Enhanced optimization of volumetric modulated arc therapy plans using Monte Carlo generated beamlets

Area of Science:

• Medical Physics
• Radiation Oncology
• Computational Biology

Background:

• Treatment planning systems (TPS) optimize volumetric modulated arc therapy (VMAT) plans using objective functions, but rapid dose calculation algorithms limit accuracy.
• Monte Carlo (MC) routines offer high accuracy but are too slow for comprehensive VMAT optimization.

Purpose of the Study:

• To introduce and evaluate a novel enhanced optimization (EO) approach for improving VMAT plan dose distributions.
• To leverage MC calculations for refining TPS-generated VMAT plans by applying small perturbations.

Main Methods:

• EO utilizes TPS VMAT plans as a starting point, applying perturbations based on MC-calculated beamlet dose matrices.
• Beamlet doses are computed using EGSnrc MC toolkit on a distributed-computing framework.
• A greedy search algorithm minimizes a ternary-valued objective function to determine optimal control point parameters.

Main Results:

• EO improved objective scores (6-60%) and DVHs for brain and head and neck VMAT plans, preserving target dose while reducing OAR doses.
• Prostate VMAT plans showed reduced objective scores (46-79%), but substantial DVH improvements were limited.
• Stricter objectives in EO lowered OAR dose in a pediatric brain case without compromising target dose.

Conclusions:

• A novel EO method using MC calculations effectively refines VMAT plan dose distributions.
• This approach is particularly beneficial for complex treatments involving critical organs at risk (OARs).
• Further development aims to reduce computation time and enhance EO sophistication for treatment planning.