JA Griffiths, GJ Royle, AM Hanby, JA Horrocks, SE Bohndiek, RD Speller
Phys Med Biol
Identification of specific tissue types in conventional mammographic examinations is extremely limited. However, the use of x-ray diffraction effects during imaging has the potential to characterize the tissue types present due to the fact that each tissue type produces its own unique diffraction signature. Nevertheless, the analysis and categorization of these diffraction signatures by tissue type can be hampered by the inhomogeneous nature of breast tissue, leading to categorization errors where several types are present. This work aims to reduce sample categorization errors by combining spectral diffraction signature collection with sample imaging, giving more detailed data on the composition of each sample. Diffraction microCT was carried out on 19 unfixed breast tissue samples using an energy resolving translate-rotate CT system. High-resolution transmission microCT images were also recorded for comparison and sample composition analysis. Following imaging, the samples were subjected to histopathological analysis. Reconstructing on various momentum transfer regions allows different tissue types to be identified in the diffraction images. Results show a correlation between measured x-ray diffraction images and stained histopathological tissue sections. X-ray diffraction signatures generated from the measured data were categorized and analysed, with a t-test indicating that they have the potential for use in tissue type identification.