gms | German Medical Science

Objective: The theoretical advantage of image-based robotic assistance in lumbar spinal instrumentation has not been scientifically proven so far. Two pilot studies focusing on the only CE approved miniature-robot for spine surgery provided excellent results for screw positions. The present study is the first prospective randomised comparison of robot-assisted versus free-hand lumbar pedicle screw implantation.

Methods: Patients were either randomized to robot-assisted lumbar pedicle screw placement using SpineAssist (Mazor Surgical Technologies, Israel) or free-hand pedicle screw placement. All patients were planned for mono- or bisegmental lumbar instrumentation with a pedicle-screw based internal fixator system. According to the study protocol the following parameters were assessed: duration of surgery, duration of preparation in- and outside the OR, radiation exposure of the patient and accuracy of the screw position evaluated by postoperative thin-cut CT scans.

Results: Fifty-four patients have been successfully randomized. The average age in the robot-assisted group with n=27 patients (17 females and 10 males) was 65 ± 21 years. In the free-hand group we could include n=27 patients (16 females and 11 males) with an average age of 65 ± 20 years. Time for preparation (robot-assisted 24 min ± 9 vs. free-hand 13 min ± 13) and the operating time itself (robot-assisted 140 min ± 24 vs. free-hand 152 min ± 43) showed equal data. But in addition, there was the necessity for planning with a mean duration of 23 min ± 12 per operation and a higher radiation exposure because of a preoperatively CT-scan (430,10 ± 159.13 [mGy*cm]) in the robot-assisted group only. Eight (5.8%) of the 130 free-hand implanted pedicle screws had a deviated position. Revision was necessary for one (0.7%) pedicle screw. In contrast, 14 (12%) of the 114 robot-assisted screws were placed suboptimal and 10 (8.7%) required revision.

Conclusions: In the first single-center prospective randomized controlled trial we could not detect an advantage of the robot-assisted lumbar pedicle screw implantation versus free-hand placement . Moreover, in the present series the robot-assisted screws showed a higher percentage of incorrect position. Underlying factors might be the method used for robot attachment to the patient and slipping of the percutaneously inserted drill guides at the entry points. We believe that an improvement of those factors could significantly increase the accuracy of the system.