Purpose: This study investigated improvement in material decomposition noise and iodine CNR by optimizing the tube voltage and filtration for photon-counting CT with a silicon detector. Silicon photon-counting detectors, which have demonstrated feasibility for human scanning, have unique spectral properties that may benefit from new tube potential and filtration approaches.
Methods: A simulation study was performed across a range of conditions to identify promising filter and tube voltage combinations. The simulations modeled x-ray transmission through a 20, 30, or 40 cm water background and 1-cm iodine task with concentration varied from 1 to 20 mg/ml. Tube voltages of 80, 100, and 120 kV were modeled. Copper, tin, cerium, dysprosium, ytterbium, and tungsten filters of varying thicknesses were modeled in addition to inherent tube filtration. All spectra were modeled with skin exposure matched to the 120 kV beam without additional filtration. A silicon detector with eight energy bins was modeled assuming realistic detection efficiency and Compton scatter effects. The energy-bin thresholds were previously optimized for each tube voltage assuming no additional filtration. The Cramer-Rao-Lower-Bound material decomposition noise variance was calculated for each case and used to estimate the CNR of the resulting optimal virtual monoenergetic images. Promising filter and tube voltage combinations, considering realistic tube power constraints, were further investigated through photon-counting CT simulations.
Results: Filters of 0.1-mm copper, 0.065-mm ytterbium, and 0.02-mm tungsten were identified as promising candidates. The study also demonstrated the improved noise and CNR when using 80 or 100 kV compared to 120 kV. For example, 100 kV spectra with 0.1-mm copper filtration demonstrated 12% CNR improvement and ~8% basis material noise reduction for the 20-cm phantom, and 22% CNR improvement for the 40-cm phantom.
Conclusion: Photon-counting CT with silicon detectors may benefit from lower tube-voltage selection and additional filtration, for example 0.1-mm copper.
Funding Support, Disclosures, and Conflict of Interest: Research supported by GE Healthcare. F.G. and J.F are employees of GE Healthcare.