TU-G-BRB-04: A Robust-CVaR Optimization Approach to Left-Sided Breast IMRT

Area of Science:

• Radiation Oncology
• Medical Physics
• Cancer Treatment

Background:

• Intensity-modulated radiation therapy (IMRT) is a standard treatment for breast cancer.
• Ensuring target coverage while minimizing dose to organs at risk, particularly the heart, is crucial for left-sided breast cancer patients.
• Uncertainty in patient breathing patterns can compromise treatment plan efficacy and safety.

Purpose of the Study:

• To evaluate the feasibility of a cardiac-sparing IMRT planning approach for left-sided breast cancer using a robust optimization model.
• To assess the effectiveness of a robust optimization model incorporating conditional-value-at-risk (CVaR) in managing breathing uncertainties.

Main Methods:

• A novel robust optimization model using conditional-value-at-risk (CVaR) was developed for breast IMRT.
• The model was tested using a 4DCT dataset simulating breathing variations (20% lung volume change across 5 phases).
• Treatment plans generated by the robust model were compared against conventional clinical IMRT methods.

Main Results:

• The robust-CVaR model achieved a 14.6% reduction in mean heart dose compared to conventional free-breathing IMRT.
• Target coverage and dose homogeneity were maintained, with clinical dose-volume limits met for the heart and target.
• A 23.9% improvement in maximum heart dose (to 25cc) was observed, with low plan quality variability.

Conclusions:

• Robust optimization with CVaR limits enhances IMRT treatment quality for breast cancer.
• The robust-CVaR approach provides high-quality plans deliverable during free breathing without external devices.
• Consistent treatment quality under irregular breathing was demonstrated; further improvements may come from incorporating lung motion metrics.