KAHIBARO
35.3. Water Phantom
Table of Contents
Phantom geometry
In proton therapy simulation, a water phantom is the standard reference object for studying depth dose and validating beam modeling. In GATE you represent the phantom as a regular volume made of liquid water, placed inside the world and aligned with the beam axis.
At the simplest level, you create one box volume that is large enough to contain the entire proton range plus some margin. The phantom dimensions along the beam direction must extend well beyond the expected Bragg peak, so that you capture the full depth dose curve and the distal falloff. For example, for clinical proton energies you might choose a few tens of centimeters along the beam axis and several centimeters laterally so that the lateral scatter remains inside the phantom. The exact size depends on the energies you want to simulate and any comparison data you have.
Within your GATE simulation, you typically define the world first, then add a box volume whose parent is the world and whose material is set to water from the Geant4 material database. You then position this water box so that the proton beam enters one face of the phantom orthogonally, with the beam axis aligned with one coordinate axis, usually the $z$ axis. Keeping the beam axis and the phantom faces aligned with the coordinate axes simplifies later depth dose analysis, because depth can be read directly as the $z$ coordinate of energy deposition or dose.
The phantom thickness along $z$ controls the maximum depth you can study. Its size in $x$ and $y$ should be large enough that the primary beam and most of the scattered protons remain inside the water for the depths of interest. If the phantom is too small in transverse directions, protons can leave the phantom and the simulated dose near the edges will not represent an infinite medium. For central-axis depth dose analysis, you usually place the proton beam so that it is centered in $x$ and $y$, entering at one face of the water box. In more advanced setups you can embed this water phantom inside other structures, but for proton physics studies and Bragg peak validation a single homogeneous water box is often sufficient.
The material choice is critical. You should use a standard Geant4 water material, for example G4_WATER or an equivalent NIST water material, rather than creating a custom material, because proton stopping powers and ranges have been extensively benchmarked for these standard definitions. This improves comparability with published range data and reference depth dose curves.
For accurate proton depth dose and Bragg peak studies, always use a homogeneous water volume that is thick enough to contain the full proton range along the beam axis and wide enough laterally to avoid significant edge escape of protons.
Dose scoring
Once the water phantom geometry is in place, you need a way to record how much energy the protons deposit as a function of depth inside the water. In GATE this is done with a dose actor that scores dose in a voxelized volume. The actor divides the phantom into a regular 3D grid of voxels and accumulates dose in each voxel over all simulated events.
To study proton therapy in water you typically configure the dose actor so that its voxel grid matches the water phantom or at least covers the region where the dose is of interest. Along the beam direction you want relatively fine sampling of depth in order to resolve the Bragg peak and the distal falloff. For example, you might choose voxel sizes of a few millimeters or smaller in the depth direction, while accepting coarser resolution laterally if you are mainly interested in central-axis depth dose. The voxel grid origin and spacing should be aligned with the coordinate system you used for the phantom, so that the $z$ index of a voxel corresponds directly to a physical depth in water.
The actor must be attached to the water phantom volume. In practice, you specify the phantom volume name as the target of the dose actor so that only energy depositions within this water volume are scored. GATE will then compute absorbed dose in each voxel using the relation
$$
D = \frac{E_{\text{dep}}}{m},
$$
where $E_{\text{dep}}$ is the sum of energy deposited inside the voxel and $m$ is the voxel mass. The actor outputs a 3D image of dose values, often in units of gray. Depending on your configuration, GATE can also store the statistical uncertainty of the dose in each voxel, usually as a relative standard deviation.
For proton beam studies you usually analyze the 3D dose data by extracting a 1D depth dose profile along the beam axis, often at the center of the phantom in $x$ and $y$. This profile shows the entrance dose, the build-up, the Bragg peak, and the distal falloff. If you configured the dose actor with sufficiently small voxels in depth, the peak position and shape will be well resolved. You can then compare the simulated depth dose with analytical range calculations or with measured data.
When choosing the dose grid resolution, there is a trade-off between spatial accuracy and computation time. Smaller voxels provide finer spatial detail of the Bragg peak and the lateral dose spread, but also increase the number of voxels and therefore the memory usage and output file size. For an introductory water phantom simulation, it is common to start with a moderate resolution, verify that the Bragg peak position is correctly reproduced, and then refine the voxel size if more detailed structure is needed.
Absorbed dose in a voxel is defined as $D = E_{\text{dep}} / m$. To obtain accurate proton depth dose curves, use a dose actor that is attached to the water phantom, align its voxel grid with the beam axis, and choose a voxel size fine enough in depth to resolve the Bragg peak.
Views: 11