Mixed Integer Programming Optimization Models for Brachytherapy Treatment Planning

Eva K. Lee and Richard J. Gallagher

March 1997

Abstract:

Mixed integer programming is proposed as an approach for generating treatment plans for brachytherapy. Brachytherapy is a type of radiation therapy that involves the placement of radioactive sources (seeds) either in tumors (interstitial implants) or near tumors (intracavitary therapy and mold therapy). In this approach radiation is emitted outward and limited to short distances. Thus, unlike external beam radiotherapy, where radiation must traverse normal tissue in order to reach the tumor, brachytherapy is much more localized and therefore reduces radiation exposure to normal tissue while allowing an escalation in the radiation dose. However, the ``optimal'' placement and dosage of the radioactive seeds in brachytherapy is a difficult problem. Two related, but distinct, mixed integer programming models are formulated and tested on data from eight prostate cancer patients. The results demonstrate that in some cases, ``good'' treatment plans can be obtained in less than one CPU minute.

Keywords: Mixed Integer Programming Models, Radiation Treatment Plan Optimization



Lee, School of Industrial and Systems Engineering, Georgia Institute of Technology, GA 30332-0205. Supported in part by NSF CAREER grant 9501584.

    
Gallagher, Department of Medical Informatics, Columbia University, New York, New York. Supported in part through an extramural training grant from the U.S. National Library of Medicine.



Eva Lee
Sun Jan 28 03:53:19 EST 1996