Click here to


Are you sure ?

Yes, do it No, cancel

Development of a TEM Image-Based Cellular Geometry Model for Geant4 Monte Carlo Simulations of GNP-Mediated Dose Enhancement/radiosensitization

S Jayarathna*, S Krishnan , S H Cho , The University of Texas MD Anderson Cancer Center, Houston, TX


(Sunday, 7/29/2018) 2:05 PM - 3:00 PM

Room: Davidson Ballroom A

Purpose: To develop a cellular geometry model based on the Transmission Electron Microscopy (TEM) images integrated with the Geant4 particle transport toolkit for Monte Carlo simulations of radiobiological phenomena such as gold nanoparticle (GNP)-mediated radiosensitization.

Methods: A given TEM image was read using a custom CERN ROOT-V.6.6 macro and saved as a Geant4 acceptable geometry file in the format of Geometry Development Markup Language (GDML). Specifically, thresholds were imposed depending on the RGB values of the TEM image and the corresponding material density was assigned to each pixel. An arbitrary height was assigned to each pixel to transform it to a voxel and the GDML file was imported into Geant4 via inbuilt I/O methods. In the current study, a TEM image showing a single cell containing internalized GNPs in the cellular endosomes was used to demonstrate the concept of the proposed cell model. A clinically acceptable Yb-169 photon spectrum was applied along the XY-plane and the doses to the nucleus and surrounding pixels were calculated. The GNPs were unloaded from the cell by replacing the corresponding voxels by water voxels and the same procedure was repeated to investigate the GNP-mediated dose enhancement to the nucleus.

Results: The constructed cellular geometry model was successfully run on a dedicated high-performance computer cluster with Geant4-V-10.3. In the case of this model, the simulation results showed a factor of 1-10 GNP-mediated dose enhancement in the cellular nucleus region.

Conclusion: The current approach can be applied to develop cellular geometry models based on TEM images for Geant4 Monte Carlo simulations. In particular, the capability of importing critical geometrical features such as GNP cluster distributions into Geant4 leads to a cellular geometry model with greater details, enabling highly sophisticated and possibly predictive computational investigations of physical contributions to radiobiological phenomena such as GNP-mediated radiosensitization.

Funding Support, Disclosures, and Conflict of Interest: Supported by NIH/NCI grant R01CA155446 and CPRIT grant RP160497


Monte Carlo, Radiobiology


TH- Radiobiology(RBio)/Biology(Bio): RBio- general

Contact Email