Influence of variance reduction techniques on conventional radiotherapy simulations with TOPAS MC

Area of Science:

• Medical Physics
• Computational Physics

Background:

• Monte Carlo simulations are crucial for accurate radiotherapy dosimetry.
• Computational efficiency in these simulations is a persistent challenge.
• Variance reduction techniques (VRTs) offer a promising solution to accelerate calculations.

Purpose of the Study:

• To evaluate the effectiveness of Geometrical Particle Splitting (GPS) and Importance Sampling (Ant Colony Method) as VRTs.
• To improve the computational efficiency of Monte Carlo calculations for the radiotherapy beam quality index, TPR20,10.
• To assess the impact of VRTs on simulation accuracy and processing time.

Main Methods:

• Simulations were performed using TOPAS with a 6 MV Elekta linear accelerator source and a water phantom.
• TPR20,10 was calculated using both VRTs and compared to baseline simulations without VRTs.
• The influence of split planes and particle selection (photons/electrons) on efficiency was analyzed.
• Z-test was used to validate simulated results against experimental data.

Main Results:

• Both GPS and Importance Sampling significantly enhanced computational efficiency.
• Importance Sampling achieved the highest efficiency gain factor (16.91) when applied to both photons and electrons.
• GPS demonstrated an efficiency gain factor of 4.01.
• VRT effectiveness was dependent on geometry and particle selection.

Conclusions:

• VRTs substantially improve computational efficiency for Monte Carlo simulations in radiotherapy.
• Importance Sampling, particularly when applied to both photons and electrons, offers superior gains.
• Careful application of VRTs enhances precision without significantly increasing computational cost.