Introduction: Although previous studies reported that robot-assisted spine surgery have helped improve accuracy of pedicle screw placement, it still remains undetermined whether such assistant devices can be alternatives to spine surgery experts. This study has verified if the accuracy varies among three groups of operators with different years of experience in this field: expertly trained spine surgeons (group A), spine surgeons less than 3 years of experience (group B), and inexperienced trainees (group C).

Methods: Forty patients (21 men), aged 12 to 89 (mean, 72.5) years underwent a total of 264 percutaneous pedicle screw (PPS) placements (40 each for cervical and thoracic spines and 184 for lumbosacral spines and pelvis) using the Cirq robotic arm system coupled with intraoperative CT-based navigation. With the patient positioned and prepared on the operating table, we attached the reference array to the two 3 mm-diameter K-wires inserted into the iliac bone unilaterally followed by registration CT scanning for the thoracic/lumbosacral PPS and S2 alar-iliac screw placements. For the cervical PPS placements, we affixed the reference array to the C2 spinous process or the Mayfield tongs. When carrying out PPS insertion, the surgeon located the entry area with blunt finger dissection through a 2-cm incision first, then he/she moved a robot arm thereto manually. Subsequently, the robot arm, when activated, automatically positioned to align to the preplanned pedicle trajectory. The arm locked in place, with a drill-stabilizing metallic tube assembled manually at its end, allowed a desired drilling and guidewire insertion. In the cervical vertebrae, we employed a posterolateral intermuscular approach through a small incision described by Tokioka et al.

Results: Considering no pedicle wall breach for cervical PPSs and both no breach and breach < 2 mm for thoracic/lumbosacral and iliac PPSs as successful placements, overall success rate accounted for 93.6% (247 of 264 screws): 95.0% for cervical, 85.0% for thoracic, and 95.1% for lumbosacral PPS and alar-iliac screw placements. No patients required revision surgery for repositioning or replacing the PPSs.

The success rates showed no significant difference among three groups of operators: 94.4% for group A, 92.7% for group B, and 93.9% for group C.

Conclusions: The surgeon-controlled robotic arm with intraoperative CT-based navigation served well for PPS placements at any spinal level including cervical spine. By using this device, safety and accuracy of PPS placements did not significantly vary depending on surgeon's years of experience in spinal surgery.