Techniques for the Enhancement of Optimization Schemes in Radiation Therapy
Spaans, Jason D.
MetadataShow full item record
With the advent of intensity modulated radiation therapy (IMRT), the process of generating a radiotherapy treatment plan has become a complex and time consuming process due to an increase in the number of variables which must be optimized. These large variable sets, in the form of beamlet fluence dose distributions, must be weight-optimized in order to produce a radiotherapy treatment plan that homogeneously delivers the prescribed dose to the tumor volume and delivers an acceptable amount of dose to normal tissue. Therefore the ability of optimization algorithms to generate radiotherapy plans that conform to these parameters and to produce these plans in a clinically-timely manner has become paramount. In addition, the introduction of multi-criteria optimization increases the number of optimized plans which need to be generated further increasing optimization time. This increase in optimization complexity can be addressed in two different ways: current optimization algorithms can be made more efficient by reducing the number of mathematical operations they require; and the possibility of applying a new optimization technology, such as a quantum annealing algorithm, may be explored. I have investigated reducing the number of mathematical operation required in optimization by sampling precomputed dose matrices. The results of this study are presented and indicate that it is a promising methodology for IMRT optimization. Additionally, the results from the application of a quantum annealing algorithm based on a novel quantum computing system to the IMRT optimization problem are discussed.