Methodology for simulating x-ray sources of computed tomography systems using GATE 10 without manufacturer data

Area of Science:

• Medical Physics
• Imaging Science
• Computational Modeling

Background:

• Manufacturer data for computed tomography (CT) X-ray sources is often unavailable.
• Accurate CT simulations require realistic modeling of X-ray source characteristics.
• Monte Carlo (MC) simulations are crucial for CT system development and optimization.

Purpose of the Study:

• To implement and evaluate an experimental source modeling method for MC simulations of CT systems.
• To enable realistic CT X-ray source simulation using experimental measurements, particularly X-ray spectrometry.
• To assess the applicability of the developed method for single-energy CT (SECT) scanners.

Main Methods:

• Implemented an experimental method combining X-ray spectra (CdTe spectrometer) and bowtie filter attenuation profiles (ionization chamber).
• Developed two GATE 10 source models: Single-spectrum (on-axis) and Multi-spectra (on- and off-axis).
• Evaluated models by comparing simulated and measured half-value layers (HVL) and CT dose index (CTDI) values.

Main Results:

• The Multi-spectra and Single-spectrum models showed improved agreement with measured HVLs (mean difference of 4%).
• Multi-spectra model simulations yielded CTDI values with <7% difference compared to measurements.
• Manufacturer source data showed the largest discrepancies, particularly at 140 kV.

Conclusions:

• Modeling spectral variations across the X-ray beam significantly enhances CT source model accuracy.
• The proposed framework provides a viable alternative to manufacturer data for CT simulations.
• This work supports the use of MC methods for integrating new CT systems.