Robust optimization for intensity modulated radiation therapy treatment planning under uncertainty

Area of Science:

• Radiation Oncology
• Medical Physics
• Computational Biology

Background:

• Intensity modulated radiation therapy (IMRT) enables precise dose tailoring to tumor geometry.
• Conventional treatment planning assumes static tumor position, which is often inaccurate due to organ motion and patient positioning variability.
• Tumor location can significantly deviate during the course of radiation therapy.

Purpose of the Study:

• To develop a probabilistic model for the intensity modulated radiation therapy (IMRT) inverse problem.
• To demonstrate the equivalence between the proposed probabilistic model and robust optimization techniques under specific assumptions.
• To evaluate the computational feasibility and clinical promise of the new model for improving treatment planning.

Main Methods:

• Formulation of a probabilistic model for the IMRT inverse problem.
• Demonstration of mathematical equivalence to robust optimization methods.
• Application and computational evaluation of the model using a sample prostate cancer case.

Main Results:

• The probabilistic model was found to be computationally feasible for a sample prostate case.
• The proposed method showed potential for improved healthy tissue sparing compared to traditional treatment planning.
• The model effectively maintained the prescribed dose to the target while enhancing organ sparing.

Conclusions:

• The probabilistic model offers a promising advancement in IMRT treatment planning.
• This approach addresses the challenge of non-stationary tumor targets in radiation therapy.
• The method has the potential to optimize radiation delivery, leading to better patient outcomes through enhanced tissue sparing.