Metal artifact reduction by dual-energy computed tomography using energetic extrapolation: a systematically optimized...

Energetic extrapolation is a promising strategy to reduce metal artifacts in dual-source computed tomography (DSCT). We performed this study to systematically optimize image acquisition parameters for this approach in a hip phantom and assess its value in a clinical study.

Titanium and steel hip prostheses were placed in a standard hip phantom and a water tank and scanned on a DSCT scanner. Tube spectra, tube current ratio, collimation, pitch, and rotation time were optimized in a stepwise process. Artifacts were quantified by measuring the standard deviation of the computed tomography density in a doughnut-shaped region of interest placed around the prosthesis. A total of 22 adult individuals with metallic implants referred for computed tomography for a musculoskeletal indication were scanned using the optimized protocol. Degree of artifacts and diagnostic image quality were rated visually (0-10) and maximum streak intensity was measured.

Sn140/100 kVp proved superior to Sn140/80 kVp. There was a benefit for increasing tube current ratio from 1:1 to 3:1, but not beyond, in favor of the Sn140 kVp spectrum. Artifacts were less severe for a collimation of 32 × 0.6 mm as compared with 40 × 0.6 mm. A pitch of 0.5 at a rotation time of 0.5 seconds per rotation was preferable to other combinations with comparable scanning times. In the clinical study, increasing the extrapolated photon energy from 64 to 120 keV decreased the severity of artifacts from 8.0 to 2.0 (P < 0.001) and decreased streak intensity from 871 to 153 HU (P < 0.001). The median diagnostic image quality rating improved from 2.5 to 8.0 (P < 0.001). The median energy level visually perceived as optimal for diagnostic evaluation was 113 keV (range, 100-130 keV).

Sn140/100 kVp with a tube current ratio of 3:1, a collimation of 32 × 0.6 mm, and extrapolated energies of 105 to 120 keV are optimal parameters for a dedicated DSCT protocol that effectively reduces metal artifacts by energetic extrapolation. The protocol effectively reduces metal artifacts in all types of metal implants. The optimized reconstructions yielded relevant additional findings.