Dynamic Radiographs Are Unreliable to Assess Arthrodesis Following Cervical Fusion: A Modeled Radiostereometric Analysis of Cervical Motion

Abstract

Study design: In vitro study.

Objective: The purpose of the present study was to utilize an idealized cervical spine model to determine whether the parallax effect or changes in the position of the spine relative to the x-ray generator influence intervertebral motion parameters on dynamic cervical spine radiographs.

Summary of background data: The utility of flexion-extension radiographs in clinical practice remains in question due to poor reliability of the parameters utilized to measure motion.

Materials and methods: A cervical spine model with tantalum beads inserted into the tip of each spinous process was utilized to measure interspinous process distance (IPD) on plain radiographs. The model was then manipulated to alter the generator angle and generator distance, and the IPD was measured. The impact of individual and combined changes in these parameters on IPD was assessed. Multivariate analysis was performed to identify independent drivers of variability in IPD measurements.

Results: Isolated changes in the generator distance and generator angle and combined changes in these parameters led to significant changes in the measured IPD at each intervertebral level in neutral, flexion, and extension, which, in many instances, exceeded an absolute change of >1 mm or >2 mm. Multivariate analysis revealed that generator distance and generator angle are both independent factors impacting IPD measurements that have an additive effect.

Conclusions: In an idealized cervical spine model, small clinically feasible changes in spine position relative to the x-ray generator produced substantial variability in IPD measurements, with absolute changes that often exceeded established cutoffs for determining the presence of pathologic motion across a fused segment. This study further reinforces that motion assessment on dynamic radiographs is not a reliable method for determining the presence of an arthrodesis unless these sources of variability can be consistently eliminated.

Level of evidence: 3.