Room: Karl Dean Ballroom A2
Purpose: The goal of this study was to implement the stoichiometric calibration method for proton stopping-power-ratio (SPR) estimation based on virtual mono-energetic CT datasets generated by Philips IQon DECT scanner and evaluate the overall uncertainty associated with it.
Methods: 14 tissue substitute inserts from Gammex RMI 467 phantom were scanned with IQon DECT under varying scanning conditions (different phantom sizes and locations within the phantom). The mono-energetic HUs across a range of energy (40-200 keV) were generated for each insert by IQon spectral CT viewer. The stoichiometric method was implemented based on the mono-energetic HUs of a selected energy. The uncertainties in SPR estimation was estimated for five independent uncertainty categories and for lung, soft and bone tissues separately.
Results: The virtual mono-energetic HU values of 100 keV were found most consistent under different scanning conditions for most tissue inserts. Thus, the CT datasets of 100 keV were selected for the stoichiometric calibration and SPR estimation. The overall uncertainty in SPR estimation using the stoichiometric approach based on 100 keV mono-energetic CT datasets was estimated to be 7.4%, 1.2% and 1.4% for lung, soft and bony tissues, respectively, which is comparable to the corresponding values for the conventional DECT approach implemented on state-of-art DECT scanners from other vendors: 3.8%, 1.2% and 2.0%, respectively.
Conclusion: We have identified the optimal energy (100 keV) for implementing the stoichiometric method on an IQon DECT scanner. Our results demonstrated the potential of the IQon DECT scanner for reducing proton SPR uncertainty because of its unique dual-layer detector design: even the stoichiometric approach implemented on IQon DECT can achieve comparable accuracy as the more complex DECT-based approach on other DECT scanners. Our next step will be exploring ways to take advantage of additional effective atomic number information provided by IQon DECT for true DECT-based approach.