Mixed Integer Programming Applied to Radiation Treatment Planning Optimization

Georgia Tech Research Corporation, Atlanta GA

Investigators

Abstract

Proposal #0098219 Eva K. Lee GA Tech Res Corp -GIT The field of mixed integer programming (MIP) involves a fascinating and lively blend of theoretical analysis, algorithm and software development, and scientific computing. Over the past thirty years, real applications have provided invaluable insight and motivation for advancing the frontiers of this evolving discipline. In particular, the MIP modeling paradigm and the associated algorithmic branch-and-bound, branch-and-cut, and branch-and-price solution strategies have been applied extremely successfully to a broad range of industrial applications, from scheduling and process planning applications, to VLSI and telecommunication network design. Integral to these successful applications of MIP were close collaborations between optimization researchers and the domain experts in each application area. In each case, much research was devoted to developing a superior model (versus simply a correct formulation, as the solution process depends heavily on the underlying structure of the formulated model) and effective computational strategies for the specific application at hand. Beyond this active research came the technological transfer of computational advances into commercial software (e.g., CPLEX, XPRESS), enabling the integration of these advances into the decision support systems used daily by practitioners. The primary focus of this research is to explore applications of mixed integer programming strategies to radio-therapy treatment planning optimization. The initial focus will be radiosurgery treatment planning. Radiosurgery involves small field, stereotactic external beam irradiation to the brain. It has evolved over the past decade into a common method for treating and controlling certain central nervous system lesions such as arteriovenous malformations, metastatic lesions, acoustic neuromas, pituitary tumors, malignant gliomas and other intracranial tumors. Although radiosurgery provides an excellent potential for effectively curing these disorders, the complexities and functional dependencies both within and between brain structures means that the procedure is inherently high-risk and can bring about severe complications. Hence, accuracy and precision are of paramount importance in both the planning and execution of stereotactic radiosurgery.