A phantom and cadaveric study of radiostereometric analysis in posterior cervical and lumbar spinal fusion

BACKGROUND CONTEXT: Detecting pseudarthrosis following spinal fusion is important for accurate diagnosis and treatment. Current diagnostic measures hold certain drawbacks. Radiostereometric analysis (RSA) is a radiographic technique with the capability to measure intervertebral segment changes and may be a novel way of assessing fusion. PURPOSE: The purpose of this work was to measure the accuracy and precision of RSA in instrumented posterior cervical and lumbar spinal fusion for measuring intervertebral movement. Further, to gain surgical practice with RSA in spine and determine optimal bead placements. STUDY DESIGN: Artificial bone and cadaveric spine models were used to simulate a 3-level cervical (C3-C6) and a 2-level (L4-S1) lumbosacral posterior spinal fusion to analyze bead placements and to measure RSA accuracy and precision. METHODS: Preliminary RSA bead placements were planned and measured in the artificial model. Secondary bead placements were adjusted slightly in the cadaveric model to consider additional fusion scenarios. Bead spread, detectability, and stability were measured to determine optimal placements. Translational and rotational precision of both models were measured. Accuracy was measured in the artificial spine model. Model-based RSA software was used for analysis. RESULTS: Optimal bead placements were found to be throughout the lateral mass of C3-C6 and in the spinous process, transverse process, and within the screw canal of L4-S1. Detectability was high among all segments. Spread was greater in L4-S1 than C3-C6 due to bead collinearity along the transverse axis of the cervical vertebrae. Translational and rotational RSA accuracy in cervical and lumbosacral regions ranged between 0.005 to 0.014 mm and 0.058 to 0.208°. Translational and rotational precision measured in the phantom models ranged 0.017 to 0.131 mm and 0.058 to 0.394° in C3-C6, and 0.086 to 0.191 mm and 0.200 to 0.369° in L4-S1. Translational and rotational precision measured in the cadaveric models ranged 0.054 to 0.548 mm and 0.148 to 1.386° in C3-C6, and 0.068 to 0.164 mm and 0.100 to 0.270° in L4-S1. CONCLUSIONS: RSA was found to be a feasible radiographic technique in C3-C6 and L4-S1 spinal fusion when measured in artificial and cadaveric models. Optimal bead placements were determined. Bead spread was shown to be better throughout the lumbar region than the cervical region due to anatomical size variations. RSA accuracy and precision were within acceptable RSA criteria. CLINICAL SIGNIFICANCE: The results from this work contribute to the accuracy, precision, and bead placements for studying RSA in cervical and lumbar spinal fusions. This work may further support the development of clinical studies to assess spinal fusion by evaluating postoperative intervertebral movement using RSA.