Fast, accurate photon beam accelerator modeling using BEAMnrc: a systematic investigation of efficiency enhancing...

Area of Science:

• Medical Physics
• Computational Physics
• Radiation Oncology

Background:

• Monte Carlo (MC) simulations are crucial for accurate radiation therapy planning.
• Optimizing MC code efficiency is essential for reducing computation time.
• Comparing different MC code systems and techniques aids in selecting the most suitable tools.

Purpose of the Study:

• To investigate efficiency-enhancing methods in the BEAMnrc MC code system.
• To compare the performance of BEAMnrc with VMC++, another MC code for treatment head simulation.
• To evaluate the impact of different cross-section data and variance reduction techniques on simulation accuracy and efficiency.

Main Methods:

• Simulated a 6 MV photon beam from a Siemens Primus linear accelerator using BEAMnrc and VMC++.
• Generated phase space files at various field sizes.
• Investigated photon/bremsstrahlung cross sections, approximate efficiency improving techniques (AEITs), and variance reduction techniques (VRTs) in BEAMnrc.
• Compared simulation efficiency (relative efficiency) and dose distributions in water phantoms against measurements.

Main Results:

• VMC++ demonstrated significantly higher relative efficiencies, simulating treatment heads in minutes.
• BEAMnrc achieved highest efficiencies with directional bremsstrahlung splitting (DBS) alone or combined with charged particle range rejection.
• Noteworthy differences (1%-3%) observed between NIST and Bethe-Heitler cross sections, but dose distributions agreed within 2% of measurements.

Conclusions:

• VMC++ offers rapid and accurate simulations of linear accelerator treatment heads.
• BEAMnrc's efficiency can be substantially improved using specific VRTs and AEITs.
• Both investigated MC systems, with appropriate techniques and cross-sections, provide dose calculations within acceptable clinical agreement with measurements.