A Greedy reassignment algorithm for the PBS minimum monitor unit constraint.

Area of Science:

• Medical Physics
• Radiation Oncology
• Computational Imaging

Background:

• Proton pencil beam scanning (PBS) treatment planning involves numerous unique beam spots with varying weights.
• Treatment planning systems optimize these weights, but some may fall below the minimum deliverable threshold of the treatment delivery system.
• This limitation can compromise the accuracy and efficacy of proton therapy treatments.

Purpose of the Study:

• To investigate a novel Greedy reassignment algorithm for post-processing proton PBS treatment plans.
• To mitigate the effects of low-weight pencil beams by ensuring all spots meet the minimum monitor unit (MU) constraint.
• To generate deliverable treatment plans for clinical implementation.

Main Methods:

• Developed a Greedy reassignment algorithm that deletes the smallest weight spot and reassigns its weight to nearest neighbors until all spots exceed the MU constraint.
• Evaluated the algorithm's performance on 496 fields from 190 patient treatment plans.
• Compared the Greedy method against two other post-processing techniques using the gamma-index pass rate and a developed planning metric.

Main Results:

• The Greedy reassignment method demonstrated a 1.8 times better planning metric at a 90% gamma-index pass rate compared to other methods.
• A strong correlation was observed between the planning metric and the gamma-index pass rate for the Greedy algorithm.
• The planning metric showed a standard deviation of 18% of the centroid value for fields with a 90% ± 1% pass rate.

Conclusions:

• The Greedy reassignment algorithm effectively generates deliverable proton PBS treatment plans by addressing low-weight spots.
• The developed planning metric can aid in optimizing treatment planning parameters and facility design to achieve acceptable gamma-index pass rates.
• This approach is crucial for facilities implementing PBS with small spot sizes and spacing, where minimum MU constraints significantly impact plan quality.