Determination of the Interaction Position of Gamma Photons in Monolithic Scintillators Using Neural Network Fitting

Conde, P.; Iborra, A.; González, Antonio J.; Hernández, L.; Bellido, P.; Moliner, L.; Rigla, J. P.; Rodríguez-Álvarez, M. J.; Sánchez, F.; Seimetz, M.; Soriano, Antonio; Vidal, Luis; Benlloch, J.

doi:10.1109/tns.2016.2515163

Cited by 21 publications

(20 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…An alternative approach is to fit an analytic model of the optical photon distribution using an optimisation algorithm which minimises the mean squared error between the observed distribution and the analytic prediction 65,66 . Where an exact analytic model is difficult or impractical to derive due to complex optical properties of the material, neural networks can be trained to automatically account for non-linearities in the relationship between the observed photon distribution and the point of interaction, particularly near slab boundaries 67 . However, the training procedure would need to be repeated if the material or slab dimensions change.…”

Section: Related Workmentioning

confidence: 99%

“…A similar outcome will result from Compton interactions where the scattered photon escapes from the scintillator slab; in this case, the number of photons emitted at the point of interaction will be proportional to the difference in energy between the incident and scattered photon. These two cases will be the easiest to localise due to the simplicity of the interaction; the scintillation event may be treated as an isotropic point source 66,67 . In a nanocomposite or ceramic garnet scintillator, an attenuative factor is included to account for the imperfect transparency of the material 68 .…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Optimisation of monolithic nanocomposite and transparent ceramic scintillation detectors for positron emission tomography

Wilson

Alabd

Safavi-Naeini

et al. 2020

Sci Rep

View full text Add to dashboard Cite

High-resolution arrays of discrete monocrystalline scintillators used for gamma photon coincidence detection in pet are costly and complex to fabricate, and exhibit intrinsically non-uniform sensitivity with respect to emission angle. nanocomposites and transparent ceramics are two alternative classes of scintillator materials which can be formed into large monolithic structures, and which, when coupled to optical photodetector arrays, may offer a pathway to low cost, high-sensitivity, high-resolution pet. However, due to their high optical attenuation and scattering relative to monocrystalline scintillators, these materials exhibit an inherent trade-off between detection sensitivity and the number of scintillation photons which reach the optical photodetectors. in this work, a method for optimising scintillator thickness to maximise the probability of locating the point of interaction of 511 keV photons in a monolithic scintillator within a specified error bound is proposed and evaluated for five nanocomposite materials (LaBr 3 :ce-polystyrene, Gd 2 o 3-polyvinyl toluene, LaF 3 :ce-polystyrene, LaF 3 :Ce-oleic acid and YAG:Ce-polystyrene) and four ceramics (GAGG:Ce, GLuGAG:Ce, GYGAG:Ce and LuAG:Pr). LaF 3 :Ce-polystyrene and GLuGAG:Ce were the best-performing nanocomposite and ceramic materials, respectively, with maximum sensitivities of 48.8% and 67.8% for 5 mm localisation accuracy with scintillator thicknesses of 42.6 mm and 27.5 mm, respectively. Detection of high-energy photons for positron emission tomography (PET) is inherently challenging due to their highly penetrating nature. Solid-state detectors are unable to directly detect such photons with high efficiency, due to the limited detector thickness (typically <1 mm) and low density and effective atomic number of most semiconductors 1,2. To increase sensitivity to high energy photons, in most applications it is necessary to optically couple a photodetector to a scintillator. When a high-energy photon deposits energy in the scintillator, the energy is absorbed by dopant atoms (commonly cerium, europium or thallium) and re-radiated as multiple lower-energy photons-typically in the optical range-which can be detected by a solid-state photodetector or photomultiplier tube 3. Many excellent scintillator materials are now available, including some with good sensitivity to high energy photons due to their high density and effective atomic number 3,4. The best of these also provide high light output, good energy resolution and fast decay time, and scintillate at wavelengths which are compatible with semiconductor photodetectors 3-5. Conventionally, high spatial resolution PET requires the use of a large number of very small, discrete, optically isolated scintillator crystals, either individually coupled to a photodetector cell or multiplexed via some sort of light-sharing scheme. High gamma photon sensitivity requires a large radial scintillator thickness, while maintaining a uniformly high spatial resolution across the field of view requires very narrow cryst...

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Optimisation of monolithic nanocomposite and transparent ceramic scintillation detectors for positron emission tomography

Wilson

Alabd

Safavi-Naeini

et al. 2020

Sci Rep

View full text Add to dashboard Cite

show abstract

“…Statistical approaches rely on training data consisting of measured light distributions with known reference positions. Several statistical methods have been presented in literature, e.g., k-nearest neighbors (kNNs) [24], neural networks [8], [25], [26], maximum likelihood [27], Voronoi This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ diagrams [28], and gradient tree boosting [11], [14].…”

Section: Introductionmentioning

confidence: 99%

“…An alternative approach is dedicated setups shaping a gamma beam usually consisting of a material with high atomic mass (e.g., tungsten or lead). First, parallel hole collimators with precise drilling of known diameter ranging from 0.5 mm to 1.0 mm were employed for PET detector calibration [9], [24], [25], [27], [31]. The thickness of the shielding material and shape of the parallel hole determine the beam profile.…”

Section: Introductionmentioning

confidence: 99%

Characterization and Simulation of an Adaptable Fan-Beam Collimator for Fast Calibration of Radiation Detectors for PET

Hetzel

Mueller

Grahe

et al. 2020

IEEE Trans. Radiat. Plasma Med. Sci.

View full text Add to dashboard Cite

Monolithic scintillators for positron emission tomography systems perform best when calibrated individually. We present a fan-beam collimator with which a crystal can be calibrated within less than 1 h when suitable positioning algorithms are applied. The collimator is manufactured from lead, features an easily adaptable slit to tune the beam width and can be operated together with a coincidence detector to select a clean sample of 511-keV annihilation photons. We evaluated the performance of the collimator with a Geant4 simulation for slit widths of 0.25 mm, 0.4 mm, and 1 mm and validated the shape of the beam profile experimentally by step-wisely moving a detector into the beam. This shows a clear narrow and box-shaped beam profile even if the collimator is operated without the coincidence setup. In the latter configuration, the fraction of gammas in the beam region on a 50×50 mm 2 large detector is between 48% and 79% which is improved significantly to more than 94% by using only coincidence events. Analyzing the energy distribution shows that the fraction of 511-keV photons is increased from less than 50% to more than 96% by selecting coincidences.

show abstract

“…Ga/ 68 Ge: El emisor de positrones 68 Ga, con un periodo de semidesintegración de 68 minutos, se produce del radioisótopo padre 68 Ge, cuyo periodo de semidesintegración es de 270.8 días. El 68 Ge decae el 100 % de las veces por captura electrónica al 68 Ga y puede producirse, bien sea mediante bombardeo con protones de 40 MeV a un blanco de galio siguiendo la reacciones 69 Ga(p,2n) 68 Ge y 71 Ga(p,4n) 68 Ge, o a través del bombardeo con partículas α a un blanco de zinc, produciéndose la reacción 66 Zn(α,2n) 68 Ge, o mediante una reacción de espalación producida por protones de altas energías que impactan sobre un blanco de molibdeno. En el generador, el 68 Ga se extrae mediante un oxido de titanio (SnO 2 ) o dioxido de titanio (TiO 2 ).…”

Section: Generador 68unclassified

Algoritmo de reconstrucción analítico para el escáner basado en cristales monolíticos MINDView

Góez¹

View full text Add to dashboard Cite

Motivación y antecedentes 3 1.2. Objetivos y organización de la tesis 5 2 Introducción 7 2.1. Radiactividad 8 2.1.1. Ley de decaimiento radiactivo 9 2.1.2. Decaimiento β + 10 2.2. Fundamentos físicos de la física médica. 11 2.2.1. Sección eficaz 12 2.2.2. Dispersión Rayleigh o dispersión coherente 13 2.2.3. Efecto fotoeléctrico 13 III IVÍNDICE GENERAL 2.2.4. Efecto Compton o dispersión no coherente 2.2.5. Producción de pares 2.2.6. Coeficiente de atenuación 2.2.7. Aniquilación electrón-positrón 2.2.8. Centello del cristal 2.3. Aspectos generales del PET: 2.3.1. Funcionamiento de la tomografía por emisión de positrones 2.3.2. Clasificación de los eventos 2.3.3. Corrección por tiempo de vuelo 2.4. Aspectos Logísticos del PET: Producción de radioisótopos 2.4.1. Ciclotrón 2.4.2. Generadores 2.5. Técnicas de imagen híbridas 2.5.1. Imagen híbrida de PET/CT 2.5.2. Imagen híbrida de PET/MRI PARTE II MATERIALES Y MÉTODOS 3 Escáner MINDView 3.1. Introducción 3.2. Escáner MINDView 3.2.1. Geometría del escáner MINDView 3.2.2. Bloque detector 3.2.3. Fotomultiplicadores de silicio SiPM 3.2.4. Determinación de las coordenadas planares x e y 3.2.5. Simulaciones de la respuesta del detector del escáner MINDView y simulaciones Monte Carlo en GATE de fuentes puntuales 4 Algoritmos de reconstrucción analíticos 4.1. Introducción 4.2. Formación de la imagen 4.2.1. Aproximación de la integral de línea 4.3. Imagen 2-dimensional 4.3.1. Teorema del corte central de Fourier en dos dimensiones 4.3.2. Suficiencia de datos 4.3.3. Almacenamiento de las coincidencias en modo 2D 4.4. Imagen Tridimensional Í NDICE GENERAL V 4.4.1. Teorema de la sección central tridimensional 4.4.2. Suficiencia de datos 4.4.3. Almacenamiento de las coincidencias en modo 3D: Michelogramas 4.5. Reconstrucción analítica Bidimensional 4.5.1. Retroproyección 4.5.2. Reconstrucción FBP2D 4.6. Reconstrucción analítica Tridimensional 4.6.1. Varianza espacial de la PSF 4.6.2. Retroproyección 3D 4.6.3. Reconstrucción FBP3D 4.7. Algoritmo de retroproyección filtrada a posteriori -BPF 4.7.1. Formato de almacenamiento en modo lista 4.7.2. Algoritmo de retroproyección de convolución 4.7.3. Algoritmo de retroproyección filtrada a posteriori en 3D PARTE III RESULTADOS 5 Adaptación de la librería STIR y propuesta de un algoritmo de BPF105 5.1. Introducción 5.2. Adaptación de la librería STIR a escáneres basados en cristales monolíticos 5.2.1. Geometría del escáner ALBIRA de pequeños animales. 5.2.2. Procedimiento de adaptación de geometrías de anillos abiertos a la geometría de anillo cerrado en STIR 5.2.3. Módulo de conversión a coordenadas locales 5.2.4. Conversión de un sistema monolítico de detección a un sistema basado en cristales pixelados 5.2.5. Módulo de conversión de formato modo lista a michelogramas115 5.2.6. Tamaño de los michelogramas 5.2.7. Tiempo de reconstrucción 5.3. Resultados de reconstrucciones en STIR 5.3.1. Resultado obtenidos con el escáner ALBIRA para pequeños animales 5.3.2. Resultados obtenidos en el escáner MINDView dedicado a cerebro 5.4. Algoritmo de r...

show abstract

Determination of the Interaction Position of Gamma Photons in Monolithic Scintillators Using Neural Network Fitting

Cited by 21 publications

References 18 publications

Optimisation of monolithic nanocomposite and transparent ceramic scintillation detectors for positron emission tomography

Optimisation of monolithic nanocomposite and transparent ceramic scintillation detectors for positron emission tomography

Characterization and Simulation of an Adaptable Fan-Beam Collimator for Fast Calibration of Radiation Detectors for PET

Algoritmo de reconstrucción analítico para el escáner basado en cristales monolíticos MINDView

Contact Info

Product

Resources

About