A robust optimization model for intensity-modulated radiotherapy: Cheap-Minimax.

Area of Science:

• Medical Physics
• Radiation Oncology
• Computational Biology

Background:

• Intensity-modulated radiotherapy (IMRT) is standard but vulnerable to planning and delivery uncertainties.
• The planning target volume (PTV) approach is common but limited, especially for superficial or near-critical structure targets.
• Robust optimization models, particularly minimax, are emerging for intensity-modulated particle therapy, but can be overly conservative in photon treatments.

Purpose of the Study:

• Introduce the Cheap-Minimax (c-minimax) robust optimization model for photon-based IMRT.
• Improve the balance between plan robustness and the 'price of robustness' (dose to organs at risk, OARs).
• Address limitations of PTV and standard minimax approaches in managing uncertainties.

Main Methods:

• The c-minimax model was implemented within the MatRad treatment planning system.
• Applied to 20 clinical cases: 5 prostate and 15 breast cancer.
• Results compared against conventional minimax and PTV-based approaches.

Main Results:

• For prostate cancer, c-minimax offered PTV-comparable robustness with 20% less rectal and 10% less bladder OAR dose than minimax.
• For breast cancer, c-minimax improved robustness by 23.7% (vs. PTV) and 18.2% (vs. minimax), reducing lung and heart doses.
• Both c-minimax and minimax reduced skin dose by ~19% compared to PTV.

Conclusions:

• The c-minimax model effectively addresses PTV limitations and minimax over-conservativeness in IMRT.
• Demonstrates significant advantages in complex cases like breast cancer, enhancing treatment quality.
• Offers a flexible, clinically feasible strategy to reduce toxicity risks and improve patient outcomes through better uncertainty management.