Incorporating uncertainties in respiratory motion into 4D treatment plan optimization.

The purpose of this work is to investigate robust 4D optimization techniques which account for respiratory motion uncertainties. Two robust optimization techniques were applied to generate 4D optimized lung treatment plans. The probabilistic optimization approach minimizes the dose variance in the target volume while the worst case optimization minimizes a weighted combination of the nominal and worst case dose distributions which occur in the presence of respiratory motion variation. The two 4D optimization approaches were compared with a margin-based midventilation planning approach in five lung patients. Respiratory motion amplitude and baseline variations were quantified from tidal volume measurements during planning 4D CT acquisition. A similar target coverage was obtained for all three approaches, although the 4D optimization methods tended to be better at sparing the organs at risk. Both robust planning methods are suited for automatic determination of treatment plans which ensure target dose conformality under respiratory motion variations, while minimizing the dose burden of healthy lung tissue.