Radiographic Evaluation of the Postoperative Interbody Fusion Patient: Is CT the Study of Choice?

Williams et al are to be complimented for an excellent and thorough review of devices available for interbody fusion and of the radiographic findings pertinent to the evaluation of fusion healing. As the authors point out, interbody fusions are being performed by spinal surgeons with increasing frequency, and the interpretation of postoperative radiographs is an increasingly common and important task for the radiologist.

The authors emphasize the importance of CT in evaluating interbody fusions and have proposed a radiographic protocol consisting of scans done at 3, 6, and 12 months, with an additional scan at 24 months if a solid fusion is not seen earlier. No conventional radiographs are obtained. No alteration in the protocol is proposed, regardless of the device or material placed in the disk space. The same protocol is recommended, whether a metal cage is placed with bone morphogenic protein inside the cage or a segment of allograft bone is placed to fill the intervertebral space.

The routine adoption of this protocol requires discussion. Replacing conventional radiographs with CT scans and reconstructions entails a substantial increase in cost, 8-fold at this writer’s institution, yet the authors provide no data to validate their protocol and few references to prove that CT offers a benefit worthy of the increased cost.

The reliability and sensitivity of CT scans or conventional radiographs for identifying fusion healing is not known. One study evaluating fusions in an animal model found a congruency between the extent of bony fusion in CT imaging and histologic assessment of only 14% (1). In that study, CT images significantly overestimated the extent of fusion but accurately identified the presence of a fusion in 83% of specimens.

There is no doubt that CT offers some advantage over conventional radiographs, especially with regard to identifying lucency around hardware and cystic changes within the endplates. One might anticipate, however, that other findings predictive of fusion failure such as subsidence, translation, or other change in alignment would be more likely to be identified on weight-bearing films, yet these are not part of the recommended protocol. Change in alignment on bending films also indicates implant loosening, but these, too, are not part of the authors’ protocol.

One may also question why the authors chose to perform scans at 3-, 6-, and 12-month intervals. Regardless of the device or graft type used (cancellous bone from the iliac crest, allograft bone, or bone morphogenic protein are the most common options), no patient is likely to show healing of the fusion as early as 3 months after surgery, so a scan at this early stage seems unnecessary. The authors indicate that this early scan is helpful in determining whether the patient can safely return to work; presumably if loosening of the implants is seen, the patient would not be allowed to return to full duty; however, is the more expensive CT really better than conventional radiographs at identifying loosening? The references provided only suggest that CT is better at demonstrating bridging of the fusion bone—in other words, solid fusion healing. There is little evidence that CT better demonstrates loosening—or, for that matter, that restricted activity will allow solid fusion once the implants have loosened.

If a solid fusion is seen at 6 months, what is the benefit of the 12-month scan? Why do both, at twice the cost? Let us accept the assumption that a scan at 3 months showed no loosening and the patient returned to full activity. If the patient’s clinical status is successful and he or she is free of pain, why obtain a scan at 6 months? For that matter, why obtain any radiographic study in this asymptomatic patient? What radiographic finding would change the asymptomatic patient’s management? If the goal of the radiographic study is to prove that solid healing of the fusion was accomplished, why not wait until the 12-month interval when a higher percentage of patients destined for success are likely to have achieved a solid fusion? If the graft material used was allograft bone, an even longer interval might be appropriate, as this is the slowest to heal.

Finally, although it is clear that CT offers an enormous advantage by making it possible to visualize bone graft within a metal cage, its benefit is less obvious when the cage is made of a radiolucent material. With carbon fiber or polyetheretherketone cages, anteroposterior radiographs obtained with the radiographic beam oriented in line with the disk space can clearly visualize the intervertebral space. If trabeculae of bone are seen bridging the disk space without a gap, it is clear that fusion has been achieved without the need for a reconstructed CT (2). This is even more convincing when bone morphogenic protein is used, because early radiographs show lucency in the disk space, which is replaced by bone as healing occurs. Even when a metal cage is used, if bone is seen bridging the interspace anterior to the cage on a lateral view, it is clear that healing of the fusion has occurred. This has been described as the “sentinel” sign (3). A CT is not needed when these findings can be visualized on conventional radiographs.

The protocol proposed by Williams et al provides a reasonable starting point for postoperative evaluation of the patient who has undergone interbody spinal fusion; however, in light of the increased costs associated with CT imaging, as well as the advantage of plain radiographs for certain findings, it seems reasonable to individualize the postoperative radiographic studies obtained, rather than routinely obtaining 3 or 4 CT scans on each patient. The type of studies obtained and the timing of those studies should be based on the patient’s clinical status, the type of device used, and the type of graft placed within the device. In this way, our limited resources for health care expenditure can be more effectively used and costs controlled.