Comprehensive Error Analysis for Robotic-assisted Placement of Pedicle Screws in Pediatric Spinal Deformity: The Initial Learning Curve.

BACKGROUND Surgical navigation improves pedicle screw insertion accuracy and reliability. Robotic-assisted spinal surgery and screw placement has not been fully assessed in pediatric patients with spine deformity undergoing posterior spinal fusion. The purpose of this study was to describe the learning curve for robotically assisted pedicle screw placement in pediatric patients. METHODS A retrospective review on a consecutive series of the first 19 pediatric patients who underwent posterior spinal fusion by a single surgeon using robotic navigation was performed. Demographics, curve parameters, pedicle diameter, vertebral rotation, and additional outcome measures were recorded. Screw position was assessed with calibrated intraoperative 3-dimensional fluoroscopic images. All complications of planned and placed robotically placed screws were recorded. RESULTS A total of 194 left-sided screws were planned as robot-assisted. One hundred sixty-eight of the robotically planned screws (86.6%) were placed with robot assistance; 29 robotically planned screws (15.0%) were abandoned or converted to freehand. The mean time per robotically placed screw was 3.6±2.4 minutes. Fifteen breaches (8.9%) and 1 anterior perforation occurred with 2 critical (>2 to 4 mm) breaches, 1 was associated with a durotomy, and both occurred in the first case. There were no intraoperative/postoperative neuromonitoring changes and no sequela from the durotomy. Six breaches occurred in the first case. The odds ratio of obtaining a breach in screws with a matched trajectory was 0.275 (95% confidence interval: 0.089-0.848). CONCLUSIONS Screw time and accuracy improved and the number of breaches decreased after 10 cases. This series had 2 critical breaches (between 2 and 4 mm) on the first case. Overall, excluding the 2 critical breaches, 98.8% of robotically executed screws were placed without a critical breach, which is comparable to previous pediatric deformity studies. Caution should be exercised during the initial training period to avoid complications as experience and training lead to an improved understanding of surgical planning, skive, and soft-tissue pressure on the end-effector all of which can impact accuracy of robotically assisted pedicle screw placement. LEVEL OF EVIDENCE Case Series, IV.