Abstract
This paper will present the results of an investigation into three iterative approaches to inverse treatment planning. These techniques have been examined in the hope of developing an optimization algorithm suitable for the large-scale problems that are encountered in tomotherapy. The three iterative techniques are referred to as the ratio method, iterative least-squares minimization and the maximum-likelihood estimator. Our results indicate that each of these techniques can serve as a useful tool in tomotherapy optimization. As compared with other mathematical programming techniques, the iterative approaches can reduce both memory demands and time requirements. In this paper, the results from small- and large-scale optimizations will be analysed. It will also be demonstrated that the flexibility of the iterative techniques can be greatly enhanced through the use of dose-volume histogram based penalty functions and/or through the use of weighting factors assigned to each region of the patient. Finally, results will be presented from an investigation into the stability of the iterative techniques.