Chao-Jen Lai PhD, Abstract Co-Author: Nothing to Disclose

Chris Chorng-Gang Shaw PhD, Abstract Co-Author: Nothing to Disclose

Lingyun Chen, Abstract Co-Author: Nothing to Disclose

Aysegul A. Sahin MD, Abstract Co-Author: Nothing to Disclose

Wei Tse Yang MD, Abstract Co-Author: Nothing to Disclose

Gary Jacob Whitman MD, Presenter: Nothing to Disclose

To investigate the effects of various reconstruction slice thickness settings for displaying cone beam breast CT images.

380 x 380 x 380 volume data with a 300-micron-voxel were extracted from cone-beam CT images of breast specimens. Spherical simulated masses of various sizes (2-8 mm) were numerically created by a computer program and inserted into the volume data at several possible locations containing no known cancers. The contrast level of the simulated mass was determined from the difference in the CT number between the duct carcinoma and the overlapping tissue in the reconstructed images. An attenuation pattern for contrast level was applied to all masses. The simulated masses and controls were alternatively displayed at the same location while display. Thin slice (0.3 mm) images were averaged to form thicker slice images (0.6, 1.2, and 1.8 mm) for display. Stationary and sequential display of the thin slice and the synthesized thick slice images were randomized and viewed by two mammographers for rating the image quality and visibility of the masses. Receiver operating characteristic (ROC) analysis has been performed. The area under the ROC curve (Az) is computed to compare different display modes, and two-tailed t-test is used to compute statistical significance.

In the pilot study, it was found that thin slice (0.3 mm) images depicted superior tissue contrast and resolution. Averaging of 5-10 slices into a thicker slice (1.5-3 mm) for display resulted in less noisy images with similar image contrast. However, excessive averaging (over 20 or more slices) resulted in blurred structures and reduced contrast with the appearance of regular mammograms.

This work was supported in part by a research grant EB-00117 from the National Institute of Biomedical Imaging and Bioengineering and a research grant CA104759 from the National Cancer Institute.

Lai, C, Shaw, C, Chen, L, Sahin, A, Yang, W, Whitman, G, Detection of Computer-simulated Masses in Cone Beam Breast CT Imaging: Performances with Various Slice Thicknesses—A Work-in-Progress.  Radiological Society of North America 2005 Scientific Assembly and Annual Meeting, November 27 - December 2, 2005 ,Chicago IL. http://archive.rsna.org/2005/4418424.html