Robotic-Guided Spine Surgery: Implementation of a System in Routine Clinical Practice-An Update.

Objective

The aim of this study is to present the initiation of robotic-guided (RG) spine surgery into routine clinical care at a single center with the use of intraoperative CT (iCT) automatic registration-based navigation. The workflow included iCT with automatic registration, fusion with preoperative imaging, verification of preplanned screw trajectories, RG introduction of K-wires, and the insertion of pedicle screws (PSs), followed by a control iCT scan.

Methods

All patients who underwent RG implantation of pedicle screws using the Cirq® robotic arm (BrainLab, Munich, Germany) in the thoracolumbar spine at our department were included in the study. The accuracy of the pedicles screws was assessed using the Gertzbein-Robbins scale (GRS).

Results

In total, 108 patients (60 female, mean age 68.7 ± 11.4 years) in 109 surgeries underwent RG PS placement. Indications included degenerative spinal disorders (n = 30 patients), spondylodiscitis (n = 24), tumor (n = 33), and fracture (n = 22), with a mean follow-up period of 7.7 ± 9 months. Thirty-seven cases (33.9%) were performed percutaneously, and all others were performed openly. Thirty-three operations were performed on the thoracic spine, forty-four on the lumbar and lumbosacral spine, thirty on the thoracolumbar, one on the cervicothoracic spine, and one on the thoracolumbosacral spine. The screws were inserted using a fluoroscopic (first 12 operations) or navigated technique (latter operations). The mean operation time was 228.8 ± 106 min, and the mean robotic time was 31.5 ± 18.4 min. The mean time per K-wire was 5.35 ± 3.98 min. The operation time was lower in the percutaneous group, while the robot time did not differ between the two groups. Robot time and the time per K-wire improved over time. Out of 688 screws, 592 were GRS A screws (86.1%), 54 B (7.8%), 22 C (3.2%), 12 D (1.7%), and 8 E (1.2%). Seven screws were revised intraoperatively, and after revision, all were GRS A. E screws were either revised or removed. In the case of D screws, screws located at the end of the construct were revised, while so-called in-out-in screws in the middle of the construct were not revised.

Conclusions

Brainlab's Cirq® Robotic Alignment Module feature enables placement of pedicle screws in the thoracolumbar spine with high accuracy. A learning curve is shown through improvements in robotic time and time per K-wire.