What is the most accurate radiographic criterion to determine anterior cervical fusion?

Abstract

BACKGROUND CONTEXT: The accuracy of radiographic criteria for determining anterior cervical fusion remains controversial, and inconsistency in the literature makes a comparison of published clinical results problematic. The descriptions of bridging bone are still lacking and subjective, and the interpretation of images can be influenced by the type of graft or cage used. PURPOSE: To assess and validate the diagnostic accuracies of four radiographic fusion criteria using the results of surgical exploration. STUDY DESIGN: Retrospective, radiographic, and comparative study. PATIENT SAMPLE: This study included patients who required anterior or posterior exploration of a previous anterior cervical arthrodesis level(s) ranging from C3-C4 to C7-T1 for suspected pseudarthrosis or adjacent-segment pathologies. They underwent radiologic examinations to determine the four fusion criteria. We included patients whose images were taken at least 1 year after the index surgery, and 82 patients with 151 cervical segments were enrolled. OUTCOME MEASURES: The inter- and infra-rater reliabilities and validity that correlated with the results of surgical exploration for the four fusion criteria were assessed using data (fusion or not) that were collected by two raters. METHODS: The four published radiographic fusion criteria were interspinous motion (ISM) < 1 mm and superjacent ISM >= 4 mm, seen on dynamic radiographs; conventional bridging bone, as seen on computed tomography (CT) scans; and extra-graft bridging bone (ExGBB) and intragraft bridging bone (InGBB), observed on multi-axial reconstructed CT scans. The criteria were evaluated by two raters (spine surgeons with 5 and 7 years of experience). The raters evaluated each criterion twice at two different time points, 3 to 4 weeks apart. First, ISM and conventional bridging bone on CT scans were evaluated, followed by ExGBB and InGBB, with a time interval of 4 months. This Research was supported by the Chung-Ang University Research Grants (less than 5,000 US dollars) in 2016. RESULTS: The inter- and intra-rater reliability values of the ExGBB (0.887-0.933) criteria were the highest, followed by those for the ISM (0.860-0.906), bridging bone (0.755-0.907), and InGBB (0.656-0.695) criteria. The validity values that correlated with the exploration results were the highest for the ExGBB criteria (k=0.889), followed by the ISM (k=0.776), bridging bone (k=0.757), and InGBB (k=0.656) criteria and ExGBB showed the highest sensitivity (91.7%) and specificity (98.4%). Regarding the graft materials that were used, all criteria had the highest values in the auto-cortical group and lowest values in the cage group. Of note, sensitivity and specificity of ExGBB were 100% in autocortical group. In the cage group, the validity values for the ExGBB (k=0.663) and ISM (k=0.666) criteria were higher than those for the bridging bone (k=0.504) and InGBB (k=0.308) criteria CONCLUSION: The presence of ExGBB (anterior, posterior, or lateral to the graft or cage) correlated the best with surgical exploration. The ISM criteria demonstrated a similar accuracy to that of conventional bridging bone criteria on CT scans. In arthrodesed segments with auto-cortical bone, criteria showed the highest validity values. In cage group, ISM and ExGBB had acceptable accuracy, but the conventional bridging bone and InGBB were worse than guessing. We recommend that ISM and ExGBB criteria should be used to increase accuracy in patients who undergo arthrodesis with cages. (C) 2018 Elsevier Inc. All rights reserved.