MCvoxEL

Statistical simulation techniques, called Monte Carlo methods, are applied in human body models in calculation of ionizing radiation doses. Monte Carlo techniques are characterized by the use of random numbers. The program "follows" the particles (photons) trajectories since they leave the source until they are completely absorbed or escape to outside of the model limits.

The distances between interactions are calculated using the cross-sections values for the three studied phenomena: photoelectric effect; Compton scattering and pair production. The interactions are simulated using sampling techniques on mathematical models that describe the physical events. The deposited energy must be computed in the place of each interaction in the interior of the anthropomorphic model.

The first heterogeneous models used were the mathematical simulators:

One mathematical simulator is composed of a set of mathematical expressions which define the volume and position of simple geometric structures (spheres, ellipsoids, cones, etc.), representing the shape and volume of organs and tissues of interest.

The advancement in the acquisition of medical images allied to the fast progress in capability and processing of the computers made possible the construction of a new type of model: the *tomographic *simulators also called *voxel-based *models.

This new technique allowed the construction of new models representing the human body in a much more realistic way. The tomographic simulators are built from the superposition of several transverse images of the modeled body. The resolution of the original images (obtained by computerized tomography, CT, or nuclear magnetic resonance, NMR) defines the horizontal dimensions, and the spacing between images (slice thickness) defines the vertical dimension of each voxel. The information about the position, density and organ/tissue of each voxel are stored in a database which will be the virtual model of the studied body.

In the construction of the models used by the program, the segmentation of organs and tissues is done just coloring the original images.

MCvoxEL program, which calculates the radiation transport, is able to read graphic files. This allows a quite optimized way of write and read geometric data:

[1] Less memory size needed to store information of the virtual model;

[2] All data are stored in Random Access Memory (RAM);

[3] Less processing time (fewer operations and straight access to data).

The program defines the dimensions of each voxel from information about scale used in the acquisition of the original images.

The user informs the number of images and the program builds automatically the model 'reading' the colorful images:

The program calculates the radiation transport using data about geometry (and energy) of the source and probability of interaction in the different tissues of the model. After a large number of simulated photons the program supplies conversion coefficients for dose in each segmented organ or tissue. Several simulations using MCvoxEL were done and results compared with data obtained by other investigators who used modern and well-supported Monte Carlo codes. In all cases, the results lead to the validation of the radiation transport calculation:

The program rebuilds internal images of the model showing regions with determined absorbed energy range defined by the user (black regions in the figure). In this specific simulation an isotropic point source was positioned at x=12, y=12 and z= 17cm coordinates and the black voxels show regions were absorbed energy was higher than one-millionth part of the energy emitted by the source:

The program builds 3-D images with gray tones. Any segmented anatomical structure can be visualized (frontal and lateral views of skeleton of a head-and-neck model):

Virtual X-ray radiographies can be done by MCvoxEL:

Virtual Computed Tomography also is simulated by MCvoxEL: