Lexicographic ordering: intuitive multicriteria optimization for IMRT.

Area of Science:

• Radiation Oncology
• Medical Physics
• Computational Biology

Background:

• Intensity-Modulated Radiation Therapy (IMRT) inverse planning involves multiple, often conflicting, objectives for target coverage and critical structure sparing.
• Clinical decision-making requires balancing these objectives, typically through iterative optimization processes that are time-consuming and complex, especially with numerous planning goals.

Purpose of the Study:

• To implement and evaluate a multicriteria optimization strategy, lexicographic ordering (LO), for addressing complex IMRT inverse planning problems.
• To demonstrate LO's ability to manage a large number of planning goals by categorizing them into hierarchical priority levels.

Main Methods:

• Lexicographic ordering (LO) was applied, involving the sequential optimization of planning goals based on predefined priority levels.
• The LO approach was tested on two clinical cases: a prostate cancer case and a head and neck cancer case, with up to 23 planning goals.

Main Results:

• The LO method successfully optimized comprehensive lists of planning goals using only a few priority levels, prohibiting tradeoffs between higher and lower priority objectives.
• Optimization times per level were practical, ranging from 26 to 217 seconds, simplifying large-scale problem representations.
• The LO approach effectively mimicked physician decision-making processes for conflicting planning goals.

Conclusions:

• Lexicographic ordering offers an intuitive and efficient solution for multicriteria optimization in IMRT inverse planning.
• This hierarchical strategy makes complex treatment planning more manageable and produces encouraging results for challenging clinical cases.