Frequency split metal artefact reduction in pelvic computed tomography
- 644 Downloads
Artefacts from total hip replacement affect image quality and the visualisation of pelvic lesions on computed tomography (CT). We propose a frequency split (FS) approach in addition to the normalised metal artefact reduction (NMAR) algorithm that aims to suppress artefacts and improves image quality in patients with orthopaedic hardware.
Data from ten consecutive patients with metal artefacts from uni- (n = 5) and bilateral (n = 4) total hip replacement or osteosynthesis (n = 1) were reconstructed with filtered back projection (FBP), linear interpolation MAR (LIMAR), NMAR, FSLIMAR and FSNMAR and analysed for image quality and severity of artefacts.
NMAR and FSNMAR significantly improved the assessment of the pelvic organs, lymph nodes and vessels compared with FBP, LIMAR or FSLIMAR (P < 0.05). Assessment of the metal hardware, joint and capsule was improved with the addition of FS (FSLIMAR, FSNMAR). No algorithm-related artefacts were detected in regions that did not contain metal.
NMAR, FSLIMAR and FSNMAR have the potential to improve image quality in patients with artefacts from metal hardware and to improve the diagnostic accuracy of CT of the organs of the pelvis. Although introducing some algorithm-related artefacts, FSNMAR most accurately displayed adjacent bone and tissue next to metal implants.
• Orthopaedic metallic hardware often creates serious artefacts in computed tomography, hindering diagnosis.
• The normalised metal artefact reduction (NMAR) algorithm was developed to suppress such artefacts.
• NMAR improves CT assessment of pelvic organs in patients with orthopaedic hardware.
• Addition of the frequency split technique (FSNMAR) helps assess tissue near metal hardware.
• NMAR and FSNMAR are robust and computationally effective sinogram interpolation algorithms.
KeywordsComputed tomography Metal artefact reduction Total hip replacement Sinogram interpolation Frequency split
The first two authors contributed equally in this study.