Incorporating organ movements in inverse planning: assessing dose uncertainties by Bayesian inference

Area of Science:

• Medical Physics
• Radiotherapy
• Computational Biology

Background:

• Fractionated radiotherapy requires accurate dose calculation, but inter-fraction organ movements introduce uncertainties.
• Estimating patient geometry probability distributions is crucial for calculating expected dose distributions.

Purpose of the Study:

• To present a novel method for calculating dose uncertainties in fractionated radiotherapy.
• To quantify dose variance arising from finite fractions and estimated geometry probabilities.
• To integrate dose variance considerations into inverse intensity-modulated radiotherapy (IMRT) planning.

Main Methods:

• Developed a method based on Bayesian inference to quantify total dose variance.
• Calculated both expectation and variance distributions of the dose.
• Incorporated dose variance minimization into the inverse planning optimization process.

Main Results:

• The proposed Bayesian inference method quantifies dose uncertainties due to inter-fraction organ motion.
• Considering dose variance in inverse IMRT planning enhances robustness.
• Bayes theorem enables interpolation between patient-specific data and population-based organ motion knowledge.

Conclusions:

• The method provides a robust way to handle uncertainties in radiotherapy dose calculations.
• Minimizing dose variance during inverse planning is essential for reliable treatment.
• Bayesian inference offers a powerful framework for integrating organ motion variability into radiotherapy planning.