Dual-energy computed tomography - How accurate is gemstone spectrum imaging metal artefact reduction: its application to orthopedic metal implants

Objective: To assess the accuracy and suitability of dual-energy computed tomography (DECT) in scanningmetals used in orthopedic implants. Materials and Methods: Four metal phantoms (Cobalt Chrome, Titanium Grade 5, Stainless Steel 316, and Stainless Steel 630), commonly usedmaterials in orthopedic implants, were scanned by conventional, polychromatic CT as well as Gemstone Spectrum Imaging (GSI) DECT, with and without metal artefact reduction software (MARS). Scans were assessed for artefact based on Hounsfield unit values; and surfaces generated, based on a Canny edge detection algorithm. Two separate metal implants were also scanned and assessed for dimensional accuracy. Results: Conventional, polychromatic CT, and GSI DECT (without MARS) scans displayed major beam hardening in the presence of all four metals. The GSI DECTwith MARS showed very clear and reproducible boundaries with minimal noise surrounding the metal phantoms. However, geometric analysis found overestimation of the dimensions, volume, and surface area for most of the metal phantoms. Titanium displayed the least artefact, compared to the other metals, in all scan scenarios. Conclusions: Although metal artefact reduction using GSI DECT looks superior to conventional CT, when measured objectively, it was shown to overestimate geometries and skew dimensions. The GSI DECT with MARS should be used with caution, especially when assessing questions of implant shape or wear.