Optimization and Systems Theory Seminar


November 25, at 11.00, room 3721, Lindstedtsvägen 25, KTH:
Timothy C. Y. Chan, Industrial Engineering, University of Toronto

Adaptive and robust optimization methods in intensity-modulated radiation therapy

Abstract: The traditional approach to robust intensity-modulated radiation therapy treatment planning involves creating an appropriate uncertainty set to model the uncertain effect, solving a single treatment planning problem, and then delivering the same treatment over multiple weeks. In this talk, I will present an adaptive robust optimization approach to IMRT optimization, where information gathered in previous treatment sessions is used to update a model of uncertainty and guide treatment plan re-optimization for the next session. Such an approach allows for the estimate of the uncertain effect to improve as the treatment progresses. This approach involves solving a sequence of linear programs, and is therefore highly tractable. I will present computational results for a lung cancer case where the dominant uncertainty is in the patient's breathing motion. Using this adaptive robust method, I demonstrate that it is possible to attain simultaneous improvement in both tumor coverage and organ sparing over the non-adaptive approach. I also show that it is possible to closely approximate "prescient" solutions, and provide some theoretical insight as to why this occurs.


Calendar of seminars Last update: November 18, 2011.