Case report Artefact or hip prosthetic fracture on multislice CT? The importance of correct positioning when scanning metal implants M.W. Kusk Hospital of Southwest Jutland, Department of Radiology and Nuclear Medicine, Finsensgade 35, 6700 Esbjerg, Denmark article info Article history: Received 27 July 2018 Received in revised form 27 November 2018 Accepted 3 December 2018 Available online 14 December 2018 Introduction Fractures of total hip arthroplasty (THA) implants, are rare but not unknown. 1 Correct diagnosis is crucial in deciding whether implant replacement surgery is indicated. Loosening of femoral stem or acetabular components are far more common, 2 but cannot be differentiated without imaging. Conventional radiography of the hip is first-line modality, but when the diagnosis is unclear or incon- sistent with clinical presentation, CT is the modality of choice. 3,4 However metal implants impart artefacts on CT images. Beam hardening artefacts occur where x-rays are filtered by passage through radiodense material, increasing average energy of the x-ray spectrum. 5 The result is a dark streak- or star-like artefact radiating from the metal into adjacent tissue, and potentially obscuring pathology. Artefact reduction can be achieved by increasing kVp, metal artefact reduction (MAR) reconstruction al- gorithms, and reducing thickness of metal in the scanplane, through careful positioning of patient and/or tilting of the scanner gantry. 6 When these precautions are not observed, there is a risk of the artefacts simulating pathology. Case report A 67 year old man presented at the orthopedic surgery depart- ment 6 months after right THA, with increasing pain when walking, and no history of hip trauma. Paraclinical findings did not indicate infection. Standard hip radiographs were performed, showing normal postoperative status. The patient was then referred to CT of the hip prosthesis, to rule out component loosening. Scanning was on a Toshiba Aquilion PRIME 80-slice CT-scanner (Toshiba Medical Systems, The Netherlands) with a dedicated protocol as follows: mA: 300, kVp 135, Pitch: 0,684, rotation time: 1 sec. Images were reconstructed in 0,5 mm thickness and 0,3 mm interval at a medium sharp kernel, for MPR processing. Images were reconstructed with and without Single Energy Metal Artefact Reduction (SEMAR). Images were then reviewed by a musculoskeletal (MSK) radi- ologist with more than 15 years of experience. Small cysts were found in the acetabulum/liner interface. At the neck of the femoral component, a dark streak raised suspicion of fractured prosthesis (Fig. 1), on both SEMAR and non-SEMAR images, especially on paracoronal MPR. Volume Rendering Technique (VRT) re- constructions were used for a global view. The irregular appearance and localization could indicate serrated edges of a material fracture. The case was discussed with another MSK radiologist. A senior CT radiographer was consulted to see if artefacts could be resolved by reconstructions at different algorithms, which was not the case. Therefore the patient was recalled for renewed CT scan of the re- gion, with identical parameters. This time the patient was posi- tioned with the scan plane as perpendicular to the femoral neck as possible, reducing the amount of metal traversed by the beam. The patient was positioned as shown in Fig. 2, as THA patients are advised not to cross their legs (to minimize subluxation risk). A board was placed under the shoulder region, and feet immobilized in a vacuum bag, taped to the patient table but extending beyond it. The dark streak could not be reproduced, and it was concluded that no fracture was present (Fig. 3). Discussion This case highlights the importance of correct patient posi- tioning in CT, when scanning metal implants, even though advanced MAR techniques are used. MAR algorithms are designed to reduce artefacts in tissue surrounding the metal, which may help to explain the failure to reduce the artefact within the metal. SEMAR works by identifying projections in raw data space containing metal artefacts, subtracting these from the sinogram and interpolating missing projections from remaining data. Following this, new images are reconstructed from the resulting sinogram, with the metal data reinserted. Other vendors offer similar solutions: MAR (GE), O-MAR (Philips), I-MAR (Siemens). 7 E-mail address: martin.weber.kusk@rsyd.dk. Contents lists available at ScienceDirect Radiography journal homepage: www.elsevier.com/locate/radi https://doi.org/10.1016/j.radi.2018.12.001 1078-8174/© 2018 The College of Radiographers. Published by Elsevier Ltd. All rights reserved. Radiography 25 (2019) e75ee77