Patients
The study was approved by the Bumrungrad International Institutional Review Board (approval number 278-01-20 CDEK-H Exp).
We retrospectively reviewed data from 2,098 pedicle screw procedures and the fusion rate in 475 consecutive patients who underwent MIS-TLIF with surgical imaging (O-arm Navigation System, Medtronic Inc., Dublin, Ireland) and computerized navigation (StealthStation Surgical Navigation System, Medtronic Inc.) systems at Bumrungrad International Hospital between October 2013 to December 2017 by the same surgical team. Patients were included if they were diagnosed with spondylolisthesis, disc herniation, recurrent disc herniation, foraminal stenosis, or recurrent facet cyst. The exclusion criteria were incomplete patient data or missing O-arm scan images.
Surgical procedure
Patients were positioned prone on the operating table with a Jackson frame. A reference pin was fixed securely to the posterior superior iliac spine. Intraoperative 3D images were made after the O-arm scan. The skin incision for screw placement was made according to the optimal entry point for pedicle screw placement indicated by a probe. The Wiltse approach was performed. The size and trajectory of the screw were selected during the tapping procedure by sagittal, axial, and coronal images on the StealthStation [Fig. 1]. The preparation of screw placement was completed before all decompression. However, only pedicle screws on the non-decompression side (non-TLIF approach side) were first manually inserted without K-wire under navigation control before the decompression procedure started.
For decompression, an incision was made according to the intraoperative imaging. The microscopic decompression started from the facet removal on the side of the symptoms to identify the traversing and exiting nerve roots. After decompression, the polyetheretherketone cage with the autogenous bone graft and the bone substitute was packed into the disc space. Cross decompression, if needed, was performed. The pedicle screw on the decompression side, prepared and selected according to size, was then inserted under navigation control. The O-arm scan was redone to check the accuracy of all screw placements. The positions of the screws were evaluated based on the O-arm scan images in the sagittal, axial, and coronal planes. Any unacceptable pedicle breach; >4mm breach lateral direction or > 2 mm breach medial direction detected was revised immediately and another O-arm scan, which was performed to confirm the final pedicle screw position. Once the positions of the screws were accepted by the surgical team, rods were inserted. All wounds were closed in layers [Fig. 2].
Evaluation of screw placement
The intraoperative O-arm images were evaluated by two independent reviewers. If the cortical layer was breached by the pedicle screw, the breaches were assessed in the axial, sagittal, and coronal planes using the Gertzbein-Robbins classification system[11]. Specifically, pedicles that did not breach the cortical layer were defined as grade A, breaches < 2 mm as grade B, breaches < 4 mm as grade C, breaches < 6 mm as grade D, and breaches ≥ 6 mm as grade E [Fig. 3a,b,c]. The direction of screw misplacement was defined as medial, lateral, inferior, or superior [Fig. 3b, c]. In cases where the breach grade was questionable, the poorer grade was selected for data analysis.
Evaluation of fusion rate
The postoperative computed tomography (CT) scan at the 2-year follow-up was assessed by two independent reviewers. The fusion was graded according to Bridwell et al. as follows [19]: fused with remodeling and present trabeculae (grade I) [Fig. 4a]; graft intact, not fully remodeled or incorporated, but no lucency present (grade II) [Fig. 4b]; and graft intact, potential lucency present at top and bottom of the graft (grade III) [Fig. 4c], and fusion absent with collapse/resorption of the graft (grade IV) [13].
Statistical analysis
Frequency tables and measures of central tendencies were used to summarize descriptive statistics. Parametric and nonparametric tests were used for continuous quantitative (visual analog scale [VAS] back, VAS leg, Oswestry Disability Index [ODI] score, length of hospital stay, operative time, and radiation time) and qualitative (intraoperative and postoperative complications) variables, respectively. The level of statistical significance was set to p < 0.05 for all statistical analyses. Statistical analysis was performed using Stata Statistics Version 14 software (StatCorp LLC, College Station, TX, USA).