KAHIBARO
14.4. Dose Actor
Table of Contents
Absorbed dose
In GATE, a dose actor is a special type of actor that calculates the absorbed dose in a region of interest. Absorbed dose is the quantity that tells you how much radiation energy has been deposited per unit mass of material. It is central in radiotherapy, internal dosimetry, and any safety or optimization study.
By definition, the absorbed dose $D$ is
$$
D = \frac{dE}{dm}
$$
where $dE$ is the energy deposited in a small mass $dm$ of material. In the SI system, the unit of absorbed dose is the gray, abbreviated Gy, where
$$
1 \,\text{Gy} = 1 \,\frac{\text{joule}}{\text{kilogram}}.
$$
In Monte Carlo simulations like those done with GATE, energy deposition happens as many small steps along particle tracks. Each step deposits an energy increment $\Delta E_i$ in some part of your geometry. The dose actor collects all these energy deposits inside a user-defined scoring region, and then divides the total energy by the mass of that region or by the mass of smaller subdivisions inside it.
The dose actor is closely related to the energy deposition actor, but it performs an extra normalization by mass and reports results in units of dose instead of energy. You will typically use the energy deposition actor when you are interested in raw energy per volume, and the dose actor when you need quantities directly comparable to clinical or regulatory metrics.
In practice, when you configure a dose actor you specify at least:
The volume where dose is scored, for example a water phantom, a patient CT volume, or a detector.
A discretization of that volume, usually a three dimensional grid of voxels.
The output units and some options like whether to record statistical uncertainty.
Internally, for each voxel the actor accumulates the sum of all deposited energies and the sum of squared deposited energies. After the simulation, it converts the total energy per voxel to dose using the mass of each voxel. Because the geometry and materials are known in GATE, the mass is computed from the voxel volume and the material density.
Key definition
Absorbed dose in a voxel:
$$
D = \frac{\sum_i \Delta E_i}{m_\text{voxel}}
$$
with result typically expressed in gray (Gy).
The dose actor can usually also record the relative statistical uncertainty in each voxel, which is important when you decide whether you have simulated enough particle histories. A large number of events will reduce the uncertainty in the mean dose estimate in each voxel.
Conceptually, the dose actor is independent of the imaging or therapy modality. The same tool is used whether you simulate external photon beams, proton beams in a water phantom, or radionuclide sources distributed in an organ. What changes between applications is the geometry where you attach the actor, the voxel grid configuration, and the interpretation of the resulting dose distribution.
Dose maps
A dose map is the spatial distribution of absorbed dose, usually represented on a three dimensional grid of voxels. The dose actor in GATE is responsible for creating these voxelized dose maps as simulation output.
When you define a dose actor, one of the central tasks is to configure the scoring grid. This grid partitions your scoring volume into voxels with a defined number of cells along $x$, $y$, and $z$, and corresponding physical voxel sizes. For example, you may choose a $200 \times 200 \times 200$ grid with $1 \,\text{mm}^3$ voxels to obtain a fine three dimensional dose map in a water phantom. For patient CT based simulations, the voxel grid is usually matched to the CT image resolution so that dose can be overlaid directly on the anatomical images.
The actor then records for each voxel:
The absorbed dose $D$ in Gy.
Optionally the squared dose or related quantities that allow calculation of statistical uncertainty.
Some implementations also record the number of hits or events that contributed to each voxel.
After the simulation finishes, the dose actor writes the dose map to disk. Output formats in GATE are typically compatible with medical imaging tools, for instance MetaImage (MHD/raw) or NIfTI, so you can open them in software like ITK Snap, 3D Slicer, or in Python with scientific libraries.
The resulting data are a 3D image where each voxel value represents dose. You can think of this as analogous to a CT or PET image, but with units of gray instead of Hounsfield units or activity concentration. This makes it straightforward to create isodose lines, depth dose curves, and dose volume histograms after the simulation.
Important points about dose maps
- Each voxel contains dose $D$ in Gy, computed from deposited energy and voxel mass.
- Voxel size and grid dimensions strongly affect resolution and noise.
- Consistent coordinate systems between geometry, dose actor, and any CT data are essential.
There is a trade off between spatial resolution and statistical noise. Smaller voxels provide finer spatial detail but receive fewer energy deposition events per voxel, which increases the relative statistical uncertainty. Larger voxels smooth out the noise but reduce spatial accuracy. In practical GATE simulations you adjust the number of voxels, the total number of simulated histories, and the source configuration until you reach a suitable balance for your application.
Dose maps are the starting point for many analyses. You can extract profiles along a line, for example to study the Bragg peak position in proton therapy. You can compute planar dose distributions for comparison with measurements in a water tank. In internal dosimetry, you can average the dose over labeled organs, using structures derived from CT segmentations, to obtain organ mean doses.
Because dose maps may be very large datasets, you also need to consider output file size and memory usage, especially for high resolution patient studies. This is another reason to tune the voxel grid, and to focus scoring only on regions where dose information is actually needed.
Views: 11