Purpose: The aim of this study was to develop a theoretical framework for the investigation of the impact of x-ray spectral shapes and dependent parameters (dose allocation, total dose, energy ratio) on the resulting dual-energy (DE) image quality.
Methods: Spectral shape incident on the object was described with an analytical fitting function, partially depicting the physical mechanisms behind the polychromatic x-ray beam generation such as Bremsstrahlung radiation and its attenuation. This function was written in terms of the physical parameters, such as x-ray tube potential Emax (kVp), parameter α (related to tube mAs), and parameter β (related to photoelectric absorption coefficient). Simple phantom was developed to evaluate image quality and this model was used in theoretical calculations. The contrast to noise ratio (CNR) between a tumor region (behind bone) and its background was used as a figure of merit to quantify image quality. Analytical CNR was derived and parametrically investigated in terms of spectral shape parameters, dose allocation, and total dose. Experimental verification of the derived equation as a function of dose allocation was performed. Various dose allocations for different energy ratios were achieved by varying beam filtrations and mAs with constant total dose.
Results: Equation for DE CNR as a function of spectral parameters, total dose, and dose allocation was derived. It was parametrically investigated and experimentally validated. It was shown that the contrast is defined by material properties only, while noise depends on relationship between spectra. Expression for dose allocation corresponding to maximum tumor CNR function was derived.
Conclusion: The derived analytical CNR function allows for investigation of the spectral parameters and their impact on the image contrast and noise. It was shown that dose allocation corresponding to maximum CNR has strong dependency on material composition of the imaged object and cannot be reduced to one near-optimal number.