Development of a randomized 3D cell model for Monte Carlo microdosimetry simulations

Abstract

Purpose: The objective of the current work was to develop an algorithm for growing a macroscopic tumor volume from individual randomized quasi‐realistic cells. The major physical and chemical components of the cell need to be modeled. It is intended to import the tumor volume into GEANT 4 (and potentially other Monte Carlo packages) to simulate ionization events within the cell regions. Methods: A MATLAB © code was developed to produce a tumor coordinate system consisting of individual ellipsoidal cells randomized in their spatial coordinates, sizes, and rotations. An eigenvalue method using a mathematical equation to represent individual cells was used to detect overlapping cells. GEANT 4 code was then developed to import the coordinate system into GEANT 4 and populate it with individual cells of varying sizes and composed of the membrane, cytoplasm, reticulum, nucleus, and nucleolus. Each region is composed of chemically realistic materials. Results: The in‐house developed MATLAB © code was able to grow semi‐realistic cell distributions (∼2 × 10 8 cells in 1 cm 3 ) in under 36 h. The cell distribution can be used in any number of Monte Carlo particle tracking toolkits including GEANT 4, which has been demonstrated in this work. Conclusions: Using the cell distribution and GEANT 4, the authors were able to simulate ionization events in the individual cell components resulting from 80 keV gamma radiation (the code is applicable to other particles and a wide range of energies). This virtual microdosimetry tool will allow for a more complete picture of cell damage to be developed.