Robotic-assisted navigation facilitates accurate and clinically safe pedicle screw placement by a well-trained orthopaedic surgery resident when compared to an attending surgeon: a cadaveric study.

Pedicle screws can be placed using freehand technique, navigation, or robotic-assisted navigation (RAN). There is a steep learning curve for RAN in pediatric spine deformity. This study aimed to compare pedicle screw accuracy, precision, and incidence of breach using RAN between a pediatric spine surgeon with 3 years of RAN experience and a fourth-year orthopedic resident who had completed pediatric- and adult-spine rotations with RAN training. A preoperative computed tomography (CT) scan was obtained on two adult cadavers. The resident and attending each placed T2-L5 screws bilaterally in one cadaver each using RAN. A computer-vision algorithm was used to compare the postoperative CT screw position to the preoperatively planned screw position. Laminectomies were performed to assess breach. Statistical analysis compared systematic error (signed mean error), precision (2 standard deviations), and accuracy (mean absolute error) between training levels for angular deviation and translational deviation at the screw tip, tail, and mid-pedicle positions in the medial-lateral (ML) and superior-inferior (SI) directions. Fifty-seven screws were placed in two cadavers. The attending screws were more precise and accurate than the resident screws in the ML direction at the tail and mid-pedicle positions, and more precise in the ML direction at the tip (P < 0.05). Resident screws had better systematic error and accuracy than attending screws in the SI direction at the tip, tail, and mid-pedicle positions (P < 0.02). There were no significant differences in systematic error, precision, and accuracy in the angular deviation between resident and attending. All breaches were <1 mm (1 attending, 2 resident). This cadaveric study identified that increased surgical experience may increase accuracy and precision when placing screws in the clinically important medial-lateral direction of the pedicle when using RAN. Although the fourth-year orthopedic surgery resident with a spine rotation as well as an introduction to robotic surgery rotation placed clinically safe pedicle screws using RAN without breach, this study demonstrates that 3D accuracy metrics paired with RAN can help identify specific opportunities for improvement in surgical technique for trainees.Level of evidence II.