:: ukrc 2004

The 2004 UK Radiological Congress (UKRC) was held at Manchester G-MEX from 6-8th June 2004. ImPACT presented a paper at the advances in CT scanning session at this meeting, the abstract for the talk is below.

The relative performance of cone-beam reconstruction algorithms and standard multi-detector algorithms (view slides)

David J. Platten, Nicholas Keat, Maria A. Lewis, Julia F. Barrett, Sue Edyvean

PURPOSE: To compare the performance of cone-beam reconstruction algorithms to that of standard multi-detector CT (MDCT) algorithms.

MATERIALS/METHODS: Several commercially available MDCT systems were used to obtain helical CT images of a simple phantom. The phantom consists of a 10 mm diameter Teflon rod supported at an angle of 60 degrees to the horizontal, immersed in a water bath. It was positioned in each scanner such that the central section of the rod was at the isocentre. Helical scans were performed, with images reconstructed using both cone-beam and standard MDCT algorithms. The scans and image reconstruction were repeated with the phantom vertically offset from the isocentre by 130 mm. The standard deviation of CT number within a region of interest around the rod was calculated for each image in order to characterise the severity of image artefact. The level of artefact was also visually assessed.

RESULTS: Both numerical and visual assessments demonstrate that the images reconstructed using cone-beam reconstruction algorithms were of a higher quality than those reconstructed with the standard MDCT algorithm.

CONCLUSIONS: Cone-beam reconstruction algorithms suppress image artefact around high contrast objects that are at an angle to the scan plane when compared to images reconstructed using standard MDCT algorithms.