Quantitative material decomposition method for spectral CT imaging

Spectral CT is a promising new imaging modality that is able to provide spectral information of several pre-selected energy ranges in one image acquisition, which allows quantitative decomposition of multiple materials. In addition to conventional reconstruction, novel decomposition techniques should be developed to realize material separation. Decomposition of multienergy X-ray data into basis materials can be performed in the projection domain, image domain, or during image reconstruction. In this work, a projection domain decomposition method was introduced and accomplished by a simulated phantom study. Its performance is evaluated not only for heavy atoms with their individual K-edge signature like gadolinium and iodine, but also for lighter atoms like iron, calcium and potassium. It is shown that this approach succeeds in the quantification of gadolinium, iodine and iron, moreover, it is also capable to discriminate iron and potassium from water and PMMA.