Visual grading evaluation of commercially available metal artefact reduction techniques in hip prosthesis computed...

Area of Science:

• Medical Imaging
• Radiology
• Computed Tomography

Background:

• Hip prostheses in CT imaging can cause significant metal artefacts, degrading image quality and potentially hindering diagnosis.
• Various metal artefact reduction (MAR) techniques are available, but their efficacy across different vendors and protocols requires thorough evaluation.

Purpose of the Study:

• To assess the effectiveness of MAR algorithms from four CT vendors in reducing artefacts caused by hip prostheses.
• To compare the performance of single-energy CT (SECT) and dual-energy CT (DECT) based MAR techniques, including monoenergetic reconstructions.

Main Methods:

• Phantom images of hip prostheses were acquired using MAR algorithms on SECT and DECT data from four vendors.
• Five radiologists performed a visual grading of image quality using 10 criteria, comparing MAR images against a reference.
• Ordinal probit regression analysis was employed to statistically evaluate image quality differences.

Main Results:

• MAR algorithms generally improved image quality across most criteria (8-10/10) with statistical significance (p < 0.001).
• Some MAR techniques introduced new artefacts or degraded image quality in specific instances.
• Monoenergetic reconstructions showed limited improvement for some criteria on one DECT scanner but worsened image quality on another.

Conclusions:

• MAR algorithms offer a significant improvement in CT image quality for hip prostheses.
• Clinical application requires careful validation of each MAR technique due to potential for new artefacts and variable performance.
• Monoenergetic reconstructions are generally not sufficient for effective metal artefact reduction in this context.