KAHIBARO
26.5. PET Physics Configuration
Table of Contents
Positron transport
In a PET simulation you must ensure that the physics list correctly transports positrons from their emission point until they slow down and annihilate. This part of the configuration decides how realistic the positron range and the annihilation position are, which directly affects spatial resolution and blurring in PET images.
In GATE, PET simulations normally reuse general electromagnetic physics lists from Geant4, for example G4EmStandardPhysics or the low energy variants. You do not build these models by hand. Instead you choose and configure them in your Python script. A typical OpenGATE style configuration selects an electromagnetic physics list that includes:
- Continuous energy loss for positrons through ionization and excitation.
- Multiple scattering that changes particle direction while slowing down.
- Bremsstrahlung for high energy positrons.
- Annihilation at rest once the positron is thermalized.
You normally do not have to explicitly enable annihilation processes. All standard electromagnetic lists include positron annihilation, and GATE uses these lists as building blocks. What you do control is which overall physics list is used and what production cuts are applied, which is covered in detail in the chapters about physics lists and production cuts.
For PET, there is a specific goal. You want to model the full positron track until annihilation to obtain a realistic annihilation distribution, but you do not want unnecessary secondary particles that waste computation. The minimum configuration for a realistic PET simulation must ensure that:
Positron transport includes ionization losses, multiple scattering, and annihilation at rest in the patient or phantom materials.
If you remove positron transport or replace the positron by a gamma source at 511 keV, the positron range will be absent, and the intrinsic PET blur will be underestimated. That kind of shortcut is sometimes useful for simplified tests but is not acceptable when you study spatial resolution, reconstruction, or quantitative performance.
A typical workflow in GATE is to select a physics list that already covers positrons correctly, such as a general electromagnetic list recommended for medical applications, and then verify that positrons are indeed produced from the radionuclide source and tracked. You can use actors to check the spatial distribution of annihilation points or to record positron tracks. If the physics list is misconfigured and annihilation is not included, you will see missing 511 keV photons and an incorrect singles energy spectrum.
The numerical details of positron range depend on the chosen radionuclide and material. For example, F‑18 has a relatively short mean positron range in water, while higher energy positron emitters such as Rb‑82 produce longer ranges. With a correctly configured positron transport, these differences appear naturally in your simulation without additional adjustments, because the physics models handle the energy loss and scattering automatically.
Production cuts also affect positron transport. If cuts are too coarse in regions where positrons slow down, very small secondary particles might be suppressed, which slightly changes the detailed track structure. For typical PET image level studies, standard cut values are sufficient, but for very high precision dosimetry with positron emitters you may need finer cuts in the phantom region.
In summary, PET physics configuration for positron transport consists of choosing a suitable electromagnetic physics list that includes all positron processes and then confirming with simple checks that annihilation and positron range behave as expected for the radionuclides and materials you are using.
Gamma interactions
After annihilation, two 511 keV photons are created and travel through the patient or phantom and then into the detector. The way these photons interact is entirely governed by your electromagnetic gamma physics configuration. This part of the PET physics configuration must correctly reproduce attenuation, scatter, and energy deposition at 511 keV in both tissue and detector materials.
In GATE, the gamma physics configuration again relies on a Geant4 electromagnetic physics list. For PET, the important processes for photons around 511 keV are:
- Photoelectric effect in high Z materials such as scintillation crystals and lead.
- Compton scattering in both tissue and detector materials.
- Rayleigh scattering, which is less dominant at 511 keV but may be included by some physics lists.
- Pair production, which has a threshold at $1.022$ MeV, so it does not occur for pure 511 keV annihilation photons, but it can still matter for higher energy photons or for some radionuclides with additional gamma emissions.
The balance between photoelectric absorption and Compton scatter in the detector controls the shape of the singles energy spectrum. For a correctly configured PET physics list, you will see a clear 511 keV photopeak, a Compton continuum extending from low energies up to near 511 keV, and a backscatter peak if you include surrounding materials. These features are used to define PET energy windows and to separate true events from scattered events.
For PET images in tissue, Compton scattering dominates at 511 keV, which creates scattered coincidences and reduces contrast. That effect should appear naturally if your gamma physics are configured properly. The fraction of scattered coincidences in your simulated data is often used as a performance metric, and you will compare it with experimental values or literature to validate your configuration.
The key configuration choice is thus the selection of an electromagnetic physics list that treats 511 keV photons accurately in the materials you use. For PET this usually means:
Gamma physics must include at least photoelectric effect and Compton scattering for 511 keV photons in both patient and detector materials.
Low energy extensions of the electromagnetic models are sometimes used in SPECT or diagnostic x ray simulations because they improve accuracy at tens of keV. For PET at 511 keV, standard electromagnetic physics are usually sufficient, but low energy models remain compatible if you also simulate non annihilation gammas with lower energies.
Another aspect of gamma physics in PET is the modeling of attenuation in the patient or phantom. The linear attenuation coefficient at 511 keV for different tissues determines how many annihilation photons reach the detector and how many are absorbed or scattered. In GATE this attenuation arises automatically from the cross sections in the physics list and the material definitions. To check that the model is reasonable, you can simulate a simple water phantom and compare the measured attenuation with the analytical law
$$
I = I_0 \, e^{-\mu x}
$$
where $I_0$ is the incident photon fluence, $x$ is the path length in the material, and $\mu$ is the linear attenuation coefficient at 511 keV. If your simulated attenuation deviates strongly from the expected value, you should first review your material definitions, then the chosen electromagnetic physics list.
Gamma interactions in the detector material are equally crucial for realistic energy response. The combination of physics and detector material, such as LYSO or BGO, defines the detection efficiency for 511 keV photons and the probability that events deposit energy in the photopeak window. Later, the digitizer configuration will add blurring and thresholds on top of this physical energy deposition, but the starting point must be a correct description of how gammas interact.
Through the choice of physics list, GATE provides a consistent description of gamma interactions from the annihilation point to the detector. In a PET physics configuration, you typically select a well tested electromagnetic list for medical imaging, verify attenuation in simple phantoms, and then inspect the singles energy spectrum in a basic detector setup. These checks confirm that gamma interactions are handled properly, which is essential before moving on to more advanced PET system simulations and reconstructions.
Views: 10