Marcel Van Straten, Presenter: Nothing to Disclose

Henk Willem Venema PhD, Abstract Co-Author: Nothing to Disclose

Joris Hartman, Abstract Co-Author: Nothing to Disclose

Gerard J. Den Heeten MD, Abstract Co-Author: Nothing to Disclose

Cornelis A Grimbergen PhD, Abstract Co-Author: Nothing to Disclose

To demonstrate the relationship between the x-ray tube rotation angle and the spiral reconstruction interpolation artifacts, and the effect of these artifacts on the reproducibility of spiral CT scans. Reproducible scans are required when image processing techniques such as subtraction or masking are used.

With the aid of a skull in a synthetic material and a small high density sphere the influence of pitch, gantry tilt and interpolation algorithm on the reconstruction artifacts is investigated. Scans were made with a spiral CT scanner (Mx8000 Quad, Philips, Best, The Netherlands) with 4x1 mm collimation. Multiple scans were made of the skull with all adjustable scan parameters kept constant. Series were made for different values of pitch and gantry tilt. Differences between scans within one series were determined by calculating the mean value of the squared differences in the subtracted images.

In the scans of the dense sphere interpolation artifacts showed up as hyper- and hypodense streaks from the center of the sphere (windmill artifacts). The streaks varied in number and intensity for varying pitch. The artifacts were more severe for higher pitch values. The orientation of the streaks was fully determined by the starting angle of the x-ray tube. This angle did not have a constant value but varied seemingly at random between consecutive scans. Consequently, in the scans of the skull the interpolation artifacts showed up differently in otherwise identical images, and showed up as substantial residual artifacts when subtractions were made. The magnitude of these artifacts could be reasonably predicted from the magnitude and phase of the windmill artifacts. Scanning with a gantry tilt resulted in a considerable decrease of the reproducibility. The influence of the interpolation algorithm (180° or 360°)was relatively small.

Multi-slice spiral CT images are in general not reproducible. Only if the x-ray tube starting angle is equal for both scans or if a low pitch value (0.375) is used, images are reproducible.

M.V.: M. Van Straten is financially supported by Philips Medical Systems corporation.

Van Straten, M, Venema, H, Hartman, J, Den Heeten, G, Grimbergen, C, Reproducibility of Multi-slice Spiral CT Scans.  Radiological Society of North America 2004 Scientific Assembly and Annual Meeting, November 28 - December 3, 2004 ,Chicago IL. http://archive.rsna.org/2004/4407586.html