Determination of parameters for a multiple-source model of megavoltage photon beams using optimization methods

Area of Science:

• Medical Physics
• Radiotherapy
• Accelerator Physics

Background:

• Accurate modeling of medical linear accelerator radiation output is crucial for effective radiotherapy treatment planning.
• A key challenge is adapting generic models to specific treatment units.
• Multiple-source models with adjustable physical parameters offer a promising approach.

Purpose of the Study:

• To develop and validate an automated procedure for deriving parameters of a multiple-source model for medical linear accelerators.
• To assess the accuracy of the model in predicting radiation output across various beam energies and accelerator vendors.
• To evaluate the model's performance for different field configurations and irregular shapes.

Main Methods:

• Utilized a multiple-source model incorporating bremsstrahlung, extra-focal photon, and electron contamination sub-sources.
• Employed optimization methods to derive model parameters from standard beam data measurements.
• Used a gamma error measure (1%, 1 mm) as the cost function for optimization.
• Applied the procedure to six beam data sets (6 MV to 23 MV) from three vendors, focusing on Varian accelerators.

Main Results:

• Achieved an average gamma error < 0.7 for all studied treatment beams and field sizes.
• A minimum of 83% of measurement points passed the gamma < 1 criterion.
• Average gamma errors were < 1.1 for varying SSDs and asymmetric fields.
• Differences in point doses for irregularly shaped MLC apertures were < 1.0%.

Conclusions:

• Source model parameters for medical linear accelerators can be automatically derived using optimization and standard measurements.
• The developed procedure demonstrates acceptable accuracy and processing time.
• The method is applicable to a wide range of accelerators, enhancing radiotherapy treatment planning.