Enhanced imaging of metallic knee prostheses by using a representative phantom

Objective: Investigate how a representative phantom can contribute to enhanced imaging of a Total Knee Arthroplasty (TKA) prosthesis in low-field MRI. Method: Ingredients of the phantom are gelatine (gelling agent), Dotarem® (T1 modifier), agarose (T2 modifier) and demineralised water. The T1 and T2 relaxation times of tissue mimicking samples are measured using a spin echo sequence in a 0.25 Tesla low field MRI. Artefacts sizes in the MR images of the phantom materials with prosthesis are measured and compared with the artefact sizes in the MR images of a human knee with prosthesis. Results: Relaxation times of the samples ranged from 210 to 480 for T1. It was not possible to calculate T2 values in the samples. The artefact sizes differ from artefacts in a real knee with prosthesis. Relaxation times of the human knee ranged between 160 and 470 for T1 and between 25 and 127 for T2. The phantom is not representative for a human knee. Comparison between phantom with prosthesis and knee with prosthesis show conflicting results. Especially the FSE X-MAR scan of the phantom shows more artefacts than the standard FSE scan. Conclusion: It was not possible to create a representative phantom. It is seen that Dotarem® and agarose change T1 relaxation times. The metal artefacts do not show consistent results. A representative phantom with complex structures should determine in what way metal artefacts can be reduced, to better image TKA problems.