(Sunday, 7/12/2020) [Eastern Time (GMT-4)]
Room: AAPM ePoster Library
Purpose: There are emerging interests and developments of other radiation types than photons such as protons and carbon ions, owing to their unique biological and dosimetric characteristics that are distinctive from photons. We investigate the feasibility of a systematic approach to optimizing treatment plans with multiple radiation modalities to identify an optimal combination of the modalities and their fractionation regimens.
Methods: We develop a framework to maximize the biological effect (BE) of multiple modalities on the tumor while constraining the BE of organs-at-risk (OAR) to their tolerance. The optimization variables are the spatial dose distribution and number of fractions of each modality. This formulation gives rise to non-convex, mixed integer programming and we propose a bilevel optimization algorithm to efficiently solve it. The upper level optimizes the fractionation using the value function defined with the spatial dose distribution optimized in the lower level. We demonstrate the feasibility of our framework and algorithms in a simple 2-dimensional head-and-neck phantom with two different radiation modalities (M1+M2). M1 used the dose deposition matrix of 6 MV photons and the alpha/beta ratio of 10Gy for the tumor, 2Gy for the cord and unspecified tissue, and 5Gy for the right and left parotids. M2 used the dose deposition matrix of 250 MeV protons and its alpha/beta ratio of the tumor and OAR were varied.
Results: The dual modalities were optimal in all our simulations and the tumor BE was increased by 7.8 to 22.9% with optimal solutions depending on the tumor doubling time compared to a conventional regimen, which assumes M1 only for 25 fractions.
Conclusion: We successfully set up a framework to optimize treatment plans involving multiple modalities. The results of our numerical simulations agree with the clinical intuition, validating our approach and showing the promise of the framework for further clinical investigation.
Funding Support, Disclosures, and Conflict of Interest: This work was funded in part by National Science Foundation (NSF) Grant #1560476.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email
