Purpose: In radiotherapy, computed tomography (CT) imaging is the gold standard method to obtain patient-specific data for treatment planning. To account for tissue heterogeneities, dose calculation engines require CT calibration curves which links, in each voxel, the HU to other physical properties such as the electron density (ED) or a stopping power (SP). So called ‘Schneider’s method’ is the most widely used calibration method in hadron therapy worldwide. However, uncertainty on HU measurements can cause tissue misallocation and yield errors on absorbed dose calculation of order of 10 % or more. Moreover, since the HU-ED( or HU-SP) relations requires performing multi-linear fits, subjectivity arises when it comes to defining the number of segments and their global trend. To overcome this current limitation, we suggest to use pseudo Triple-Energy CT (pseudo TECT), three CT images scanned with three different kVps, for more accurate stopping power estimation.
Methods: Commercial dual energy CT scanner was used for having pseudo TECT images and Gammex 476 phantom was scanned. CT scanning with 80, 100, 120 and 140 kVp were done and all possible energy groups were tested. The core idea was inspired by stoichiometric calibration using dual energy CT-based stopping power estimation methods, then expanded and applied to pseudo TECT.
Results: The residuals(in %) for the estimated electron density, atomic number, mean excitation energy, and stopping power were -0.31, 0.09, 1.11, and -0.43 for DECT based estimation, whereas -0.41, 0.00, 0.10, and -0.42 for pseudo-TECT based. That is, the results were similar or worse for electron density, but extremely improved for atomic number, mean excitation energy and stopping power eventually.
Conclusion: Out suggested pseudo-TECT based method for hadron therapy has shown significant improvement for the accuracy of stopping power estimation compared to the DECT based method.