A column generation heuristic for VMAT planning with adaptive CVaR constraints.

Area of Science:

• Medical Physics
• Computational Optimization
• Radiation Oncology

Background:

• Volumetric Modulated Arc Therapy (VMAT) is a complex radiation therapy technique requiring precise treatment planning.
• Optimizing VMAT plans involves balancing dose delivery to the tumor while minimizing dose to healthy tissues.
• Current treatment planning systems often require significant manual intervention and computational time.

Purpose of the Study:

• To develop an efficient, automated computational procedure for generating VMAT treatment plans.
• To minimize the total monitor units (MUs) in VMAT plans while adhering to strict dose-volume constraints.
• To compare the performance of the developed automated method against a commercial treatment planning system.

Main Methods:

• A column generation heuristic based on a mixed integer linear programming model was developed.
• The objective function minimized total monitor units.
• Conditional value-at-risk constraints were used to ensure dose-volume requirements were met, with a two-phase approach for plan generation and refinement.

Main Results:

• The automated procedure successfully generated medically acceptable VMAT treatment plans for prostate cancer.
• Plans generated by the new method required significantly fewer monitor units (approximately [Formula: see text] less on average) compared to those from the Eclipse system.
• The procedure is fully automated, requiring no human intervention for plan generation.

Conclusions:

• The proposed computational procedure offers an efficient and automated alternative for VMAT treatment planning.
• This method demonstrates potential for improving treatment plan quality by reducing monitor units, which can lead to faster treatment delivery and potentially reduced delivery uncertainties.
• The findings suggest a promising direction for advancing radiation oncology treatment planning through computational optimization.