The objective of this study was to assess the discriminative power of dual-energy computed tomography (DECT) versus single-energy CT (SECT) to distinguish between ferromagnetic and non–ferromagnetic ballistic projectiles to improve safety regarding magnetic resonance (MR) imaging studies in patients with retained projectiles.
Materials and Methods
Twenty-seven ballistic projectiles including 25 bullets (diameter, 3–15 mm) and 2 shotgun pellets (2 mm each) were examined in an anthropomorphic chest phantom using 128-section dual-source CT. Data acquisition was performed with tube voltages set at 80, 100, 120, and 140 kV(p). Two readers independently assessed CT numbers of the projectile’s core on images reconstructed with an extended CT scale. Dual-energy indices (DEIs) were calculated from both 80-/140-kV(p) and 100-/140-kV(p) pairs; receiver operating characteristics curves were fitted to assess ferromagnetic properties by means of CT numbers and DEI.
Nine (33%) of the projectiles were ferromagnetic; 18 were nonferromagnetic (67%). Interreader and intrareader correlations of CT number measurements were excellent (intraclass correlation coefficients, >0.906; P < 0.001). The DEI calculated from both 80/140 and 100/140 kV(p) were significantly (P < 0.05) different between the ferromagnetic and non–ferromagnetic projectiles. The area under the curve (AUC) was 0.75 and 0.8 for the tube voltage pairs of 80/140 and 100/140 kV(p) (P < 0.05; 95% confidence interval, 0.57–0.94 and 0.62–0.97, respectively) to differentiate between the ferromagnetic and non–ferromagnetic ballistic projectiles; which increased to 0.83 and 0.85 when shotgun pellets were excluded from the analysis. The AUC for SECT was 0.69 and 0.73 (80 and 100 kV[p], respectively).
Measurements of DECT combined with an extended CT scale allow for the discrimination of projectiles with non–ferromagnetic from those with ferromagnetic properties in an anthropomorphic chest phantom with a higher AUC compared with SECT. This study indicates that DECT may have the potential to contribute to MR safety and allow for MR imaging of patients with retained projectiles. However, further studies are necessary before this concept may be used to triage clinical patients before MR.