Abstract
The purpose of this research was to determine the feasibility of radiostereometric analysis (RSA) as a diagnostic tool for assessing non-union following spinal arthrodesis procedures. Further, to estimate clinical thresholds for precision and accuracy of the proposed method in the cervical and lumbar spine.
A three-level lumbo-sacral and a four-level cervical posterior arthrodesis procedures were performed on an artificial spine model (Sawbones, WA). Using a spring loaded inserter (RSA Biomedical, Sweden), eight to ten RSA markers were placed within each of the L4 and L5 segments in the spinous process (L4 only), lamina, transverse processes, posterior and anterior (down the pedicle) wall of the vertebral body. Eight to ten markers were placed within the proximal sacrum (S1) at the medial and lateral crests, tuberosity, and within the sacral canal wall. Four to eight RSA markers were placed into the C3-C6 lateral masses. Titanium screws and rods were applied to the spinal segments. Identical procedures were then performed on a cadaveric spine using similar bead placement and hardware.
RSA imaging consisted of 12 double exams (24 exams) of the cervical and lumbar regions for both the Sawbones and cadaveric spine to assess precision of measurement under zero-displacement conditions. The most distal vertebrae were considered the datum against which the movement of all other vertebrae was compared.
The artificial spine was then dismantled for accuracy assessment in which the middle vertebrae (L5 and C4-C5) were moved relative to the superior (L4 and C3) and inferior (S1 and C6) vertebrae by known, incremental displacements on an imaging phantom device. Displacements occurred along the superior-inferior, anterior-posterior, and flexion-extension (rotational) axes of motion. RSA images were obtained at each displacement.
Image analysis was performed using model-based software (RSACore v3.41, Leiden, Netherlands) to visualize implanted RSA beads in 3-D space. Precision was defined as the 95% confidence interval of error in measuring zero-displacement. Accuracy was defined as the mean difference (with 95% confidence interval) between the known and measured displacement.
The rate of RSA bead detection was high with 5–8 implanted beads being visible in both the lumbar and cervical regions of the artificial and cadaveric spines.
Translational RSA precision for both spines was better than 0.25 mm and 0.82 mm for the lumbar and cervical regions, respectively. Rotational precision was better than 0.40° and 1.9° for the lumbar and cervical regions, respectively. RSA accuracy for the artificial spine overall demonstrated less than 0.11 mm translational bias (margin < ±0.02 mm) and less than 0.22° rotational bias (margin < ±0.15°).
This study demonstrates that RSA achieves sufficient precision and accuracy to detect intervertebral micromotion for the purpose of assessing arthrodesis. Well dispersed RSA bead placement is critical to achieving sufficient accuracy as well as avoiding occlusion by metal hardware. Cervical bead implantation is particularly sensitive to bead clustering due to small vertebrae size and proximity to critical structures. The results of this work will aid in the development of a clinical study to assess arthrodesis in patients.
