An automated optimization strategy to design collimator geometry for small field radiation therapy systems

Area of Science:

• Medical Physics
• Radiotherapy Technology
• Computational Optimization

Background:

• Small-field radiotherapy systems require precise collimator design for accurate dose delivery.
• Optimizing collimator geometry is crucial for achieving desired beam characteristics like penumbra and dose rate.
• Existing systems like CyberKnife provide benchmarks for small-field collimator performance.

Purpose of the Study:

• To develop an automated optimization strategy for designing collimators in small-field radiotherapy.
• To link dose profile characteristics with treatment head geometric parameters.
• To apply and validate the strategy using a linac-based radiosurgery system.

Main Methods:

• Developed an objective function integrating dose profiles (FWHM, penumbra, central dose rate) and geometric parameters (collimator thickness/radii, SDC).
• Employed a downhill simplex algorithm for optimization.
• Applied the strategy to design collimators for 5, 10, 15, and 20 mm photon beams, comparing results to CyberKnife.

Main Results:

• Penumbra-prioritized optimization yielded parameters closely matching CyberKnife (radii within 3-14%, output factors within 0-8%, penumbras within 2%).
• Dose rate-prioritized optimization increased central dose rate by 1.5-2 times for larger beams (10-20 mm) but widened penumbras by 1.5-2 times.
• No improvement in central dose rate was observed for the 5 mm beam size.

Conclusions:

• An automated optimization strategy for collimator design in small-field radiotherapy was successfully developed.
• The strategy's penumbra-prioritized results align well with the CyberKnife system, indicating its design likely prioritizes sharp penumbras.
• This work demonstrates the principle of applying automated optimization to complex collimator designs.