Assessing stray neutron dose variability across Monte Carlo codes in a proton therapy scenario

Area of Science:

• Medical Physics
• Radiation Dosimetry
• Computational Science

Background:

• Monte Carlo (MC) simulations are crucial for proton therapy dosimetry.
• Significant discrepancies exist in neutron dose estimates across various MC codes, nuclear models, and cross-section libraries.

Purpose of the Study:

• To compare neutron fluence and dose equivalent calculations across five MC codes (PHITS, MCNP, FLUKA, GATE, TOPAS).
• To evaluate the impact of different nuclear data libraries and models on these calculations for a clinical proton beam.
• To provide a reference for expected variations in neutron dose predictions.

Main Methods:

• Simulated a spread-out Bragg peak (SOBP) proton beam in a water phantom.
• Evaluated neutron fluence spectra and dose equivalents at out-of-field positions using default and alternative nuclear data.
• Investigated the influence of proton vs. neutron cross-sections and different cross-section libraries (JENDL-5, JEFF-3.3, TENDL-2021).

Main Results:

• Proton dose distributions showed high consistency (<2 mm spatial variation) across codes.
• Neutron fluences varied by up to 130% and dose equivalents by 88% based on nuclear data.
• Proton cross-sections significantly influenced neutron spectra more than neutron cross-sections.

Conclusions:

• Cross-section library selection is a primary driver of variability in MC neutron dose calculations.
• While default MC settings show internal consistency, experimental benchmarking is essential for accurate neutron dose predictions in proton therapy.