Zhihui Sun MS, Abstract Co-Author: Research grant, Siemens AG

William Paul Segars PhD, Abstract Co-Author: Nothing to Disclose

Elliot K. Fishman MD, Abstract Co-Author: Research support, Siemens AG Advisory Board, Siemens AG Research support, General Electric Company Advisory Board, General Electric Company Co-founder, HipGraphics, Inc

Jeffrey A. Brinker MD, Abstract Co-Author: Nothing to Disclose

Katsuyuki Taguchi PhD, Presenter: Research grant, Siemens AG Research grant, Koninklijke Philips Electronics NV

One of the problems with the current cardiac CT imaging is banding artifacts, which are caused by non-periodic heart motion. The purpose of our study was to develop a method to reduce banding artifacts.

The heart is scanned over several heart beats, where data from each heart beat cover a limited height of the heart. In reconstruction, those heart beats are “connected” to image the entire heart. All of the current methods (e.g., Taguchi, et al., Med Phys 33: 512-39) (1) choose a cardiac phase with respect to electro-cardiogram (ECG) signals and (2) use the fixed cardiac phase for the entire heart. To reduce banding artifacts, our proposed method (1) chooses cardiac phases to maximize the consistency of the anatomical structure between adjacent heart beats and (2) allows the phases to smoothly vary in space. We evaluated the proposed method using patient data collected from a CT scanner (Siemens, Sensation 64). Images were reconstructed using both the proposed method and current methods. A mean absolute difference (MAD) between each pair of adjacent slices was calculated throughout the volume and an abrupt change in MAD along the longitudinal direction was used as a banding artifact index.

Image reconstructed by the proposed method showed significantly reduced banding artifacts. The anatomical structure in the images was much more consistent and physiologically reasonable than that in images reconstructed by the current methods. Particularly, in some cases with strong non-periodic heart motions, the banding artifact index was reduced by ~60% (to ~40% of the current level). The remaining 40% of the banding artifact index was mainly caused by the uncorrelated image noise.

The proposed method significantly reduces the banding artifacts in 3D cardiac CT images, especially when patients' heart motions are strongly non-periodic.

The new cardiac CT reconstruction algorithm reduces banding artifacts and improves the quality of cardiac CT images. This may also allow us to relax patient exclusion criterion for cardiac CT exam.

Sun, Z, Segars, W, Fishman, E, Brinker, J, Taguchi, K, Banding Artifact Reduction for Cardiac CT.  Radiological Society of North America 2008 Scientific Assembly and Annual Meeting, February 18 - February 20, 2008 ,Chicago IL. http://archive.rsna.org/2008/6011040.html