Non-Circular Scans and Image Reconstruction for Breast Ct
Junguo Bian,Nathan J. Packard,Kai Yang,Dan Xia,John M. Boone,Xiaochuan Pan
DOI: https://doi.org/10.1117/12.772093
2008-01-01
Abstract:Current dedicated, cone-beam breast CT scanners generally use a circular scanning configuration largely because it is relatively easy to implement mechanically. It is also well-known, however, that a circular scanning configuration produces insufficient cone-beam data for reconstrucing accurate 3D breast images. Approximate algorithms, such as FDK has been widely applied to reconstruct images from circular cone-beam data. In the FDK reconstruction, it is possible to observe artifacts such as intensity decay for locations that are not within the plane containing the circular source trajectory. Such artifacts may potentially lead to false positive and/or false negative diagnosis of breast cancer. Non-circular imaging configurations may provide data sufficient for accurate image reconstruction. In this work, we implement, investigate innovative, non-circular scanning configurations such as helical and saddle configurations for data acquisition on a dedicated, cone-beam breast CT scanner, and develop novel algorithms to reconstruct accurate 3D images from these data. A dedicated, cone-beam breast CT scanner capable of performing non-circular scanning configurations was used in this research. We have investigated different scanning configurations, including helical and saddle configurations. A Defrise disk phantom and a dead mouse were scanned by use of these configurations. For each configuration, cone-beam data were acquired at 501 views over each turn. We have reconstructed images using our BPF algorithm from data acquired with the helical scanning configuration.