A guide to diagnose failed total knee arthroplasty: a case study comparing low-field and high-field MRI

C. Post, F.F. Schröder, F.F.J. Simonis, A. Peters, R. Huis in 't Veld, N. Verdonschot

Research output: Contribution to journalMeeting AbstractAcademic


Introduction: Fifteen percent of the primary total knee arthroplasties (TKA) fails within 20 years. Among the main causes for revision surgery are instability and patellofemoral pain. Currently, the diagnostic pathway requires various diagnostic techniques to reveal the original cause for the failed knee prosthesis and is therefore time consuming and inefficient.

Accordingly, there is a growing demand for a diagnostic tool that is able to simultaneously visualize soft tissue structures, bone and TKA. Magnetic resonance imaging (MRI) is capable of visualising all the structures in the knee although a trade- off needs to be made between metal artefact reducing capacities and image quality. Low-field MRI (0.25T) results in less metal artefacts and a lower image quality compared with high-field MRI (1.5T). The aim of this study is to develop a MRI imaging guide to image the problematic TKA and to evaluate this guide by comparing low-field and high-field MRI on a case study.

Method: Based on literature and current differential diagnostic pathways a guide to diagnose patellofemoral pain, instability, malposition and signs of infection or fracture with MRI was developed. Therefore, methods as Insall Salvati, patellar tilt angle and visibility of fluid and soft tissues were chosen. Visibility was scored on a VAS scale from 0 to 100mm (0mm zero visibility, 100mm excellent visibility).

Subsequently, this guide is used to analyse MRI scans made of a volunteer (female, 61 years, right knee) with primary TKA (Biomet, Zimmer) in sagittal, coronal and transversal direction with a FSE PD metal artefact reducing (MAR) sequence (TE/TR 12/1030ms, slice thickness 4.0mm, FOV 260×260×120mm3, matrix size 224×216) on low-field MRI (Esaote G-scan Brio, 0.25T) and with a FSE T1-weighted high bandwidth MAR sequence (TE/TR 6/500ms, slice thickness 3.0mm, FOV 195×195×100mm3, matrix size 320×224) on high-field MRI (Avanto 1.5T, Siemens).

Scans were analysed three times by one observer and the intra observer reliability was calculated with a two-way random effects model intra class correlation coefficient (ICC).

Results: Due to less metal artefacts on the low-field MRI scans the angle, distance and ratio measurements were more consistent: Insall Salvati low-field 0.97–0.99, Insall Salvati high-field 1.05–1.12, patellar tilt angle low-field 2.1–2.8°, patellar tilt angle high-field 2.4–7.6°.

Over all, the VAS scores are higher on the high-field MRI scans; VAS medial collateral ligament high-field 26–45, VAS medial collateral ligament low-field 24–34, VAS popliteus tendon high-field 15–27, VAS popliteus tendon low-field 2–7.

The ICC values of the VAS scores, angle measurements and ratio measurements were excellent, ICC > 0.9. The ICC values of the distance measurements were moderate, ICC > 0.6.

Conclusion: MRI offers possibilities to simultaneously differentiate underlying causes of the failed knee prosthesis. The structures of interest were more clearly visible on the high-field MRI scans due to higher image contrast. The angle, distance and ratio measurements were more consistent on the low-field MRI scans due to less metal artefacts. Further research should focus on a larger group of patients with complaints after TKA to verify the analysis methods.
Original languageEnglish
Pages (from-to)135-135
JournalOrthopaedic proceedings
Issue numberSuppl. 5
Publication statusPublished - 2019
EventISTA 31st Annual Congress 2018 - Queen Elizabeth II Centre, London, United Kingdom
Duration: 10 Oct 201813 Oct 2018
Conference number: 31


  • NLA


Dive into the research topics of 'A guide to diagnose failed total knee arthroplasty: a case study comparing low-field and high-field MRI'. Together they form a unique fingerprint.

Cite this