The accuracy of 3D image navigation with a cutaneously fixed dynamic reference frame in minimally invasive transforaminal lumbar interbody fusion

Objective: In contrast to preoperative image-based 3D navigation systems, which require surgeon-dependent registration, an intraoperative cone-beam computed tomography (cb-CT) image-based 3D navigation system allows automatic registration during the acquisition of 3D images intraoperatively. Thus, the need for spinal exposure for point matching is obviated, making a cb-CT image-based navigation system ideal for use in minimally invasive spinal procedures. Conventionally, the dynamic reference frame (DRF) is mounted to an adjacent spinous process or iliac bone through a separate incision. However, the close proximity of the DRF to the surgical area may result in its interfering with the surgical procedure or causing streak artifacts on the navigation images. Cutaneous placement of the DRF overlying the sacral hiatus is one possible solution to these problems, but such a placement does not provide a solid bony fixation point and is distant from the surgical area, both of which factors may hinder the accuracy of the navigation. The purpose of this study was to evaluate the accuracy of a novel idea for DRF placement in a series of mini-open transforaminal lumbar interbody fusion (TLIF) procedures performed with intraoperative cb-CT image-based 3D navigation. Methods: From June 2009 to December 2009, 20 patients underwent mini-open TLIF for a total 82 pedicle screws placed in the lumbar spine with cutaneous placement of the DRF overlying the sacral hiatus. The pedicle screws were inserted under navigational guidance using cb-CT data acquired intraoperatively with a Medtronic O-arm. Screw positions were subsequently checked with a final intraoperative cb-CT scan. Nineteen patients underwent single-level fusion (8 at L4–5, 6 at L5–S1, 4 at L3–4, and 1 at L2–3) and one patient underwent two-level fusion (from L3–5). Results: There were 4 (4.9%) pedicle perforations greater than 2 mm out of the 82 pedicle screw insertions in the 20 patients. Two of these breached screws were repositioned and confirmed to be in place with a final intraoperative cb-CT. There were no complications of neural injury associated with these perforations. Conclusion: A cutaneously mounted DRF overlying the sacral hiatus provides accuracy in intraoperative 3D image guided navigation for mini-open TLIF that is comparable to that obtained in other reported series using a fixed bony attachment point for the DRF.