Optimization of Monoenergetic Extrapolations in Dual-Energy CT for Metal Artifact Reduction in Different Body Regions and Orthopedic Implants

Abstract

To define an optimal monoenergetic extrapolation (ME) in dual-energy computed tomography (DECT) for metal artifact reduction (MAR) including different body regions and orthopedic implants. DECT scans were acquired with dual-source CT (SOMATOM Force, Siemens, Germany) at tube voltage A 80-100 kV/B Sn150 kV from 39 patients (mean 54.1 ± 20.7 years, 23 male vs. 16 female) with orthopedic implants ranging from wires to joint implants. Scans were assembled in four groups based on scan regions and volume. Single- and weighted-energy images at a ratio of 0.3 and MEs at 100, 130, 160, and 190 keV were produced using vendor-specific postprocessing software (Syngo.Via, Siemens, Germany). Artifact degree was assessed quantitatively by metal-induced Hounsfield unit changes in relation to reference tissues. Visibility of screw-bone interface, hardware integrity, adjacent bone, and soft tissues were visually rated on a four-point Likert scale (0, none; 3, strong artifacts with nondiagnostic quality). Optimal energy was visually determined by side-by-side comparisons. Artifact degree was statistically compared between regions and energies. Metal-induced attenuation changes were most severe in large scan volume groups for all energies. Reference tissue attenuation outside metal artifacts was not affected by ME (p = 0.57). Independent of region, ME at 130-190 keV quantitatively performed significantly better for MAR than the remainder. ME 130 keV showed the highest frequency (54%) in optimal energy ratings based on qualitative image criteria. DECT significantly reduces image artifacts in patients with orthopedic hardware and prospective choice of ME at 130 keV may suit best for optimal MAR, independent of region or implant.