Do CT Scans Overestimate the Fusion Rate After Anterior Cervical Discectomy and Fusion?

Design: This study is a radiographic analysis. Objective: To compare the fusion rates after anterior cervical discectomy and fusion (ACDF) using x-rays versus computerized tomography (CT). Background: Although fusion status may be obvious when evaluating ACDFs performed in the remote past, determining the presence of a solid fusion at earlier time points after ACDF is often ambiguous but a necessary part of practice. Commonly used tools include radiographs and CT scans. Currently, there is no gold standard imaging modality to determine fusion status. Methods: Twenty-two patients status post-ACDF (cortical allograft with anterior plates) at 34 levels with CT scans and dynamic x-rays obtained at 3, 6, and 12 months postoperatively were included. Four spine surgeons blinded to the time point independently determined fusion status according to the criteria. Results: On the basis of the x-ray criteria, the fusion rates were 26%, 41%, and 65% at 3, 6, and 12 months, respectively, postoperatively. On the basis of CT criteria, the fusion rates were 79%, 79%, and 91% at 3, 6, and 12 months, respectively. There was a significant difference in the predicted fusion rate at each time point comparing x-ray versus CT criteria. In addition, at 3 months, 41% of the levels (11/27) thought to be fused by CT criteria demonstrated >1 mm motion on dynamic x-rays. At 6 months, 33% (9/27) of the levels thought to be fused by CT demonstrated persistent motion of >= 1 mm. At 12 months, 23% (7/31) of the levels considered fused by CT still had persistent motion. Discussion: X-ray criteria for fusion, which incorporate both static and dynamic factors, predicted lower fusion rates at each time point when compared with CT scans, which evaluate only static factors. Depending on the time point, anywhere from 23% to 41% of levels thought to be fused by CT criteria demonstrated persistent motion on dynamic x-rays. Although <1 mm motion is not a sufficient criteria for fusion by itself, levels demonstrating >1 mm motion are less likely to be solidly fused. Thus, we conclude that CT scans may overestimate the fusion rate during the early stages of ACDF healing with cortical allograft, and that CT scans alone may not accurately determine fusion status. Reliable determination of fusion may thus require dynamic information obtained from flexion-extension x-ray in association with high-resolution static information from CT.