Design Optimization for Enhanced Energy Separation In a Kv Dual-Layer Imaging
Abstract
Purpose
Dual-layer imagers (DLIs) enable simultaneous acquisition of low- and high-energy images in a single exposure, offering detector-based spectral imaging capability for the on-board kV imaging systems of conventional linear accelerators towards avoiding motion and registration artifacts. Although a clinical DLI prototype has demonstrated feasibility on a TrueBeam system, achievable spectral separation and image quality depend on the detector stack design, particularly scintillator-layer thickness. This study develops and experimentally validates a Monte Carlo (MC) model of a clinical DLI to systematically investigate scintillator thickness effects on spectral separation.
Methods
A detailed MC model of a DLI was implemented in GATE/Geant4, incorporating x-ray transport, scintillation, and optical photon propagation. The detector model was validated against a clinical prototype by comparing simulated and measured MTFs. Top and bottom scintillator thicknesses were independently varied from 0.3 to 1.0 mm, yielding 64 configurations. Spectral separation was quantified using the mean energy difference (ΔE) and spectral overlap (IoU) between layers. Spatial resolution was evaluated using modulation transfer functions (MTFs) derived from a slanted-edge, and detective quantum efficiency (DQE(0)) was calculated from layer-specific quantum efficiency and Swank factor. Trade-offs were analyzed using Pareto optimization and TOPSIS ranking.
Results
Simulated and measured MTFs agreed closely (NRMSE < 0.02), confirming model accuracy. Across designs, increasing scintillator thickness improved spectral separation, with ΔE rising from ~13 to 25 keV and IoU decreasing from ~0.55 to 0.34. Thickness variations introduced trade-offs between spatial resolution and DQE(0) due to competing x-ray absorption and optical spread. Multi-objective optimization identified a subset of high-performing designs, among which a 0.8-mm top / 1.0-mm bottom configuration provided the most balanced performance.
Conclusion
The validated MC framework demonstrates that optimization of scintillator thickness pairs can enhance spectral separation while preserving clinically acceptable image quality, underscoring the role of prospective detector design in dual-layer radiotherapy imaging systems.