Industrial
and
EIN 6918: Graduate Seminar
Spring 2008
February 21, 2008
3PM, MAEB 211
The Economic Benefits
of Green Buildings
Kim
LaScola Needy, Ph.D., P.E., CFPIM
Department of Industrial Engineering
University of Pittsburgh
Abstract
The goal of radiation therapy is to deliver a prescribed dose to targets while sparing critical structures. We focus on a radiation delivery technique called Intensity Modulated Radiation Therapy (IMRT), which delivers a treatment plan to a patient through a sequence of many (about 25-50) daily treatments. The core problem in IMRT treatment plan optimization is the problem of finding an optimal intensity profile for each beam. While traditional models account for setup errors as well as changes in patient geometry between daily treatments by enlarging the targets with a margin, we instead use a stochastic programming approach in which interfraction motion is incorporated via randomly generated scenarios. Our model employs the fact that (i) for targets, the daily delivered dose distribution is most important while (ii) for critical structures, the aggregate dose distribution over the duration of the treatment is most important. We show that robust treatment plans can be obtained with only a modest number of scenarios, allowing for treatment plan optimization in a clinically acceptable amount of time. Our results show that, compared to traditional margin-based models we can improve the quality of the delivered dose distribution to the targets while maintaining the quality of the dose distribution delivered to critical structures.