A fast GPU-accelerated Monte Carlo engine for calculation of MLC-collimated electron fields

Area of Science:

• Medical Physics
• Radiation Oncology
• Computational Science

Background:

• Photon external beam therapies have advanced significantly, while electron beam therapy has lagged.
• Modern linear accelerators offer potential for advanced electron treatments, but dose calculation methods for multi-leaf collimator (MLC) electron beams are lacking.
• Clinical adoption requires dose calculation times comparable to current algorithms.

Purpose of the Study:

• To develop a graphics processing unit (GPU)-accelerated Monte Carlo (MC) engine for rapid dose calculation of electron beams collimated by a conventional photon MLC.
• To incorporate the Varian TrueBeam linear accelerator head geometry into the MC engine.
• To achieve clinically relevant dose calculation speeds.

Main Methods:

• Developed a compute unified device architecture (CUDA) framework for simulating particle transport (electrons and photons) through linac head and CT geometries, including various interactions.
• Modeled the linac head collimating geometry using vendor specifications and phase-space files.
• Benchmarked the MC engine against established codes (EGSnrc/DOSXYZnrc/GEANT) and validated dose distributions against experimental measurements in water and with radiochromic film.

Main Results:

• The GPU-based MC engine achieved dose distributions in good agreement with benchmark codes and experimental measurements for both MLC and jaw-collimated electron beams.
• Dose profiles showed average absolute differences of 1.1 mm (FWHM) and 1.9 mm (80%-20% penumbra) compared to measurements.
• Achieved a dose uncertainty of <1% in approximately 2.5 minutes on an NVIDIA Tesla V100 GPU, demonstrating a speed improvement of ~300 times over single-CPU core methods.

Conclusions:

• The developed GPU-based MC engine enables rapid and accurate dose calculation for electron beams collimated with conventional photon MLCs.
• The accelerated computation times facilitate the rapid calculation of electron fields, paving the way for mixed photon and electron particle therapy.
• This technology addresses a critical gap in electron beam therapy planning and clinical implementation.