KAHIBARO
34.5. Beam Geometry
Table of Contents
Source
In external beam radiotherapy simulations the source defines how the treatment beam leaves the accelerator head and enters the treatment field. In GATE you typically do not model every detail of a clinical linac head in a beginner project, but you still need a source that realistically represents the beam that the patient or phantom receives.
For most photon beam treatments the source is placed at the nominal source to axis distance, usually 100 cm from the isocenter. In a simple setup you can define a point source at that location and give it a narrow angular distribution to mimic a collimated beam, or a small disk or rectangle to represent a finite focal spot. For electron beams the source is usually closer to the phantom surface and often has a larger lateral size and divergence.
The key source geometry parameters are position, field size, and divergence. The position is typically defined by the isocenter and the beam direction. In a standard head‑on beam the source lies on the negative z axis, aiming toward the origin, so the beam central axis crosses the center of the phantom. The field size is controlled either by the geometric size of the source or by the collimators. In many radiotherapy simulations you let collimators define the field and keep the source relatively small. Divergence is determined by the finite distance from source to phantom. If the source is at a finite distance, a uniform field at isocenter becomes slightly larger at deeper depths, which is important for realistic dose profiles.
Energy and spatial distribution belong to other chapters, but here you should be aware that the beam geometry is closely linked to these distributions. A typical photon treatment beam is not a perfectly parallel monoenergetic pencil. The central axis receives a slightly different spectrum compared to off axis regions, and the fluence profile is not perfectly flat. When you approximate a clinical beam as a simple geometric source, you should document which geometric aspects are idealized, for example pointlike emission at fixed energy and perfectly symmetric field.
In more advanced simulations, a phase space source is often used. In this approach you record particles exiting a detailed linac head once, and then reuse that recorded distribution as a source in subsequent patient or phantom simulations. Geometrically the phase space plane replaces a simple point or disk source. The phase space plane is usually placed just below the accelerator head collimators and above any patient specific beam shaping. Its position and orientation again define the beam axis and field coverage.
For non‑coplanar or multiple fields, you do not change the source type, but its position and direction. For each field you rotate or translate the source so that the central axis and isocenter match the planned beam geometry. In a gantry rotation, you can keep the phantom fixed and rotate the source and collimators around the isocenter, or equivalently keep the source fixed and rotate the phantom. The geometric relationships are the same, but your choice affects how convenient it is to compare with clinical coordinates.
::danger
In radiotherapy simulations always verify:
- Source to axis distance (for example 100 cm for standard photon beams).
- Central axis direction and intersection with the phantom center.
- Consistency between source position, phantom position, and any rotations.
::danger
Collimators
Collimators define the treatment field by blocking parts of the beam, which makes them central to beam geometry. Even in simplified simulations you should represent them with realistic shapes and positions, because they strongly influence the lateral and penumbra dose distributions.
The basic geometric idea is straightforward. The collimator is a high density volume, such as lead or tungsten, placed between the source and the phantom. Open regions allow particles to pass, while solid regions absorb or scatter them. In GATE you usually create collimators as boxes or composite shapes within the world volume, with the phantom as a child volume placed downstream. The geometric alignment must follow the beam axis so that the open aperture projects the desired field size at isocenter.
Primary collimators define a roughly circular or rectangular large field, but in many treatment beams the main field definition comes from secondary jaws or multileaf collimators, called MLCs. Jaws are usually modeled as thick rectangular blocks that move symmetrically along transverse axes. To obtain a field size of for example 10 cm by 10 cm at isocenter, you place the jaw edges such that the projected opening at the isocenter plane covers 10 cm in each direction. This means the physical jaw separation in centimeters at the jaw plane is larger than 10 cm, since the jaws lie upstream of isocenter and the beam diverges.
MLCs are more complex geometrically because they consist of many individually movable leaves. Each leaf is a long, narrow, dense block whose length runs along the beam direction and whose width and thickness define the leaf pitch and transmission. In GATE you can build one leaf as a box volume, then repeat it many times laterally to form a full bank. The open shape that the MLC creates is defined by which leaves are retracted and which are closed into the beam. All leaves share the same orientation along the beam axis but have different lateral positions.
In a simplified geometry you might not model every mechanical detail like rounded leaf ends or tongue and groove, but you still need the main geometric parameters: leaf width at isocenter, leaf thickness, and leaf span. These determine how well the beam conforms to the target shape and what penumbra you get. Even a coarse MLC model can provide useful insight into field shaping effects if the geometry is positioned correctly relative to the source and phantom.
For electron beams, applicators and cutouts serve a similar geometric role. An applicator is often a square or rectangular metal structure attached to the head, which brings the field defining opening closer to the skin. A cutout is a custom insert placed in the applicator, typically modeled as a sheet of high density material with a patient specific opening. Its placement is again chosen so that the projected opening at the phantom surface matches the clinically desired field.
::danger
For collimator geometry always check:
- The projected field size at isocenter or phantom surface matches the intended treatment field.
- Collimator edges align with the beam central axis and desired coordinates.
- Thick high density regions fully block the beam outside the opening.
::danger
Phantom
The phantom represents the patient geometry that the treatment beam irradiates, and its geometric definition is crucial for realistic dose distributions. In GATE you typically start with a simple water phantom, then move to more complex phantoms or patient images.
A basic water phantom is often a rectangular box large enough to cover the expected high dose region and some surrounding low dose area. For example, a 30 cm by 30 cm by 30 cm cube of water is a common starting point for photon beam depth dose studies. Geometrically, you place the phantom so that its center coincides with the isocenter, or for depth dose curves, so that its surface is at a defined distance from the source. For a beam incident along the positive z direction, you might put the entrance face at z equal to 0 and let the beam cross from air into water at that plane.
Inside the phantom, you may define additional structures, such as bone, lung, or target volumes, as child volumes embedded in the main water volume. Geometrically these are boxes or other shapes placed at specific coordinates that correspond to off axis or different depth locations. The nesting of volumes lets GATE assign different materials and therefore different dose responses at each location. When you design these internal geometries, ensure they do not overlap each other or extend outside the main phantom, because that would create invalid geometry and unreliable results.
In more realistic patient dosimetry you replace the simple geometric phantom with a voxelized one built from CT images, which is discussed in other chapters. From a geometry perspective, the key point is that the voxelized phantom must be aligned carefully with the beam. The isocenter location, the patient orientation, and the couch position all affect where the beam enters and where the high dose region lies. You typically ensure that the isocenter coordinate in GATE matches the isocenter recorded in the treatment plan and that the phantom is placed accordingly in the world.
The relative geometry between phantom and source also determines the depth along the central axis. Depth dose curves are usually measured along the central axis from the entrance surface to deeper regions. If your phantom is shifted incorrectly, your simulated depth coordinate may not correspond to the experimental or clinical one. Similarly, lateral profiles at specific depths depend on the correct position of the central axis relative to the phantom boundaries.
Boundary conditions at the phantom edges matter as well. If the phantom is too small laterally, the beam will suffer excessive lateral leakage and you will underestimate scatter dose compared to an infinite medium. To reduce such edge effects, you make the phantom significantly larger than the primary field. For example, a 10 cm by 10 cm field might be simulated in a 30 cm by 30 cm phantom, not a 12 cm by 12 cm one.
::danger
For phantom geometry verify:
- Entrance surface position relative to the source and central axis.
- Phantom size is large enough to avoid unrealistic edge effects.
- Internal structures are correctly placed and non‑overlapping inside the main phantom.
::danger
Views: 12