Metal implants on CT: comparison of iterative reconstruction algorithms for reduction of metal artifacts with single energy and spectral CT scanning in a phantom model.

Abstract

To assess single energy metal artifact reduction (SEMAR) and spectral energy metal artifact reduction (MARS) algorithms in reducing artifacts generated by different metal implants. Phantom was scanned with and without SEMAR (Aquilion One, Toshiba) and MARS (Discovery CT750 HD, GE), with various metal implants. Images were evaluated objectively by measuring standard deviation in regions of interests and subjectively by two independent reviewers grading on a scale of 0 (no artifact) to 4 (severe artifact). Reviewers also graded new artifacts introduced by metal artifact reduction algorithms. SEMAR and MARS significantly decreased variability of the density measurement adjacent to the metal implant, with median SD (standard deviation of density measurement) of 52.1HU without SEMAR, vs. 12.3HU with SEMAR, p<0.001. Median SD without MARS of 63.1HU decreased to 25.9HU with MARS, p<0.001. Median SD with SEMAR is significantly lower than median SD with MARS (p=0.0011). SEMAR improved subjective image quality with reduction in overall artifacts grading from 3.2±0.7 to 1.4±0.9, p<0.001. Improvement of overall image quality by MARS has not reached statistical significance (3.2±0.6 to 2.6±0.8, p=0.088). There was a significant introduction of artifacts introduced by metal artifact reduction algorithm for MARS with 2.4±1.0, but minimal with SEMAR 0.4±0.7, p<0.001. CT iterative reconstruction algorithms with single and spectral energy are both effective in reduction of metal artifacts. Single energy-based algorithm provides better overall image quality than spectral CT-based algorithm. Spectral metal artifact reduction algorithm introduces mild to moderate artifacts in the far field.