Dual-Energy Computed Tomography in Loosening of Revision Hip Prosthesis: A Comparison Between MARS and non-MARS Images.

The aim of the study was to assess potential endoprosthesis loosening in patients after revision total hip arthroplasty, based on monochromatic dual-energy computed tomography (DECT) images obtained with and without metal artifact reduction software (MARS) and comparison with the clinical examination and hip function. Twenty-five consecutive patients underwent DECT examinations. Two monochromatic data sets were generated: with MARS (75-keV MARS) and without MARS (140-keV non-MARS) and evaluated for signs of loosening, using a dedicated radiological score. The Harris Hip Score (HHS) was used to evaluate patients' hip function. The mean radiological score for loosening in the non-MARS group was 16.2, in the MARS group 17.0 and was significantly higher (P < 0.001). Radiological loosening evaluation in non-MARS images correlated with the HHS score (ρ = 0.43, P = 0.03), whereas there was no correlation between MARS images and HHS (ρ = 0.15, P = 0.47). New MARS-related artifacts, which hinder implant-bone interface evaluation, were found in 75-keV MARS images ("pseudo-loosening" in 33 of 34 prostheses, "ground-glass blurring" in 20 of 32). High-energy monochromatic DECT images are superior to low-energy DECT MARS images in assessment of the potential loosening and correlate better with the clinical examination outcomes. For revision total hip arthroplasty evaluation, 2 data sets should be assessed inseparably because of extra artifacts in MARS images.