[19a-A304-1~11] 2.3 Application, radiation generators, new technology

△ [19a-A304-4] Effective atomic number measurement with energy-resolved computed tomography using a flat panel detector

〇(D)Hiraku Iramina^1,2, Mitsuhiro Nakamura^2,3, Takashi Mizowaki², Ikuo Kanno¹ (1.Engineering, Kyoto Univ., 2.Medicine, Kyoto Univ., 3.Human Health, Kyoto Univ.)

Keywords:Energy-resolved computed tomography, Effective atomic number

Introduction: We have previously developed an effective atomic number (Z_eff) measurement method using linear attenuation coefficients (LACs) obtained by energy-resolved computed tomography (CT) with one-dimensional (1D) detector. The energy-resolved CT was performed with a “transXend” detector, which measured X-rays as electric current and then gave X-ray energy distribution with unfolding analysis using pre-estimated response function (RF). The purpose of this study is to measure Z_eff by the energy-resolved CT using a flat panel detector (FPD). Methods: To demonstrate a 2D transXend detector, we developed the stripe absorbers for the FPD. Eleven human tissue-equivalent material rods which were grouped into four material categories were measured by X-rays with 120 kVp tube voltage, 2.3 mA tube current, and 1.0 s exposure time. Z_eff is measured by the ratio of LACs with two different pseudo-monochromatic X-ray energies. RFs of each rod material were estimated by numerical calculation. First, we employed the RF estimated for the same rod material (self-RF scenario). Second, we employed the RF estimated for the different rod materials in the same material category (cross-RF scenario). The purpose of the cross-RF scenario was to find representative rod materials in each material category. Results: Upon the self-RF scenario, measured Z_effs were systematically underestimated. Median relative error to theoretical Z_eff was -6.92% (range: -7.89% - -4.60%). After normalizing measured Z_effs to the theoretical one for Breast, median relative error improved to -0.75% (range: -1.79 - +1.73%). Upon the cross-RF scenario, the representative rod materials were found in two material categories. Conclusion: Z_eff measurements were performed by energy-resolved CT using 2D transXend detector with numerically-estimated RF data. Normalized Z_effs for all rod materials in the self-RF scenario were in good agreement with the theoretical ones.