Xinhui Duan PhD, Presenter: Nothing to Disclose

Shuai Leng PhD, Abstract Co-Author: Nothing to Disclose

Jia Wang PhD, Abstract Co-Author: Nothing to Disclose

Cynthia H. McCollough PhD, Abstract Co-Author: Research grant, Siemens AG

To determine the magnitude of errors introduced into quantitative material decomposition analyses by inaccurate knowledge of the x-ray spectra using in dual-energy CT.                  

Material decomposition was investigated for polychromatic photon energies in both image and projection space. Analytical analysis was performed to develop a closed-form expression of error propagation for image space decomposition. Simulations were performed for both spaces. X-ray spectra corresponding to 80 and 140 kV (with tin filtration at 140 kV) were evaluated. Inaccurate spectra were emulated by 1) altering the spectra shape but maintaining the same mean energy 2) altering both the spectra shape and mean energy. In projection space, material decomposition was performed by directly resolving the nonlinear equations using a trust-region method (Matlab, optimization toolbox). Two pairs of materials were included in the simulations: 1) calcium hydroxyapatite (HA) and water, 2) HA and iron oxide (Fe) for a range of material concentrations (i.e. mass fractions). The decomposition error for HA was calculated.            

Results for image and projection space showed that the relative error in calculated HA mass fraction was negatively correlated with the HA mass fraction and the difference in effective atomic number between the paired materials. Simulation results showed that the absolute magnitude of the relative error in HA mass fraction was dramatically increased as its mass fraction decreased. For the same amount of spectrum change of 80 kVp for HA mass fraction of 50%, while a relative error was 11% in the HA/water decomposition, the relative error increased to 50% in the HA/Fe simulation. For the same amount of spectrum change, decomposition errors in projection were smaller than image space.            

Errors in material decomposition analysis caused by inaccurate knowledge of the spectra in dual-energy CT mainly depended on the mass fraction of the basis materials and their effective atomic number difference.            

For clinical applications using quantitative material decomposition and dual-energy CT, accurate knowledge of the spectra is required when quantifying small mass fractions of materials.

Duan, X, Leng, S, Wang, J, McCollough, C, Errors Introduced by Inaccurate Knowledge of the X-ray Spectra in Quantitative Material Decomposition Using Dual-Energy CT.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11034421.html