Validation of a freehand technique for cortical bone trajectory screws in the lumbar spine

Restricted access

OBJECTIVE

The cortical bone trajectory (CBT) technique for pedicle screw placement has gained popularity among spinal surgeons. It has been shown biomechanically to provide better fixation and improved pullout strength compared to a traditional pedicle screw trajectory. The CBT technique also allows for a less invasive approach for fusion and may have lower incidence of adjacent-level disease. A limitation of the current CBT technique is a lack of readily identifiable and reproducible visual landmarks to guide freehand CBT screw placement in comparison to the well-defined identifiable landmarks for traditional pedicle screw insertion. The goal of this study was to validate a safe and intuitive freehand technique for placement of CBT screws based on optimization of virtual CBT screw placement using anatomical landmarks in the lumbar spine. The authors hypothesized that virtual identification of anatomical landmarks on 3D models of the lumbar spine generated from CT scans would translate to a safe intraoperative freehand technique.

METHODS

Customized, open-source medical imaging and visualization software (3D Slicer) was used in this study to develop a workflow for virtual simulation of lumbar CBT screw insertion. First, in an ex vivo study, 20 anonymous CT image series of normal and degenerative lumbar spines and virtual screw insertion were conducted to place CBT screws bilaterally in the L1–5 vertebrae for each image volume. The optimal safe CBT trajectory was created by maximizing both the screw length and the cortical bone contact with the screw. Easily identifiable anatomical surface landmarks for the start point and trajectory that best allowed the reproducible idealized screw position were determined. An in vivo validation of the determined landmarks from the ex vivo study was then performed in 10 patients. Placement of virtual “test” cortical bone trajectory screws was simulated with the surgeon blinded to the real-time image-guided navigation, and the placement was evaluated. The surgeon then placed the definitive screw using image guidance.

RESULTS

From the ex vivo study, the optimized technique and landmarks were similar in the L1–4 vertebrae, whereas the L5 optimized technique was distinct. The in vivo validation yielded ideal, safe, and unsafe screws in 62%, 16%, and 22% of cases, respectively. A common reason for the nonidealized trajectories was the obscuration of patient anatomy secondary to severe degenerative changes.

CONCLUSIONS

CBT screws were placed ideally or safely 78% of the time in a virtual simulation model. A 22% rate of unsafe freehand trajectories suggests that the CBT technique requires use of image-guided navigation or x-ray guidance and that reliable freehand CBT screw insertion based on anatomical landmarks is not reliably feasible in the lumbar spine.

ABBREVIATIONS CBT = cortical bone trajectory; LIS = less invasive surgery.

Article Information

Correspondence Joel Finkelstein: Sunnybrook Health Sciences Centre, Toronto, ON, Canada. joel.finkelstein@sunnybrook.ca.

INCLUDE WHEN CITING Published online April 19, 2019; DOI: 10.3171/2019.1.SPINE181402.

Disclosures Institutional educational support was received from Stryker Canada and Zimmer/Biomet.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    The CBT was defined using a 3D virtual probe attached to a fiducial entry point. The position of the probe is shown as a 2D projection on the CT scan to identify an optimal trajectory similar to using image-guided navigation.

  • View in gallery

    Direct lateral view of L3 using the 3D volume rendering of a nonpathologic thoracolumbar spine CT. Note the overlapping transverse foramina. The spinous process is divided into thirds along the superior to inferior axis. These spinous process landmarks are used to determine the trajectories of screw insertion for each corresponding vertebral level.

  • View in gallery

    Representative insertional screw trajectories of L3 using the combined 3D surface model with volume rendering of a nonpathologic lumbar spine CT. The entry point as previously defined for L1–4 has been utilized. The virtual awl trajectory is to touch the base of the probe to barely come in contact with the inferior 1/3–2/3 junction of the spinous process of the same level. The start point of the CBT screw for the L1–4 is 2.5 mm medial to the lateral margins of the inflection point between the pars interarticularis and the superior articular facet.

  • View in gallery

    Direct posterior view of L5 using the combined 3D surface model with volume rendering of a nonpathologic lumbar spine CT. The entry point of the transcortical screw for L5 is defined as the 75% mark between the midline and the inflection point between the pars interarticularis (PARS) and the superior articular facet (FACET), biased laterally. SAP = superior articular process.

  • View in gallery

    Representative insertional screw trajectory of L5 using the 3D volume rendering of a nonpathologic lumbar spine CT. The start point as previously defined for L5 has been utilized. The virtual awl trajectory is to touch the base of the probe to barely come in contact with the inferior 1/3–2/3 junction of the spinous process of the same level. The base of the probe is lateralized until positioned at the 50% mark between contact with the spinous process and fully vertical in the axial plane.

  • View in gallery

    Axial (left) and sagittal (right) virtual projections and actual placement of L4 CBT screws. The left pedicle screw has a perfect match between the freehand directed CBT “test” screw (yellow) and the actual placement of the screw (blue). The right pedicle screw has overlateralization of the freehand test screw (pink) compared to the actual placement of the screw (red). This is safe, but not ideal.

References

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 1300 1300 1262
Full Text Views 93 93 87
PDF Downloads 87 87 80
EPUB Downloads 0 0 0

PubMed

Google Scholar