Investigation of optimal combination of monochromatic image of dual-energy CT system for proton range calculation

Abstract

PurposeTo explore the optimal combination of monochromatic images derived from dual-energy CT (DECT) in calculating the stopping power ratio (SPR) of human tissues. MethodsMonochromatic CT numbers ranging from 40 keV to 140 keV (in intervals of 10 keV) for 34 standard human tissues were theoretically computed based on the NIST database. These values were then paired, resulting in 55 different dual-energy combinations for calculating the stopping power ratio (SPR). The combinations adhered to the rule of a lower energy (Elow) paired with a higher X-ray energy (Ehigh). Subsequently, each energy combination was utilized to predict the SPR of human tissues, and the accuracy of each pairing was assessed. Additionally, an analysis of the impact of uncertainty in the attenuation coefficient was conducted. ResultsWhen Elow is ≤ 70 keV, an optimal energy combination emerges. These optimal energy combinations within the 40 keV, 50 keV, 60 keV, and 70 keV groups are 40–100 keV, 50–90 keV, 60–80 keV, and 70–80 keV, respectively, with corresponding mean absolute errors (MAE) of SPR at 0.12%, 0.11%, 0.12%, and 0.15%, respectively. The impact of attenuation coefficient uncertainty on these findings is minimal; even with a 1–2% uncertainty, the maximum change in MAE for SPR error is only 0.04%. ConclusionsDual-energy CT utilizing energy pairs within the low energy range (40–70 keV) exhibits greater advantages in predicting the stopping power ratio of human tissues. In our investigation, the four energy combinations—40-100 keV, 50–90 keV, 60–80 keV, and 70-80 keV—demonstrated the highest accuracy in predicting SPR.