Quantitative evaluation of metal artefact reduction using iterative reconstruction on a clinical Computed Tomography (CT) scanner

Congress:

ECR 2012

Poster Number:

C-0439

Type:

Scientific Exhibit

Keywords:

Radiation physics, CT, Physics, Technology assessment, Artifacts

Authors:

I. Delakis, E. Kulama, D. McRobbie; London/UK

DOI:

10.1594/ecr2012/C-0439

DOI-Link:

https://dx.doi.org/10.1594/ecr2012/C-0439

Fig. 1: Example of a metal artefact in a CT image of a spine with metallic implant [2]

Fig. 2

Fig. 3

Fig. 4: Filtered back projection reconstruction for 1 rod in the phantom

Purpose

Metallic objects can introduce artefacts in CT imaging, mainly because of photon starvation and beam hardening [1]. Metallic objects have a very high attenuation coefficient and the number of photons passing through them is much lower than the number of photons passing through neighbouring tissue. Due to this photon starvation, noise is high in the projection data and is further magnified when the data are used for reconstruction, resulting in streak artefacts. At the same time, the beam hardening effect from the metallic object makes...

Methods and Materials

We used a CTDI Perspex phantom of 15 cm depth and 32 cm diameter, which contained holes of 1cm diameter. The phantom was imaged with a metallic (aluminium) insert rod positioned in the central hole, and then with an additional insert rod positioned in a peripheral hole. The phantom was also imaged with Perspex rods inserts in the same holes in order to enable comparison of image noise characteristics between different reconstructions. Data were acquired on a 128-channel CT scanner (Definition AS+, Siemens Healthcare, Erlangen,...

Results

Fig. 4 and Fig. 5demonstrate the metal artefact for one and two rods for filtered back projection reconstruction, respectively. The metal artefact was measured on all images using the MSD, NAD and BGP metrics.Fig. 6 shows the metal artefact as a function of iterative reconstruction weight, for images of the phantom with one metallic rod insert. All metrics have been normalised to their value for the image reconstructed with filtered back projection only.Fig. 7 shows the metal artefact as a function of iterative reconstruction weight...

Conclusion

Our work demonstrated quantitatively that iterative reconstruction can be an effective way of reducing CT metal artefacts. The improvement was more significant when higher iterative reconstruction weights were used. Therefore, iterative reconstruction can further improve image quality and benefit diagnosis.

References

[1] Jeong KY and Ra JB. Reduction of artifacts due to multiple metallic objects in computed tomography. Proc. SPIE 7258, 72583E. 2009.Lake Buena Vista,FL,USA [2] Barrett JF and Keat N. Artifacts in CT: Recognition and Avoidance. RadioGraphics. 2004; 24:1679-1691. [3] Kalender WA, Hebele R and Ebersberger J. Reduction of CT artifact caused by metallic implants. Radiology. 1987;164: 576-577 [4] Zhao S, Robertson DD, Wang G, Whiting B, andBae KT. X-ray CTmetal artifact reduction using wavelets: An application for imaging total hip prostheses. IEEE Trans. Med....