Author: Afrouz Ataei, Xinhui Duan, Andrew R. Godley, Mi-Ae Park, Mahbubur Rahman, Liqiang Ren, Chenyang Shen, Gary Xu 👨🔬
Affiliation: Department of Radiology, UT Southwestern Medical Center, UT Southwestern Medical Center, Rush University, Imaging Services, UT Southwestern Medical Center 🌍
Purpose:
Electron density (ED) phantoms are crucial for calibrating CT number (HU) to relative ED in treatment planning. This study evaluates the differences in HU-to-ED calibrations between two commercially available phantoms on two diagnostic CT scanners.
Methods:
Two electron density phantoms were evaluated: Phantom A (Model 467, Gammex) and Phantom B (Model 1467, Sun Nuclear), each containing 16 slots. Key differences include shape and size (Phantom A: cylindrical, 33 × 5 cm; Phantom B: elliptical cylinder, 40 × 30 × 16.5 cm), slot configuration (Phantom A: two concentric circles; Phantom B: central, near central, and two circular arrangements), and material composition of rods, including nominal ED and physical density. Both phantoms were scanned on two CT scanners from different vendors using 120 kV and clinical imaging protocols. Calibration curves (HU vs. ED) were generated, and piecewise linear fits were applied separately for lung-, soft tissue-, and bone-equivalent materials. Slope and intercept values were compared across scanners and phantoms.
Results:
For Phantom A, slope differences between two scanners were 2.3%/5.0%/2.1%, and intercept differences were 2.3%/1.0%/0.4% for lung-/soft tissue-/bone-equivalent materials. For Phantom B, slope differences were 1.7%/2.6%/0.0%, and intercept differences were 1.9%/0.1%/0.1%. Comparing the two phantoms, slope differences were 8.3%, 2.6%, and 3.2% for lung-, soft tissue-, and bone-equivalent materials, respectively, while intercept differences ranged from 0.7% to 1.9% across all material groups.
Conclusion:
Minimal differences were observed between scanners for the same phantom. However, differences between the two phantoms were more pronounced for low-density (lung-equivalent) materials compared to soft tissue and high-density (bone-equivalent) materials. To ensure calibration consistency, the same phantom should be used for cross-calibration of CT scanners, particularly for treatment planning. Further validation is needed to assess the impact of calibration curve variations on dose accuracy.