MR of the Spine in the Presence of Metallic Bullet Fragments: Is the Benefit Worth the Risk?

In general, radiologists are reluctant to perform MR imaging for patients with retained metallic bullet fragments (1–3). This reluctance is heightened for those patients with retained fragments near vital and susceptible anatomic structures such as the spinal cord. This issue led us to review the medical records of 19 patients (December 1991 to May 1996) at our institution who had retained metallic fragments (presumed lead) in the region of the spine, and who were studied with MR imaging. This study was done to assess whether the information obtained from the images could justify the potential hazards related to the procedure.

There were retained metallic fragments in the cervical spine in six patients, the thoracic spine in eight patients, and the thoracolumbar and lumbar spine in two and three patients, respectively. Fifteen patients had retained bullet fragments larger than 1 cm (mean, 1.5 cm), four were inside the spinal canal, six were embedded in the spinal bony structures, and five were in the paraspinal region. Three smaller metallic fragments (0.4–0.9 cm) were located in the paraspinal soft tissues. Metallic fragments smaller than 0.4 cm were located inside the spinal canal in six patients, in the bony spinal structures in six patients, and in the paraspinal region in six patients. Six patients were quadriplegic and 12 were paraplegic; these conditions were unrelated to the history of gunshot wound injury. MR examinations were performed within 3 weeks after the time of injury in six patients while in the remaining 13, a time period of 1 month to 6 years had elapsed.

MR imaging was performed at either a 1.0 or 1.5 T with T1-weighted spin-echo and gradient echo images in the sagittal and axial plane. Additional sagittal or axial T2-weighted spin-echo and T2-weighted fast spin-echo sequences were performed in six and four patients, respectively. None of the patients imaged had retained metallic fragments. Fragments, such as the ball-bearing or Prometheus type of air gun pellets, have been shown to be strongly ferromagnetic (4).

All procedures were performed without the patients experiencing any untoward effect. All metallic foreign bodies showed mild artifact (1) approximately the same size as the metallic object imaged. In two cases, however, metallic fragments smaller than 0.4 cm were confirmed by plain radiography or CT, but were not visible on the MR images. This led to the conclusion that these bullet fragments were not ferromagnetic (1).

MR studies established diagnoses in 27 patients (five acute/subacute and 12 chronic). In those two cases in which the MR image was deemed suboptimal, artifact precluded exclusion of a spinal lesion in the region of interest. Imaging artifacts in these two patients were the result of multiple small metallic fragments associated with a dominant (>1 cm) metallic fragment. In the six recently injured patients, two had a cord contusion and two had an epidural hematoma with cord compression that was subsequently treated surgically. One patient had no identifiable lesion and one had a nondiagnostic study owing to extensive artifacts induced by the metallic fragments. In the 13 chronically injured patients, four studies yielded negative results for cyst, scarring, atrophy, or a compressive lesion. Atrophy and myelomalacia were identified in six patients, one of whom had a small nonsurgical extramedullary cyst. One patient had a bullet embedded in a cord with a cyst above and below it, and one patient was diagnosed with an epidural abscess that was subsequently treated surgically. One study was nondiagnostic because of extensive metallic artifact.

On the basis of our experience, we support the use of MR imaging for patients with retained metallic ballistic fragments in the region of the spine; the information we gained would have been difficult to obtain with other imaging techniques. Certainly, as with any patient with a metallic implant or any potentially hazardous medical device, serious consideration should be given to determine which (if any) of these patients should enter the MR environment. Results obtained from MR led to surgical intervention in three of 19 patients. No untoward effects were seen.