Particle Swarm Optimisation Applied to the Direct Aperture Optimisation Problem in Radiation Therapy.

Area of Science:

• Radiation Oncology
• Computational Optimization
• Medical Physics

Background:

• Intensity Modulated Radiation Therapy (IMRT) is a standard cancer treatment.
• IMRT planning is complex, often divided into sequential Beam Angle Optimization (BAO), Fluence Map Optimization (FMO), and Sequencing problems.
• Sequential approaches can lead to treatment plans with excessive apertures and long delivery times.

Purpose of the Study:

• To address limitations of sequential IMRT planning by proposing a Direct Aperture Optimization (DAO) approach.
• To integrate FMO and Sequencing problems into a single DAO problem, considering physical delivery constraints.
• To solve the DAO problem using Particle Swarm Optimization (PSO) for aperture shape optimization and mathematical programming for intensity optimization.

Main Methods:

• Developed a hybrid approach combining mathematical programming for intensity optimization and PSO for aperture shape optimization.
• Integrated a reparation heuristic to refine aperture shapes with minimal impact on treatment efficacy.
• Applied the proposed algorithm to prostate cancer cases for comparative analysis.

Main Results:

• The PSO-based DAO approach achieved competitive results compared to the traditional sequential method.
• The proposed method resulted in reduced beam-on time, indicating shorter treatment delivery durations.
• Demonstrated efficient utilization of Multi-Leaf Collimator (MLC) apertures.

Conclusions:

• The Direct Aperture Optimization (DAO) strategy, particularly with the proposed PSO integration, offers an effective alternative to sequential IMRT planning.
• This approach can generate clinically acceptable treatment plans more efficiently, reducing patient treatment time.
• The findings suggest potential for improved IMRT delivery efficiency and patient outcomes.