Herbert K. Bruder PhD, Abstract Co-Author: Employee, Siemens AG

Rainer Raupach PhD, Abstract Co-Author: Employee, Siemens AG, Forchheim, Germany

Karl Stierstorfer PhD, Abstract Co-Author: Employee, Siemens AG

Thomas G. Flohr PhD, Presenter: Employee, Siemens AG, Forchheim, Germany

We present a four-dimensional, non-linear, iterative image filter which significantly reduces image noise in cardiac CT while preserving temporal and spatial resolution.

A non-linear, iterative, four-dimensional image filter operates on a set of image volumes, reconstructed at the desired cardiac phase and at several adjacent cardiac phases. Substantial image noise reduction without smoothing edges is achieved by correlating adjacent image signals in four-dimensional cubes. The iterative range filter that constrains adjacent image signals in four dimensions smoothes low-contrast signals while preserving or even enhancing edges. Due to the four-dimensional iteration also edges in the temporal domain are preserved, hence the temporal resolution of the images is maintained. We demonstrate the efficiency of our approach with images of a moving coronary artery phantom at 60, 70 and 90 bpm, with images of the high resolution insert of the CATPHAN, and with clinical dual source CT (DSCT) data.

Images of the moving coronary artery phantom were visually free of motion artifacts before and after image filtration, while image noise could be reduced by 80%. Spatial resolution in the high resolution phantom remained unaffected or could even be improved with the four-dimensional image filter. Clinical data demonstrated noise reduction by up to 80% at preserved sharpness or, alternatively, significantly increased sharpness and substantial reduction of Ca blooming at similar image noise.

The presented image filter can substantially reduce image noise in ECG-synchronized CT images, while maintaining spatial and temporal resolution.

In ECG-gated cardiac CT examinations, substantial noise reduction without loss of spatial and temporal resolution is possible at reduced patient dose.

Bruder, H, Raupach, R, Stierstorfer, K, Flohr, T, Four Dimensional Iterative Image Filter for Noise Reduction at Maintained Temporal Resolution in Cardiac CT.  Radiological Society of North America 2009 Scientific Assembly and Annual Meeting, November 29 - December 4, 2009 ,Chicago IL. http://archive.rsna.org/2009/8011166.html