Constrained Kalman filtering for IMRT optimization

In this paper, we develop a new optimization algorithm for Intensity Modulated Radiation Therapy (IMRT) planning using a dynamical systems approach. IMRT is a technique used to deliver precise radiation doses to a targeted tumor while avoiding surrounding critical structures. Modeling and optimizing an IMRT plan is a challenging task since tumors can change in size, shape, and position during the course of fractionated treatment. Accordingly, re-optimization of radiation beam weights is generally needed for each treatment fraction, which is computationally expensive and practically inefficient. We establish a recursive algorithm, which is based on the development of a constrained Kalman filter, to develop a complete treatment plan that takes changes in the tumor geometry into account. We implement this new algorithm to a model case demonstrating its robustness for IMRT planning.