Related Experiment Video
Updated: Aug 21, 2025
14:19
A Basic Positron Emission Tomography System Constructed to Locate a Radioactive Source in a Bi-dimensional Space
Published on: February 1, 2016
8.6K
Position estimation using neural networks in semi-monolithic PET detectors.
M Freire1, J Barrio1, N Cucarella1
1Instituto de Instrumentación para Imagen Molecular (I3M), Centro Mixto CSIC - Universitat Politècnica de València, Camino de Vera s/n, E-46022 Valencia, Spain.
Physics in Medicine and Biology
|November 16, 2022
Summary
This study compares crystal treatments and silicon photomultiplier (SiPM) models for total-body positron emission tomography (TB-PET) detectors. Results show promising spatial and energy resolution, indicating suitability for advanced TB-PET scanner development.
Area of Science:
- Nuclear physics
- Medical imaging technology
- Detector physics
Background:
- Total-body positron emission tomography (TB-PET) requires high-resolution detectors for improved imaging.
- Optimizing detector performance involves careful selection of crystal materials, surface treatments, and photodetectors.
- Semi-monolithic detector designs offer potential advantages for TB-PET scanner construction.
Purpose of the Study:
- To experimentally compare the 3D spatial and energy resolution of semi-monolithic detectors for TB-PET scanners.
- To evaluate the impact of different crystal surface treatments (Enhanced Specular Reflector (ESR) with/without black paint and retroreflector) on detector performance.
- To assess the performance of two different silicon photomultiplier (SiPM) models (S13361-3050AE-08 and S14161-3050AS-08).
Main Methods:
- Utilized an array of lutetium yttrium oxyorthosilicate (LYSO) slabs coupled to SiPM arrays.
- Investigated three crystal surface treatments: ESR + Retroreflector (RR) + Black paint (B), ESR + RR, and All ESR.
- Employed a neural network (NN) for estimating impact position in x and depth-of-interaction (DOI) directions, using data acquired with a 22Na source and pinhole collimator.
Main Results:
- Energy resolution varied between 11 ± 1% and 16 ± 1% depending on the SiPM model and crystal treatment.
- Positioning accuracy in the x-direction showed a mean average error of 1.1 ± 0.5 to 1.3 ± 0.5 mm.
- Depth-of-interaction (DOI) resolution achieved a mean average error of 1.7 ± 0.8 to 2.2 ± 1.0 mm, irrespective of the SiPM model.
Conclusions:
- The tested semi-monolithic detectors demonstrate competitive spatial and energy resolution suitable for TB-PET applications.
- Crystal surface treatment significantly influences energy resolution, with the ESR + RR configuration showing the best performance.
- These findings support the use of these detectors in the development of next-generation TB-PET scanners.
