Room: AAPM ePoster Library
Stopping power ratios (SPRs) are needed for proton dose calculation in treatment planning system. In previous study, Saito (Med. Phys. 44, 4179 (2017)) performed SPR calibration on a Gammex RMI 467 phantom using dual-energy CT images. However, due to beam hardening effect, location of material would have influence on HU value. In this work, we followed Saito’s methodology and present SPR comparison of three different rod arrangements.
16 material rods of Gammex RMI 467 were scanned with 90 and 140 kVp on a Philips Big Bore CT. Three rod arrangements following (A) phantom user’s guide, (B) Landry et al (Phys. Med. Biol., 58, 6851 (2013)) and (C) individual rod scan at center of the uniform CIRS phantom (062MA-01 and 062MA-02) were used in this investigation. In-house MATLAB codes following Saito’s methodology were validated using their published data. Calibration curves of relative electron density (RED), effective atomic number (Zeff¬), SPRs vs. ?HU were obtained from the in-house codes for 100 ICRU-46 human tissues after phantom calibration.
For final evaluation, selective outliers were excluded in overall root-mean-square error (RMSE) calculation due to their possible biasing impact on the calibration results. These included: Lung (Adult: healthy) for all evaluations in (A) and (B), lung (Adult: healthy) for Zeff evaluation in (C) and breast calcification and bone mineral-hydroxyapatite for RED and SPR evaluations in (C). Overall relative RMSE differences between calibrated and theoretical values of (A), (B) and (C) after the exclusion are 0.20%, 0.74%, 1.29% on RED, 0.94%, 2.22%, 2.14% on Zeff, 0.98%, 1.02%, 1.38% on SPRs, respectively.
Calibration results of rod arrangement recommended by user’s guide are the closest to theoretical values. Different rod arrangement schemes can lead to different calibration curves. Thus, rod arrangement scheme should be considered when performing CT to material calibration.