Technical note: A GPU-based shared Monte Carlo method for fast photon transport in multi-energy x-ray exposures

Area of Science:

• Medical Imaging Physics
• Computational Science
• Particle Physics

Background:

• Monte Carlo (MC) methods offer accurate particle transport calculations but are computationally expensive.
• Graphics Processing Unit (GPU) acceleration improves MC efficiency for single-energy CT.
• Multi-energy CT simulations face redundant calculations with conventional individual simulation approaches.

Purpose of the Study:

• To develop a novel GPU-based shared MC (gSMC) scheme for efficient scatter estimation in multi-energy x-ray imaging.
• To reduce redundant photon simulations across different energy spectra.

Main Methods:

• The gSMC method employs spectral region classification to group photons with identical initial energies.
• Multi-directional sampling identifies photons with shared energy and direction for parallel simulation.
• Shared photons undergo combined simulations, while unique photons are processed individually.

Main Results:

• gSMC achieved a ~37.8% speedup in digital phantom simulations and ~20.6% in clinical cases.
• The method maintained high accuracy, with total scatter differences within 0.09% compared to conventional MC.
• Demonstrated significant efficiency gains in multi-energy CT scatter estimation.

Conclusions:

• The proposed GPU-based shared MC simulation method enables rapid photon transport calculations for multi-energy x-ray exposures.
• gSMC effectively addresses computational bottlenecks in multi-energy CT imaging.