with and without the use of CT navigation in pediatric spinal deformity surgery. Methods A systematic review of the literature was conducted. After screening, 13 articles were qualitatively and quantitatively analyzed to be used for the review. A random effects meta-analysis using REML methodology was employed to compare outcomes of screw accuracy, estimated blood loss, radiation exposure, and surgical duration. Results Screws placed with CT navigation surgery were three times as likely to be deemed "acceptable" compared with screws placed with freehand and 2D fluoroscopy assistance, twice as likely to be "perfect", and only 1/3 as likely to be potentially unsafe (all p value < 0.01). EBL was not significantly different between groups; however, operative time was roughly thirty minutes longer on average. Random effects analysis showed no significant difference in effective dose radiation while using CT navigation (p = 0.06). Conclusion This systematic review of the literature demonstrates that intraoperative navigation results in more accurate pedicle screw placement compared to non-navigated techniques. We found that blood loss was similar in navigated and non-navigated surgery. Operative time was found to be approximately a half hour longer on average in navigated compared to non-navigated surgery. Effective radiation dose trended higher in navigated cases compared to non-navigated cases but did not reach statistical significance.