Between 11/2005, and 9/2006 a first series of patients has undergone transpedicular instrumentation with 3D robotic assistance in the lumbar spine at our Orthopaedic Department. This technology must not be confused with standard spine navigation and will be presented in detail. 16 patients (12m, 4f, avg. age 55 yrs.) were randomly selected from our clientele for lumbar fusion or dynamic stabilization via transpedicular instrumentation. After informed consent they obtained thin slice CT scans of the operating field prior to surgery. The Mazor computer system then imported the scans, allowing 3-D planning of screw placement. A fixation device was then attached to the patient and the system was calibrated in connection with a standard fluoro-scope. On the device a robotic device with a working arm was mounted. Automatic matching algorithms then moved the robot, pointing its arm towards the designated pedicle screw portals. The screws could then be placed through the working arm, either cannulated (ICON) via K-wires, or solid (XIA) via standard awls. Percutaneous MIS insertion was also feasible. Instrumentation was then set forth after removal of the robot as usual. The CT accuracy of screw placement in all robot-assisted patients was scored according to Mattes et al. postoperatively. 1 patient had to be instrumented manually for reasons unrelated to the system. In 2 early obese patients the system denied robotic access due to insufficient imaging, thus enforcing standard manual technique. In the remaining patients a total of 58 screws had to be placed. No clinical complication related to the Mazor system occurred. A total of 6 screws could not be placed by the system due to steep lumbosacral angles. Additional time of surgery could be reduced to 40 minutes per case during the series. None of the robotic screws was misplaced in the final CT. 1 of the 6 non-robotic screws was misplaced at the S1 level and needed replacement due to apparent nerve contact without palsy. The robotic screws reached an average Mattes score of 1.5 which can be considered superior to sole fluoroscopic techniques (2.5). Additional decompression did not impede the system which does not rely on surface matching. On the basis of the clinical application, additional features were developed, e.g. robot mounting wedges for hyperlordosis, and oblique fluoro view acquisition. The planning software also avoids “supercharging” of the pedicle due to screw oversize. In one case this inevitably would have happened in conventional technique. This is the first report worldwide about the beginning of robotic assisted pedicle screw placement in Europe in daily routine. The Mazor System now has proven its usefulness and potential. Additional most recent data will be available at time of presentation as the system is further evolving and under continuous use.
