Metal Artefact Reduction by Dual-energy Computed Tomography Using Monoenergetic Extrapolation: In-vitro Determination of Optimal Monoenergetic Level with Different Metallic Implants Using a Phantom Body

Vyk To,Cy Lee,Wy Wong,J. Tsang,W. Y. Chan,Oc Leung,M. Yuen,Ny Sin,Oc Li,Py Chu,Tw Yeung

DOI: https://doi.org/10.12809/HKJR1615356

2016-04-13

Abstract:Objective: To identify the optimal monoenergetic level, balancing metal artefacts, and the amount of noise present for imaging of metal implants using dual-energy computed tomography (CT) and focusing on the assessment of peri-prosthetic soft tissue. Methods: Four metallic implants commonly used in the hips were placed in a phantom body: unipolar hemiprosthesis, dynamic hip screw (DHS), intra-medullary (IM) nail, and titanium insert. The unipolar hemiprosthesis was imaged at two points: the head and stem. The head of the hemiprosthesis and DHS were imaged in two axes: one axial to and one resembling the angle at its expected position in the hip with respect to the scanner. The IM nail was assessed both at the level with and without a screw inserted. A region of interest to measure the noise level of the images was first performed with different monoenergetic levels (70-170 kV with increments of 10 kV). Four monoenergetic levels were then chosen (80, 90, 105, 120 kV) for each implant and were assessed and scored (presence of least to most artefacts: score 1-4) by nine radiologists who were blinded to the monoenergetic level. A total of eight sets of images were assessed. The scores for different monoenergetic levels were compared using analysis of variance. Results: In the first part of the experiment, the images with the least amount of noise were in the range of 85-95 kV, thus we included 90 kV among the images for subsequent scoring. The mean score for different monoenergetic levels for all implants was as follows: 3.94 for 80 kV, 2.68 for 90 kV, 1.50 for 105 kV, and 1.88 for 120 kV (p <0.001), with 105 kV having the least metal artefacts. For subgroup analysis of individual implants, 105 kV was found to produce the best quality images with a statistically significant better score for hip stem, DHS, and IM nail. 120 kV trended towards being the best monoenergetic level when imaging the hip head and the IM nail with screw where relatively more artefacts were present. 90 kV trended towards being the best monoenergetic level when imaging the titanium insert where artefacts were nearly absent. Conclusion: With regard to imaging the soft tissue around a metallic implant, the overall optimal monoenergetic level for reduction of metal artefacts using dual-energy CT is 105 kV. When more artefacts are inherently present, 120 kV trended towards being the best monoenergetic level. When artefacts are minimal, 90 kV trended towards being the optimal monoenergetic level with the least amount of noise present.

Engineering,Medicine