Intensity modulated arc therapy is a novel treatment technique that has shown great potential to be superior to conventional intensity modulated radiotherapy, both in terms of treatment plan quality as well as treatment delivery. Based on previous literature, a simplified technique called two-step intensity modulated arc therapy (2-step IMAT) was implemented into a treatment planning system. In order to automatically generate treatment plans for this technique, a beam portal shaping method was developed to generate beam segments. A sensitivity analysis was carried out on a geometric phantom to determine optimal parameters for the 2-step IMAT implementation for that particular phantom. The segment weights were optimized using the dose-based and dose-volume-based objective functions. The optimal solution search was based on the gradient-descend algorithm. The dose-based objective function was implemented using a so-called lambda-value-dose-based objective function developed in this work in order to increase both speed and flexibility of the optimization. The successful implementation demonstrated the feasibility of automatic 2-step IMAT treatment planning.
A comparison of conventional arc therapy and 2-step IMAT showed improvements in the target dose uniformity by about 50% for both geometric phantom and clinical paraspinal tumor case, whilst also improving the organ sparing. The comparisons between the lambda-value-dose-based and dose-volume-based optimizations showed a speed advantage of the former by a factor of over five in the phantom study.
The current beam portal shaping approach can be improved by optimizing the segment width and including multiple organs-at-risk in the segment generation algorithm. Future work will also include the implementation of a stochastic optimization to minimize the chance of getting trapped in local minima during the segment weight optimization. In summary, the work of this research showed that the automatic 2-step IMAT planning is a viable technique that can result in highly conformal plans while keeping the treatment planning and delivery simple and straightforward.
Identifer | oai:union.ndltd.org:canterbury.ac.nz/oai:ir.canterbury.ac.nz:10092/3844 |
Date | January 2010 |
Creators | Sun, Jidi |
Publisher | University of Canterbury. Department of Physics and Astronomy |
Source Sets | University of Canterbury |
Language | English |
Detected Language | English |
Type | Electronic thesis or dissertation, Text |
Rights | Copyright Jidi Sun, http://library.canterbury.ac.nz/thesis/etheses_copyright.shtml |
Relation | NZCU |
Page generated in 0.0011 seconds