KAHIBARO
47.1. Project Overview
Table of Contents
Build a complete PET detector simulation
In this practical example you will build, run, and analyze a complete PET scanner simulation in GATE, starting from an empty world and ending with coincidence data that can be used for image reconstruction or performance studies. The goal is not to create a fully realistic commercial scanner, but a simplified and transparent model that teaches how the main PET components fit together in GATE.
You will construct a ring of detector crystals, place a radioactive F‑18 source in a phantom, configure PET‑specific physics, and model the detector response through the digitizer chain. At the end of the project you will have a set of ROOT files with hits, singles, coincidences, and derived information such as lines of response. This project connects many earlier concepts, including geometry, materials, particle sources, physics lists, actors, and digitizers, into one coherent workflow.
What you will build
The simulated scanner will be a cylindrical PET system composed of identical detector modules arranged in a ring around the patient. Each module will contain an array of scintillation crystals, for example LYSO blocks, which detect the 511 keV annihilation photons produced by positron emission from F‑18. The geometry will be built hierarchically: first the ring, then repeated modules, and finally repeated crystals inside each module. This approach mirrors how real PET systems are designed, and prepares you for more complex scanner layouts in later chapters.
Inside the ring you will place a simple activity distribution such as a uniform cylindrical phantom filled with F‑18 or a point source at the center. The project will focus on typical whole‑body or brain PET configurations, without introducing features like time‑of‑flight or depth‑of‑interaction yet. Those topics will be addressed in later, dedicated chapters.
On the detection side you will simulate the conversion of incoming photons to detector signals using the digitizer chain. You will start from raw energy deposits in the crystals, group them into hits, transform hits into singles that represent individual detected events, and finally sort singles into coincidences. Basic detector effects such as energy resolution and an energy window selection will be applied to approximate a realistic PET detector response.
Learning objectives
By the end of this project you should be able to describe and implement the full PET acquisition chain in GATE, from radioactive decay to coincidence data. You will learn how to define a PET ring geometry using repeated modules and crystals, and how to assign consistent detector identifiers that are required for later data analysis and for constructing lines of response. You will practice configuring an F‑18 source, including its activity and spatial distribution, and understand how this affects counting statistics and simulation time.
On the physics side you will configure a physics list appropriate for PET, including positron transport and gamma interactions that produce the characteristic 511 keV photons and their subsequent interactions in the detector crystals and phantom. You will see how these physics choices influence image quality and quantitative accuracy.
You will also learn to record different types of simulation output. Hits will capture raw interaction information in the crystals, singles will represent processed detector events after digitization, and coincidences will encode pairs of nearly simultaneous detections that correspond to candidate annihilation events. You will generate lines of response from the coincidence data, which are the starting point for PET image reconstruction or sensitivity calculations.
Throughout the project you will see how to balance realism and computational cost by choosing the number of events, acquisition time, and output content. You will also gain experience checking the reasonableness of intermediate results, such as energy spectra of singles and basic coincidence count rates.
How this project connects to other chapters
This project assumes you are comfortable with the basics covered earlier in the course, such as creating a world volume, defining materials like air, water, and LYSO, setting up particle sources, and using actors and digitizers in simpler contexts. Here you will re‑use those skills in a focused PET application.
Chapters on PET Simulation Fundamentals provide the physical background, such as positron emission, annihilation photons, PET detector geometry, and detector response. In this project you put that theory into practice by implementing a specific scanner layout and source in code.
Later chapters on PET Coincidence Analysis, Time‑of‑Flight PET, and Python or ROOT data analysis will extend what you build here. The PET coincidence data that you generate in this project will be a natural input for those advanced analyses. Similarly, concepts from multithreading, performance optimization, and simulation validation can all be applied to refine and extend this PET example once you are comfortable with the basic workflow.
Throughout this project, you will move step by step: first defining the PET ring and detector modules, then adding crystals and an F‑18 source, configuring PET physics, attaching actors for hits and singles, applying energy blurring and an energy window, sorting coincidences, identifying true and scattered events, building lines of response, and finally saving and analyzing the PET data using Python or ROOT. The individual steps appear as separate sections in this part of the course, but together they form a single, complete PET detector simulation that you can adapt to your own research or teaching needs.
In this project always keep the following in mind:
- Use a PET‑appropriate physics list that includes positron transport and detailed gamma interactions.
- Make sure detector materials, crystal dimensions, and ring radius are physically reasonable and consistent.
- Configure the F‑18 activity and number of events so that statistical uncertainty is acceptable for your goals.
- Apply realistic energy blurring and a suitable PET energy window around 511 keV before sorting coincidences.
- Record enough information in your output (hits, singles, coincidences) to allow later analysis, but avoid unnecessary data that slows the simulation.
By carefully following the steps in this project and revisiting these rules, you will gain a practical and coherent understanding of how PET simulations are constructed in GATE and how they can be used for studying scanner performance, reconstruction algorithms, or protocol optimization.
Views: 10