Determination of the absorbed dose in the brain with DICOM images using GATE/GEANT4 and MATLAB

TL;DR

This study refines brain dose calculation using GATE/GEANT4 simulations and MATLAB, integrating PET images to optimize and validate absorbed dose estimates in the cerebral region.

Contribution

It introduces a detailed Monte Carlo simulation approach combined with image processing to accurately determine brain absorbed doses from PET data.

Findings

Accurate absorbed dose estimation in brain tissue achieved.

Simulation results validated against established dosimetry guidelines.

Enhanced dose visualization through integration of DICOM images and simulation data.

Abstract

In this study, the absorbed dose calculation for the cerebral region is refined using the Geant4 software, along with the ROOT and GATE programs. This approach provides a detailed statistical analysis of positron trajectories within a water medium, using a voxel-sized source based on PET imaging. By simulating positron displacement within a 4x4x4 mm3 voxel containing radioactive material mixed with water, and bordered by voxels of identical density, the exact absorbed dose in the brain can be precisely determined. PET images, processed in MATLAB, will complement this dosimetric analysis. Results are compared to established guidelines, such as MIRD Pamphlet 17 and ICRP 106, to validate the accuracy and consistency of the absorbed dose calculations. The primary aim of this research is to optimize the absorbed dose in the cerebral region using digital images from the Philips PET-CT Gemini GXL6 scanner at the Radiotherapy and Nuclear Medicine Service, University Hospital of Caracas. The absorbed dose is calculated with the GATE simulator, a Monte Carlo-based tool supported by the Geant4 and ROOT frameworks. The focus is on evaluating the positron displacement from the central voxel to neighboring voxels. Post-simulation, the program algorithms provide detailed data on absorbed doses for both gamma and beta radiation emitted by 18F. This simulation data is then integrated into PET DICOM images within MATLAB. Geometric calculations are applied to further enhance absorbed dose information by leveraging the voxel characteristics of the images, thus optimizing the absorbed dose depiction. Results will be benchmarked against standards like ICRP 109, MIRD Pamphlet 2, and ICRP 53 for validation and optimization.