SDE-based Monte Carlo dose calculation for proton therapy validated against Geant4.

Area of Science:

• Medical Physics
• Computational Physics
• Radiotherapy Physics

Background:

• Proton therapy requires accurate dose calculation for effective cancer treatment.
• High-fidelity Monte Carlo simulations (e.g., Geant4) are accurate but computationally intensive.
• Stochastic differential equations (SDEs) offer a potential framework for faster dose calculations.

Purpose of the Study:

• To evaluate the accuracy and computational performance of a novel SDE-based model for proton beam dose calculation.
• To benchmark the SDE model against Geant4 in various phantom geometries.
• To assess the SDE model's ability to handle material heterogeneities and complex dose distributions.

Main Methods:

• Implemented an SDE model using approximations for interaction cross sections and mean excitation energies.
• Benchmarked the SDE model against Geant4 in homogeneous, longitudinally heterogeneous, and laterally heterogeneous phantoms.
• Assessed depth-dose behavior, lateral transport, and the impact of heterogeneities.

Main Results:

• The SDE model accurately reproduced Geant4's depth-dose characteristics, with proton range agreement within 0.6 mm.
• Gamma pass rates exceeded 95% under strict 2%/0.5 mm criteria for all configurations.
• The SDE model demonstrated a 2.5-3x speed-up compared to single-threaded Geant4, with potential for further acceleration.

Conclusions:

• The SDE-based model provides accurate dosimetric predictions comparable to Monte Carlo simulations.
• The SDE approach offers a moderate reduction in computational cost with potential for significant speed improvements.
• This SDE model is a promising candidate for fast and accurate dose calculations in proton therapy.