ESTRO 36 Abstract Book

S112 ESTRO 36 2017 _______________________________________________________________________________________________

interfaces, stopping power ratios are playing a role.

Conclusion As the range of the interface effects is less than commonly used voxel sizes for MV photon beams, one can in general use the ratio of attenuation coefficients to convert dose to medium to dose to water. As a direct consequence, uncertainties on stopping powers (due to uncertainties on the ionization potential of tissues) have negligible impact on the dose to medium to dose to water conversion. Dose to water obtained by multiplying dose to medium with stopping power ratios is another concept not consistent with how dose to water is calculated by conventional TPSs. OC-0227 The heterogeneous multiscale model for efficient computation of microscopic dose metrics M. Martinov 1 , R. Thomson 1 1 Carleton University, Physics, Ottawa, Canada Purpose or Objective In the field of radiation therapy, there is increasing interest in the effects of ionizing radiation on short (micrometre, nanometre) length scales within macroscopic (~centimetre) volumes of interest. A common technique for studying radiation transport and energy deposition is via Monte Carlo (MC) simulations, requiring complex simulation geometries that are both reliable and efficient, two traits often in contention. This work introduces a general MC framework, the Heterogeneous MultiScale (HetMS) model, to address these challenges. Material and Methods The HetMS model involves combining distinct models of varying level of detail on different length scales into a single simulation. We implement and present the HetMS model using EGSnrc for the context of gold nanoparticle (GNP) radiation therapy. Our model is verified via comparison with recently-published results of other MC GNP simulations. We then consider two scenarios: (A) 20 keV to 1 MeV photon beams incident on a 1 cm radius and 3 cm long cylindrical phantom; (B) low-dose rate brachytherapy sources at the centre of a 2.5 cm radius sphere with GNPs diffusing outwards from the centre (Fig. 1). In each simulation, homogenized tissue/gold mixtures are employed in larger volumes, with distinct subvolumes containing GNPs discretely modelled in pure tissue. Dose scored in pure tissue within the subvolumes is compared to dose scored in homogeneous tissue/gold mixtures and dose to pure tissue (to compute Dose Enhancement Factors (DEFs)).

Results HetMS simulations are able to efficiently account for important macroscopic and microscopic effects, successfully modelling the competing effects of photon fluence perturbation (due to modelling of gold/tissue mixtures in macroscopic volumes) coupled with enhanced local energy deposition (due to discrete modelling of GNPs within subvolumes). Energy deposition is most sensitive to these competing effects for lower energy sources, with considerable variations in DEFs for different source energies, depths in phantom, gold concentrations, and GNP sizes. For the cylinder phantom with 20 mg Au/g tissue, DEFs near 3.1 are observed near the phantom surface and decrease to less than one by 7 mm depth (i.e. dose decreases, not enhancements). Within the spherical phantom, DEFs vary with time for diffusion, radionuclide, and radius; DEFs differ considerably compared to those computed using a widely-applied analytic approach (Fig. 2). Compared with discrete modelling of GNPs within entire macroscopic geometry, HetMS simulations offer efficiency enhancements of up to a factor of 120.

Conclusion The HetMS framework enables efficient simulation of both macroscopic and microscopic effects that must both be considered for accurate simulation of radiation transport and energy deposition. The HetMS model allows for MC simulations, typically prohibited by dense parameter spaces, to be employed in diverse radiotherapy and radiation protection scenarios.