Raoul M. S. Joemai BSC, Presenter: Nothing to Disclose

Paul De Bruin PhD, Abstract Co-Author: Nothing to Disclose

Wouter J.H. Veldkamp PhD, Abstract Co-Author: Nothing to Disclose

Jacob Geleijns PhD, Abstract Co-Author: Nothing to Disclose

To improve quality of CT images for patients with metal hip or shoulder implants by developing a metal artifact reduction algorithm that operates on original sinograms (raw transmission profiles) and that can be implemented in routine clinical practice.

After the CT acquisition, a standard reconstruction is performed. In the reconstructed images the metal object is segmented by a series of thresholding and morphological operations. The segmented images are submitted to a Forward Projection (FP) algorithm that performs an accurate, virtual CT scan of the imaged volume. The FP thus creates a mask (secondary sinogram) that identifies pixels that are associated to the projection of the metal implant in the original sinograms. Next, interpolation for the metal artifact is performed in the original sinogram using an algorithm based on Laplace’s equation. Pixel values within the mask before interpolation are decreased to 10% of the original value and are added to the pixel values within the mask after interpolation. Thus the metal object remains recognizable after application of the metal artifact reduction process. The modified sinograms are transferred to the CT scanner and artifact reduced images are reconstructed.

Metal artifact reduction applied to five CT scans of patients with hip or shoulder prostheses showed a decrease of noise and artifacts. Artifacts that appeared as bright and dark streaks were reduced or eliminated and image quality improved. Improvements in image quality were better for patients with one prosthesis compared to patients with two prostheses.

Modifying the original sinograms is an effective method to reduce metal artifacts. Regions around shoulder or hip prosthesis can be better evaluated. The automated algorithm can be implemented in clinical practice and can be executed automatically, immediately after the CT acquisition. A study to rigorously evaluate this method is currently being setup.

Metal artifacts can strongly detoriorate images in CT. This software can improve CT image quality for patients with metal hip or shoulder prostheses.

Joemai, R, De Bruin, P, Veldkamp, W, Geleijns, J, Metal Artifact Suppression in Original Raw Data Using an Accurate Forward Projector.  Radiological Society of North America 2009 Scientific Assembly and Annual Meeting, November 29 - December 4, 2009 ,Chicago IL. http://archive.rsna.org/2009/8014497.html