Fast VMAT Planning with Interactive Real-Time Dose Manipulation
K Otto*,University of British Columbia, BC, CanadaTH-A-137-1 Thursday 8:00AM - 9:55AM Room: 137
Purpose: Develop a planning platform for real-time interactive manipulation of dose distributions including DVHs and other dose metrics.
Methods: Intercomparisons of fixed beam IMRT, VMAT and Tomotherapy have been performed by several investigators. Other studies have compared number of beams, number of arcs and type of MLC. Generally these studies have shown only small differences in dose distribution quality. The hypothesis presented here is that delivery systems for photon external beam radiation therapy have reached a fundamental limit in their ability to create arbitrary dose distributions and that achievable dose distributions may be modeled more efficiently without requiring an exact representation of delivery parameters and beam characteristics. An interactive dose manipulation system was developed that incorporates a novel method for modeling achievable dose distributions. Computationally efficient methods for 3D fluence projection, photon scatter and electron transport were developed. Graphical navigation of dose distributions is achieved by a sophisticated method of identifying contributing fluence elements, modifying those elements and re-computing the entire dose distribution.
Results: Full 3D dose distributions are calculated in ~15-20 milliseconds. Including graphics processing overhead, clinicians may visually interact with the dose distribution (e.g. "drag" a DVH) and display updates of the dose distribution at a rate of ~ 20 times per second. Testing on various sites shows that interactive planning is completed in ~2-5 minutes, depending on the complexity of the case (number of targets and OARs). Dose distributions were shown to be achievable through a final plan optimization step using a conventional VMAT planning system.
Conclusion: Real-time interactive planning using DVHs and other dose metrics was realized through the use of a novel dose modeling and dose manipulation system. Trade-offs between Target(s) and OARs may be evaluated efficiently providing a better understanding of the dosimetric options available to each patient in static or adaptive RT.
Funding Support, Disclosures, and Conflict of Interest: K Otto has a commercial interest in the presented materials