A novel automated framework for multi-engine Monte Carlo model commissioning in proton therapy.

Area of Science:

• Medical Physics
• Computational Biology
• Radiotherapy

Background:

• Proton beam scanning (PBS) proton therapy requires accurate commissioning of beam models, which presents a significant challenge.
• Current methods can be time-consuming and prone to variability.

Purpose of the Study:

• To develop and validate an automated Monte Carlo (MC) modeling framework for standardizing and optimizing beam model commissioning in PBS proton therapy.
• To improve the efficiency and reproducibility of the commissioning process.

Main Methods:

• An automated, Python-based framework was developed, incorporating optimization algorithms (particle swarm optimization, Nelder-Mead).
• The framework utilizes a modular pipeline for data preparation, phase space fitting, energy spectrum tuning, and dose calibration.
• Validation was performed using 20 clinical cases and over 100 measurement planes with TOPAS and MCsquare MC engines.

Main Results:

• The automated framework achieved high accuracy, with maximum range errors of 0.3% (TOPAS) and 0.6% (MCsquare).
• Quality assurance plans showed excellent agreement, with median gamma pass rates of 100.0% (TOPAS) and minimum pass rates above 94.3% (MCsquare) using the 3%/3mm criterion.

Conclusions:

• The open-source, Python-based framework offers a robust and extensible solution for automated multi-engine MC beam commissioning.
• This automation enhances reproducibility and efficiency, benefiting both clinical and research applications in medical physics.