Incorporating dose-volume constraints in linear fluence map optimization for intensity modulated proton therapy

Prescriptions for radiation treatment planning are given in terms of dose-volume constraints (DVC). In our current linear programming (LP) approach to satisfy DVCs for solving the fluence map optimization (FMO) problem, very strict and conservative "hot spot" control parameters are imposed for organs-at-risk (OARs) as well as the target. However, it is not clear how to set these control parameters in our current FMO model. In some cases, the present values of the parameters are shown to be unnecessarily conservative, which results in inferior solution quality. Therefore, we have developed a heuristic algorithm that aims to satisfy DVCs based on an iterative linear fluence map optimization process. This algorithm begins by setting loose bounds on these parameters; then it gradually tightens the bounds, if necessary, through an iterative solution process. This method finds appropriate parameter values through the trade off between OARs sparing and target coverage to improve solution iteratively. Experiments are made to show the computational advantage of the algorithm using actual clinical patient cases for proton therapy.