Optimizing CT Slice Width Measurements
M Marteinsdóttir1,2,3*, H Olafsdottir2,3, J Fredriksson2,3, A Andresson3, A Healy2,4, A Kristbjornsson2,3, S Kristinsson2,3, J Levy2,4, D Goodenough 2,4,5, (1) Stockholm University, Stockholm, Sweden,(2) Image Owl, Salem, NY, (3) Raforninn, Reykjavik, Iceland (4) The Phantom Laboratory, Salem, NY (5) George Washington University, Washington DC.SU-E-I-38 Sunday 3:00:00 PM - 6:00:00 PM Room: Exhibit Hall
Purpose: For almost four decades bead and wire ramps have been used in the evaluation of slice width as part of QA testing for CT scanners. While each of these approaches have been recognized and accepted as reliable, in this paper we investigate the differences, advantages and limitations of these tools. Moreover, we study the effect of varying the field of view (FOV) and focal spot size.
Methods: The Catphan® 700 phantom includes two pairs of bead ramps (coarse and fine) and a pair of wire ramps in the same module providing an ideal setting for comparing bead ramps and wire ramps. The phantom was scanned using three devices from two different manufacturers. The data set consisted of 428 slices of 0.5,1,2,4,8 and 10 mm thickness. For the study of FOV and focal spot, 512 slices from the Catphan® 600 were acquired. All images were analyzed using Image Owl Catphan® QA software.
Results: For 0.5mm slices, bead ramps gave more accurate and precise (lower variance) estimation of the thickness than wire ramps. For 2-4 mm slices, the two approaches performed on equal terms while for the thickest slices (8 and 10mm), the wires gave more precise results. For thin slices, a small FOV (100mm) gave better results and lower spread than a large FOV (240mm). Finally, a small focal spot gave significantly better results than a large one using wire ramps for 0.5 and 1mm slices.
Conclusions: For measuring thin slices, the use of bead ramps, with adequately small FOV and a small focal spot should be advised. For measuring thick slices, wire ramps will give less variability although bead ramps give equally accurate results on average.