Daniel Prell Dipl Phys, Presenter: Nothing to Disclose

Yiannis Kyriakou PhD, Abstract Co-Author: Nothing to Disclose

Willi A. Kalender PhD, Abstract Co-Author: Consultant, Siemens AG Founder, VAMP GmbH Scientific Advisor, VAMP GmbH Shareholder, VAMP GmbH

The aim of this study was to improve metal artifact reduction (MAR) in flat-detector CT (FD-CT) by combining interpolation and physical correction methods to counterbalance their specific drawbacks.

Our approach uses an image-based weighting concept for superimposing two MAR algorithms: A 3D linear interpolation (3D LI) scheme of reprojected metal traces in the rawdata domain and a combination of multi-dimensional adaptive filtering (MAF) with a beam hardening correction (BHC) method. 3D LI interpolates the metal traces along the 2D projection data and as a third dimension along the projection angle. MAF uses an explicit filtering of metal traces only in the rawdata domain. BHC uses an image-based material segmentation followed by a forward projection of each material. Combining the material rawdata thus obtained and back projecting each sum yields correction images which are linearly combined with the in itial image to correct for beam hardening artifacts. The weighting concept uses directional weighting of metal traces to pay special attention to lines of highest attenuation. Performance was evaluated in cadaver measurements and on patient data for a C-arm FD-CT system (Artis Zeego, Siemens Healthcare Sector, Forchheim, Germany). The accuracy of correction was investigated in subtraction and in noise images (σ-images). Different regions of interest (ROIs) were evaluated in terms of CT values and image noise. Results were compared to results achieved by only one of the correction methods.

The combination method achieved an appropriate metal artifact reduction and noise reduction to a normal noise level. It minimized secondary artifacts introduced by the interpolation scheme in greater distance to the implant and also corrected for corrupted CT values in the close vicinity. The imaging of tissue surrounding the implant was highly improved. Our method was always superior to the application of one single method and offered a better detail preservation.

Our approach combines the advantages of fundamentally different algorithms for artifact reduction and improves the quality of images that are affected by metal artifacts.

Metal artifacts may prevent proper diagnosis. The reduction of obscuring artifacts may be of great importance in such cases.

Prell, D, Kyriakou, Y, Kalender, W, Improving Metal Artifact Reduction for Flat Detector Computed Tomography by Combination of Interpolation and Physical Correction Methods.  Radiological Society of North America 2009 Scientific Assembly and Annual Meeting, November 29 - December 4, 2009 ,Chicago IL. http://archive.rsna.org/2009/8015868.html