HUP0900574A2 - Detector modul for pet as well as pet equipped with the same - Google Patents

Description

(Figure 1)

PUBLICATION

COPY

7020R ^-9005'4

Detector module for a positron emission tomograph, and a positron emission tomograph having such a detector module

The invention relates to a detector module for a positron emission tomograph, and to a positron emission tomograph having such a detector module.

Positron emission tomography (PET) was developed in 1969. For examination with a PET device, substances are introduced into the human body that emit positrons (β ⁺ ) (e.g. ¹⁸ FDG). The mean free path of β ⁺ particles in body tissues is about 1 mm, i.e. on average they encounter an electron within this distance, resulting in annihilation. During this process, the electron-positron pair disappears, and instead of them, 2 511 keV gamma photons with energy corresponding to their mass are generated. According to the law of conservation of momentum, these two photons travel in opposite directions. The γ-photons are detected in scintillation detectors placed around the patient. Since the photons are generated at the same time, measuring the time difference between the impacts (coincidence) allows the identification of related photon pairs. In the ring-shaped detector system, the locations of the impacts are also determined, resulting in a straight line connecting the points of impact of the photon pair. Based on the knowledge of a large number of such lines, the location of gamma flashes in the patient can be determined. This is the basis for the operation of PET devices.

The most important technical characteristics of PET devices are sensitivity, energy resolution, temporal resolution (coincidence), photoelectric ratio and spatial resolution. The spatial resolution of PET devices is significantly limited by the fact that the recorded image consists of random discrete events of gamma flashes, which makes the image noisy. The quality of the imaging (contrast, signal-to-noise ratio) is significantly improved by the TOF-PET (time of flight) technique. In these, not only the coincidence of gamma impacts is examined, but also the time difference between the photon pair impacts is measured. With this, it is theoretically possible to determine the position of the gamma flash.

-2 PET scanners usually consist of pixelated scintillation detectors. The detectors are arranged in parallel rings, and the resulting cylindrical envelope surrounds the patient (or a part of their body, usually the head) 360°.

Scintillation detectors are usually composed of elongated scintillator elements with small cross-sections. These pixels are small in the axial and trans-axial directions to achieve good spatial resolution and much longer in the radial direction to absorb as many gamma photons as possible. The design of the scintillation detector in PET devices significantly affects the spatial resolution of the device.

The scintillation detectors of PET devices detect 511 keV gamma photon pairs resulting from positron annihilation. The scintillation material is usually a single crystal (NaI, BGO, GSO, LSO, YSO, LYSO, in TOF-PET: LYSO, BaF ₂ , LaBr ₃ ), in which the entire energy of the incident gamma photon is directly transferred to an electron during a photoeffect. This electron excites the scintillator atoms over a path of about 0.5 mm and releases its energy in the form of UV photons. This energy must be detected by a light detector. The metal detector used must be very sensitive and extremely fast, which is why photomultipliers (PMTs) are often used in PETs. In order to improve the resolution, avalanche photodiodes (APDs) with a much smaller surface area than PMTs are often used, but they are also less sensitive and noisier. The latest developments are aimed at bulk silicon detectors, where the detector material is silicon itself.

Scintillation detectors, as shown in Figure 1, generally comprise an array of scintillator elements 10 and a light guide 16 that transmits the light to light detectors 20. The scintillator elements 10 have an input end 12 and an output end 14, and opposite the output ends 14 of the scintillator elements 10 is an array of light detectors 20 that detect the light emitted from the scintillator elements 10. The scintillator elements are small in axial-transaxial size (typically 1-6 mm), but are radially longer (about 20 mm). The individual pixels are separated by a material that does not allow UV radiation to pass from one pixel to another, but reflects it, so the effective location of the scintillation is always as if it were at the end of the pixel. The

In Figure 3, the incoming arrows show the location of the scintillation, and the outgoing arrows show the UV light rays traveling towards the photodetectors 20.

One of the main classification aspects of scintillation detectors is how the scintillation material is coupled to the photodetector. Based on this, three categories can be established: a) using a light guide, the light from a scintillator pixel is distributed to multiple photodetectors, b) without using a light guide, the light from multiple pixels is directly guided to a position-sensitive detector, c) without a light guide, the light from each pixel is directly guided to a photodetector individually.

The price of detectors increases according to a) b) c). Since solution a), also called light splitting, can be produced at the lowest cost, it seemed appropriate to use this variant as the basis for our developments.

In one of the important practical applications of the light-splitting technique, the scintillation detector is assembled from detector modules (block detectors, detector blocks). A schematic drawing of such a detector module is shown in Figure 2. In these solutions, a given number, according to the figure e.g. 8 x 8 pcs. 10 scintillator elements are connected to a given number, usually 2 x 2 pcs. 20 light detectors, preferably PMTs. The 20 light detectors together define a detection surface 22 (marked with hatching). The light guide 16 is interrupted at the boundary of such a block by a suitable light-reflecting layer, 24 reflector. An important advantage of the detector module is that since the processing of a scintillation event requires a small number of 20 light detectors, during the processing time (so-called dead time) the other detector modules can receive new scintillations, which leads to a higher load capacity of the device. In addition, the use of detector modules is also advantageous due to simpler data processing, simpler construction and efficient serviceability. However, a disadvantage of the modular, block solution is that there is a relatively large dead area between the light detectors, where there is no photon detection. This dead area significantly degrades the spatial resolution and light efficiency of the detector in state-of-the-art PET devices.

-4Scintillation position determination actually means the identification of the given crystal pixel. Various solutions are known to make this as reliable as possible. In one known solution, a checkerboard matrix arrangement is used, where scintillators made of two types of materials (e.g. GSO, LSO) are used. In this way, the pixels can be identified more precisely based on the different photopeak locations of the materials and the time course of the scintillation pulse. In another known solution, notches are made in the light guide, into which the separating layer between the pixels extends. In this way, the width of the light cone starting from the pixels can be somewhat reduced, which leads to more accurate pixel localization. In another known detector module, the light guide is notched to different degrees, which can influence the size of the light cones arriving from each pixel to the light detector. There are also known solutions where separating layers with different reflection properties are used between the pixels, or where the side surfaces of the pixels are painted in different colors to help with pixel localization. In other solutions, reflector foils of different heights are placed between the pixels to improve pixel localization.

Development solutions related to PÉT detectors are described, for example, in US 4531058, US 4733083, US 4749863, US 4929835, US 5091650, US 5300782, US 5329124, US 5453623, US 6087663, US 6462341 Bl, US 2004/0021082 A1 and US 2007/0221856 A1.

The disadvantage of known solutions aimed at improving the spatial resolution of PÉT detector modules is that they are complicated and can only be implemented at a very high cost. Another disadvantage is that the known solutions cannot eliminate the resolution reduction or sensitivity reduction caused by the dead spaces of the light detectors. Another disadvantage is that the increase in spatial resolution is usually achieved at the expense of sensitivity. In the known solutions, the output signal of the detector module is nonlinearly related to the position of the flashing scintillator element, so complex calculations are also required for signal processing.

One of the aims of the invention is to create a detector module for a positron emission tomograph that is free from the disadvantages of the prior art solutions. Another aim was to create a detector module, or a PET using such a module,

-5which enables high spatial resolution, preferably with linear position sensing characteristics, while the light efficiency and sensitivity do not change in an unfavorable direction.

The object of the invention is achieved by the detector module according to claim 1 and the positron emission tomography according to claim 13. Preferred embodiments of the invention are defined in the subclaims.

Preferred embodiments of the invention are described hereinafter with reference to the drawings, in which:

Figure 1 is a schematic drawing of a detail of a prior art PET detector,

Figure 2 is a schematic diagram of a detector block viewed from the sensing direction,

Figure 3 is a representation of the detection bias occurring in a prior art detector block,

Figure 4 is a diagram representing the phenomenon shown in Figure 3,

Figure 5 is an explanatory diagram of the perceived center of gravity of a scintillation spot, the

Figure 6 is a diagram of the detection bias calculated for the detector module corresponding to Figure 2,

Figure 7. Diagram showing the random error of the measured coordinate of the centroid of a scintillation spot,

Figure 8 is a diagram of the distance of the resolvable points as a function of the scintillation position,

Figure 9 is a representation of the light cone reaching the sensing surface from the scintillator element in a prior art arrangement,

Figure 10 shows the irradiation of the crystal needle according to Figure 9 in the sensing plane,

Figure 11. Illustration of the light cone reaching the sensing surface from a scintillator element using an air-gap light deflector,

Figure 12. Irradiation of the crystal needle according to Figure 12 in the sensing plane, the

Figure 13. Enlarged drawing of the air-gap light deflector used in the solution according to Figure 1,

Figure 14 is a representation of the light cone produced by a light deflecting device used in a detector module according to the invention,

Figure 15 shows the irradiation of the arrangement according to Figure 14 in the detection plane,

Figure 16 is an enlarged drawing of the light deflecting device used in the solution according to Figure 14,

Figure 17 is a representation of the corrected detection characteristic in a detector module according to the invention,

Figure 18 is a diagram representing the corrected detection characteristic of Figure 17, Figure 19 is a diagram of another preferred embodiment of the detector module according to the invention,

Figure 20 shows the amount of light deflection angle required for the linearization of the sensing function shown in Figure 6 as a function of position,

Figure 21 shows the measurement characteristics obtained for the arrangement according to Figure 2 for the light deflection according to Figure 20,

Figure 22 is a diagram illustrating the random error in the measured coordinate of the centroid of the scintillation spot in the case of compensation according to Figure 20,

Figure 23 shows the spatial resolution as a function of the scintillation position in the case of compensation according to Figure 20,

Figures 24-26 are schematic representations of embodiments also provided with an intensity modulation device, the

Figure 27 is a diagram of the relative light efficiency resulting from the intensity modulation device according to Figures 24-26 as a function of the internal opening angle of the light cone,

Figure 28 is a diagram modeling the path of a light beam in a scintillation crystal polished on its surface and diffusely reflecting at the entrance end, and

Figure 29. A diagram modeling the path of a light ray in a scintillation crystal polished on all sides

For simplicity and frequent practical application, the invention will be described in the following for the detector module shown in Figure 2, using four 20 light detectors, preferably PMTs. (In the case of a block detector containing several PMTs, the part of it containing any four selected light detectors can be discussed similarly.) From the point of view of spatial resolution, the use of a small number of PMTs is a disadvantage: the real (x) and measured (ξ) positions of the scintillation in front of the detector module are related to each other in a nonlinear ξ = f(x) function, and this effect is so significant that the positions of scintillators coming from neighboring crystal needles from the center of the PMTs towards the edge of the block can no longer be distinguished from each other. The direction tangent of the tangent of the function f(x) is now a constant zero. Figure 3 shows a representation of such a detection distortion occurring in a detector block according to the state of the art, and Figure 4 shows a diagram representing the phenomenon according to Figure 3. Below the detection positions shown in Figure 3, a vertical line indicates the true position of the scintillation, and an X indicates the measured position determined as the center of gravity of the resulting 30 light spots.

The explanation of this phenomenon is discussed in connection with Figure 5. For the sake of simplicity, we calculate with square 20 light detectors and homogeneous 30 light spots. The center of gravity of the 30 light spots resulting from scintillation in the plane of the detectors can be determined based on the amount of light power falling on the four 20 light detectors. According to the calculation of the first moment, the x-coordinate (ξ) of the center of gravity of the light spot coming from the x-position can be determined based on equation (1) (similarly to y).

ξ - ^a' ^x a ⁺ Pb ' ^x b ⁺ ' ^x c ⁺ Pp' ^x d _ Pa ' ^x a ⁺ Pb ' ^x b ⁺ Pc ' ^x c ⁺ Pp ' ^x dz, x P _A +P _B +P _C +P _D p ' where Pa,b,c,da are the light power measured by the detectors, and xa.bcd is the location coordinate of their center. If the coordinates of the detectors are the same (xb = xd ⁼ -Xa ^{= -} xc = ^x det), the equation can be transformed into the following form:

_? _-Λ+Λ-^+Λ> .. .(^+^)-(^+^), ,,,

P ^A det — Xáet

If the x-position of the light spot changes by Δχ, the measured position will also be different (ξ+Δξ), since different light powers fall on the detector segments: if the total power of detectors B-D increases by ΔΡ, the total power detected by A-C also decreases by the same amount:

..ν-'Λ-'·» + AP)-(P, +P _C -ΔΡ) _(P,+P _D )-(P _A +P _C ) _r 2ΔΡ ^b p ^X det--- ^X _áet + · *det (?)

From this, the change in the measured position can be calculated. If, in the case of a homogeneous light spot, we denote the section formed at the intersection of the light spot and the y-axis with “d” (see Figure 5):

-8Δ^ =

2ΔΡ _2·<Ζ·Δχ·Ρ/£ _2·ί/·Δχ

Ρ '*det ^— ρ '*det ^— $ * *det ' (4) where “S” denotes the area of the light spot. For a square light spot with side length 2-Xdet:

</ = 2x _da ; S = 4xi, 3 Af = _=Δϊ , ₍₅₎ ö

i.e. the measured center of gravity coordinate exactly coincides with the real position of the scintillation. In the case of a circularly symmetric spot, d(x) dependence exists, so Δξ(χ) is also location dependent, i.e. ξ = f(x). For example, at x = 0, in the case of a homogeneous circular spot with radius R:

d = 2R ; S = S ^! x- ; Λ = χ„ => ίξ = ±·Δχ (6) ο π χ - R, however, d = 0, i.e. Δξ = 0. For an inhomogeneous spot, the relationship can be calculated more complicatedly. Instead of analytical expressions, we present Figure 6, in which, as an example, scintillators from a 20 x 3.5 x 3.5 nnn LYSO crystal were modeled using the ZEMAX optical design program. The detectors are PMTs with a circular cross-section of 040 mm, so the size of the entire block is 80x80 mm. A 30 mm thick homogeneous polymethyl methacrylate light guide was placed between the exit surface of the crystal needles and the detector plane, the edges of which were provided with a reflective coating. A narrow air gap was formed between the exit surface of the crystal needle and the light guide. When preparing the figure, the crystal needle was continuously moved along the x-axis, and the measured center of gravity coordinate, ξ, was calculated at each position based on (2) (for 10 scintillators randomly positioned within the crystal). Each point on the curve represents the average of 10 counts.

In Figure 7, we can see the standard deviation of the 10 measurements made at each x-point, i.e. the noise burdening the measurement, Δξ^ε-ΐ. (We did not take electronic and other noises into account during our studies.) This noise is caused by the fact that we launched 16,000 photons per scintillation, of which approx. 3% are converted into photoelectrons in the PMTs. In other words, each measurement is made based on less than 500 photoelectrons, and the random location of their positions makes the measured ξ-coordinate of the centroid of the entire light spot uncertain. The

It can also be seen from Figure -97 that the noise is smallest when the scintillation location is |x| > x _det . This is a consequence of the fact that in this position most of the light falls on the BD (or AC) detector pair and not on the dead zone between the circular PMTs.

Scintillation from two nearby scintillator elements, crystal needles, can be distinguished from each other if they are separated by a distance equal to or greater than the detector resolution (Ax _res ). The resolution can be determined by the following formula:

A*„,M = 2-(^W ·Δ^(χ). (7) \ ebe /

The resolution diagram obtained from our ZEMAX model is shown in Figure 8. It can be seen that the resolution completely deteriorates as x approaches the value of x<i _e t. Since it is advisable to use crystals of the same size in a scintillation detector, the size of the crystal needles will be determined by the largest Ax _res value. In Figure 8, the curve takes on enormous values towards the edge of the block detector. In spite of working with larger crystal needles, x cannot be resolved over a significant range.

It follows from equation (7) that the resolution depends on two things: the amount of noise and the slope of the f(x) function. The noise can be reduced (and thus the resolution can be improved) by improving the light efficiency of the scintillation detector, or e.g. by using square PMTs instead of circular ones. The slope of the f(x) function can be increased (and thus the resolution can be improved) by e.g. by using a square spot instead of a circular one. However, these are very difficult to implement, so we chose a different development direction when creating the invention.

According to one of the objectives of the invention, the scintillation light sensitivity of the detector module must be increased. This is necessary because the small number of photodetectors used for cost reduction results in dead spaces in the center and at the edges. However, scintillator crystals also fall on the dead spaces, i.e. areas outside the detection surface 22, and the light emitted by them must also be detected with good efficiency. According to the invention, the light emerging from the crystals must be shaped and deflected so that as little scintillation light as possible falls on the parts not covered by the photodetectors 20. This path leads to the achievement of the best possible light efficiency.

-10To achieve higher light efficiency and higher spatial resolution, the invention includes light deflecting devices that are assigned to individual scintillator elements, depending on the position of the given scintillator element in the arrangement, for example in a matrix arrangement, to deflect the light emerging from the scintillator element in a predetermined direction and angle from the direction of the scintillator element, and thereby direct the light emerging from the scintillator element onto the sensing surface.

To solve the problem shown in Figure 3, we have created light-deflecting devices characterized by variable parameters from crystal needle to crystal needle, or a light-deflecting unit consisting of them. Depending on the parameters, this light-deflecting device changes the direction of the scintillation light cone axis starting from the crystal needles, thereby being able to influence the lateral (x-y) position of the light spot in the plane of the PMTs. The principle of operation is shown in Figure 17, where the light spot position 31 without light deflection (i.e. the real location of the scintillation) is marked with a dashed line. The shifted light spot 30 is detected by the PMTs as if the photons were coming from a virtual point (x') instead of the real position (x) of the scintillation crystal. That is, if the x'-x dependence is chosen as follows:

x' = /-'(x/c) => í = g(x) = /(x') = /(/- ¹ (x/c)) = x/c, (8) where “c” is a suitably chosen multiplication factor, then by applying the inverse function the measured centroid - scintillation position dependence can be linearized (ξ = g(x)), the slope of the resulting line can be influenced by the factor “c”.

In the following, we will detail how to optically create the desired x’-x function. To do this, we will first examine the arrangement shown in Figure 9, where we can see the path of photons originating from scintillation from the 10 scintillator elements, crystal needles, towards the PMT detectors. There are 32 coupling media between the crystal needles and the 16 light guides in order to achieve the highest possible light efficiency. Figure 10 shows the radiation distribution corresponding to this case in the plane of the PMT detectors, i.e. on the 22 detection surfaces. The resulting light spot is inhomogeneous (shows a Lambert distribution), and its total power is 4.4740 W (the total power radiated during scintillation is considered to be 1 W). The

- 11 light spots are too wide, so the calculation determining the center of gravity is not very sensitive to lateral displacement.

If we place 34 air-gap light deflectors between the 16 light guides and the 10 scintillator elements as shown in Fig. 12, instead of the Lambert light spot, we obtain a well-defined, more or less homogeneous irradiance luminosity in the detector plane. The light loss is relatively low: in the case of 16 light guides of the given thickness, the light efficiency is reduced to only 92% compared to the matched case (see Fig. 10).

Due to the advantages detailed later, we propose the design of an optical unit, a so-called light deflector, which is shown enlarged in Figure 13. In the solution according to Figure 13, an air-gap light deflector 34 is fitted to the exit end 14 of the scintillator element 10 as a light deflector attachment. The air-gap light deflector 34 is made of a material whose refractive index - preferably 1.5 - differs from the refractive index of the scintillator element 10, preferably 1.8. As a result, the surface between the scintillator element 10 and the lower part of the air-gap light deflector 34 functions as an edge filter due to total internal reflection; it reflects light rays with a large angle of incidence and transmits those with a small angle of incidence. The path of the light rays is indicated by the lines in the figure; It can be seen that the lower part of the air-gap light deflector 34 is less saturated with lines than the scintillator element 10, i.e. only light rays with a small angle of incidence reach there. The air gap 36 formed in the air-gap light deflector 34 functions as a layer with direction-dependent transmission.

The reflector layer 38 between the scintillator elements 10 preferably extends to the lower part (below the air gap) of the air-gap light deflector 34, while its upper part is provided with a light-absorbing surface 39 on its side surfaces, preferably the side surfaces are blackened.

In a preferred embodiment of the invention, the solution according to Figure 13 is used with a modification that the air gap therein is tilted according to Figure 14. The light deflection device 40 according to the invention then generally comprises a layer with directional transmission 36 arranged at the exit end 14 of the scintillator element 10, which is arranged at an angle with a plane perpendicular to the longitudinal direction of the scintillator element 10.

- 12A 36 directional transmission layer, which can be produced preferably by means of an inclined air gap, e.g. two prisms assembled with an air gap as shown in the enlarged version in Figure 16. The axis of the scintillation light cone will follow the normal of the air gap surface, i.e. the angle of the light cone can be controlled by the angle of the light deflector with a small loss of light efficiency. The resulting light spot is shown in Figure 15, the light efficiency is 85% of the unaligned case.

The light deflecting device 40 is suitably surrounded by the reflector layer 38 (reflective surface, e.g. 3M ESR foil) covering the crystal needles for greater light efficiency, and the side surface of the prism closer to the light guide 16 is blackened so that the ghost image coming from the reflective surface is suppressed by the light absorbing surface 39. Of course, a case is also conceivable in which the reflector layer 38 does not overhang the crystal needles, and one in which the edge of the prism is not blackened.

The linearly compensated detection characteristic achieved by the light deflection devices 40 according to the invention is shown in Figures 17 and 18.

In a further advantageous embodiment, instead of using light deflectors consisting of individual prisms, we incorporate the light deflector prisms of several crystal needles into a component, which is formed, for example, by injection molding. Such a solution is shown in Figure 19. Here, the light deflector devices 40 are formed in a light deflector unit 44 fitted to the arrangement of the scintillator elements 10. The inclined surfaces are then not on separate prisms, but on two parts that fit together, for example, formed by injection molding. Of course, any combination of the individual prism and the combined solution is also conceivable.

The opening angle of the scintillation light cone depends, among other things, on the refractive index of the filling material of the directional transmission layer 36 formed as an oblique slit. Thus, in the light deflecting device 40, that layer can be filled with a medium suitable for the given application, e.g. an adhesive or some liquid.

The light cone shown in Figures 14-15 cannot be produced only with an air gap. An embodiment is also conceivable in which the inclined surface between the two prisms shown in Figure 16

-13 coincide, and between them there is a layer with some other type of directional transmission. Such layers may be, for example: a thin film structure or subwavelength grating, a photonic crystal, a gradient index layer, etc. The light deflecting device 40 may also be designed as a microlens arrangement.

The operation of our proposed solution is illustrated by the following modeling result obtained with the ZEMAX optical design program. The light spot shift from crystal needle to crystal needle was implemented with an air-gap light deflector with a changed tilt angle according to the function shown in Figure 20. Instead of ξ = f(x), the new function ξ = g(x) shown in Figure 21 is visibly linear (c = */ ₂ , according to our choice). In the latter figure, the light spot position obtained by centroid calculation is shown as a function of the scintillation location according to the arrangement according to Figure 2. The average of 10 measurements is reported at each x-point. The angle (a) of the light deflector shown in Figure 16 was changed according to Figure 20.

The measured centroid uncertainty (Δξηοίίβ) due to the small number of photoelectrons is shown in Figure 22. Here, the random error of the measured coordinate of the centroid of the scintillation spot is shown, at a single standard deviation, based on 10 measurements per x-point, in the case of light deflection compensation corresponding to Figure 20. The resolution curve obtained based on equation (7) is shown in Figure 23, where the resolution calculated from the average of 10 measurements per x-point is displayed as a function of the scintillation position in the case of light deflection compensation corresponding to Figure 20. The maximum distance of the resolvable points is around 2 mm, so 40x40 crystal needles can be placed in the 80 mm detector module using our method. If we refine the shape of the curve shown in Figure 20, which deviates slightly from the linearity of Figure 21, the graph of Figure 23 can be transformed into a horizontal line, the expected value of which is Ax _res ~ 1.2 mm, which allows more than 60x60 crystals to be placed in the block.

As can be seen from the above, the invention has two extremely important advantages. On the one hand, it results in the best possible utilization of the light generated during scintillation in the crystal needles by redirecting the illumination that would otherwise fall on the dead spaces to the light detectors, and on the other hand, it allows the detection to be compensated with an arbitrary, pre-selected function, thus achieving the desired detection characteristic.

- 14 However, during the detection process of photons generated during scintillation, significant light loss may occur — even with the above measures. Not all of the photons decoupled from the crystal into the light guide 16 reach the active surface of the PMT photodetectors; due to the light cone-like projection, a part of the photons fall into the dead spaces between the photodetector surfaces or outside the photodetectors, even in the case of a directional design. To eliminate this, the most obvious solution is to place a reflecting layer on the side of the light guide facing the photodetector between the photodetector surfaces. This solution reduces the loss, but the reflected photons may only reach the photodetector surface after multiple reflections, which may distort the center of gravity of the detected light, which adversely affects the determination of the position.

A more sophisticated way to reduce the loss is to modify the intensity distribution of the light cone produced by the air gap. The essence of the modification is that the intensity is reduced inside the light cone, which usually falls between the detectors, while it is increased outside the cone. During the solution, the resolution of the crystal needles does not change, while the utilized light power increases significantly. The mentioned intensity distribution can be implemented, for example, with a dielectric thin-film system, an optical grating, a photonic crystal or a gradient index layer. Coatings with special reflection properties can be implemented with dielectric layer systems and the other solutions listed above. The value of the reflection coefficient depends on the wavelength and the angle of incidence. The wavelength of the photons generated during scintillation is almost the same, therefore the reflection will mainly depend on the angle of incidence. The layer structure must be prepared in such a way that the reflection coefficient value is high for small angles of incidence, while it is low for larger angles of incidence. The layer reflects photons reaching the layer at small angles back into the crystal, from where, after multiple reflections, the layer system transmits them at larger angles of incidence, i.e., the intensity on the surface of the PMT detector decreases at small angles of incidence, while it increases at large angles of incidence.

Such solutions are shown in Figures 24-26, where the operation of the intensity modulation device 46, preferably a layer (marked with a thick line in the figures) creating a 48-hole light cone is illustrated in the following configurations:

(a) Figure 24 for a crystal in the center of the crystal matrix, (b) Figure 25 for a light deflector with a tilted surface, and (c) Figure 26 shows another arrangement of the intensity modulation device 46, where the inner, “holey” region of the light cone is tilted along with the entire light cone.

The intensity modulation device 46, preferably a layer system, should be placed between the crystal needle and the light detector, preferably close to the crystal needle. The intensity distribution created by this solution for a central crystal needle is shown in Figure 24. If we want to apply a layer system to a crystal needle that is not located in the center of the matrix, then two different cases are possible. In the first case, the location of the hole created in the light cone does not change with respect to the normal to the incidence, depending on the angle of the prisms defining the air gap. In this case, the layer is between the prism and the crystal needle. This solution is shown in Figure 25. In the second case, the hole is always located in the center of the light cone, regardless of the angle of the prisms. In this case, the layer is located between the first prism and the air gap. This solution is shown in Figure 26.

The aforementioned solutions can be applied individually to each crystal needle, since the thickness of the layer system is only a few micrometers. It is not necessary to apply the layer to each crystal needle. The value of the light efficiency varies as a function of the critical angle, which results in a high reflection coefficient at smaller angles of incidence. The dependence of the relative light efficiency on the critical angle is shown in Figure 27, which shows the light efficiency calculated from the total scintillation light per 4 pcs. 040 mm PMTs as a function of the internal aperture angle of the “holey” central light cone for the crystal in the center of the detector module, in the case of an unaligned, 30 mm thick, mirror-edged light guide, optimized reflector configuration.

The invention therefore also extends to preferred embodiments in which the intensity modulation of the beam emerging from the crystal needles is implemented according to the direction, i.e. where the detector module also includes intensity modulation means 46 assigned to each scintillator element 10, which create a predetermined intensity distribution of the light emerging from the scintillator element 10 in a plane perpendicular to the direction of the light.

The selection of materials to be used for the arrangement of the 10 scintillator elements should also be made based on the above considerations. The materials that result in the best light efficiency

-16 configuration was determined as a result of a series of measurements. Based on ZEMAX modeling results, it was likely that the configuration with the cladding polished (i.e. utilizing total reflection) and the entrance (source-side) side diffusely reflecting, thus preventing light confinement, would be the most favorable by using a reflector that does not impair total reflection on the cladding side and maximizes light efficiency on the entrance side as well. Figure 28 shows the path of a single light ray in the crystal needle if the source side is diffusely reflecting, and Figure 29 shows the case if the source side is also polished. It can be seen that while in the first case the light ray “figures out” after a few reflections, in the second case we see only a large number of reflections and the light ray does not exit the crystal needle. On the light detector side, light loss must be minimized, therefore the exit surface must be polished. Thus, we examined samples with different surfaces (polished or ground) using different reflectors on the cladding and entrance side.

To determine the best configuration, the light efficiency of 10x10x20 and 3.5x3.5x20 mm LYSO crystal samples was measured using a Hamamatsu H2431-50, a Pozitron HV-1601 power supply, a Linearlab TIA2 amplifier, a Tukan 8k USB analyzer, and a ~5 mm diameter collimated ²² Na isotope. During the measurement, the location of the 511 keV annihilation photopeak was determined in the fixed arrangement of the source - sample - detector based on a large number of scintillators, which is proportional to the detected fraction of the energy of the γ-photons (light efficiency). The detector-side surface of the samples was polished in all cases. There was an air gap of a tenth of a mm between the detector and the measured crystal needle. In order to compare measurements made at different times and with different devices, we always measured the simplest configuration, i.e. the reference crystal with a size of 10x10x20 mm polished on all sides, and compared the light efficiency of the other configurations to this, i.e. the light efficiency was measured in relative units. The reliability of the measurement results is characterized by the fact that the standard deviation was ±5%. Table 1 shows some of the results of the comparative measurements. In the table, P denotes the polished surface, and C denotes the ground surface, the reference sample is number 3. We tested Teflon tape, aluminum foil, paint (BaSO ₄ ), Lumirror, and 3M as reflectors. The light efficiency with the aluminum foil reflector is significantly lower, so we omitted it from the table.

-17The samples (1) - (4) do not have a reflector, only their surfaces are different. It can be seen that it is advantageous if the cladding is polished and the source-side surface is diffusely reflective (4). In the case of the sample with a smaller cross-section (5), the light efficiency is slightly higher than in the sample with a similar design (4), i.e. the geometry of the crystal needle also affects the light efficiency. In the sample groups (6) - (11), we changed the state of the source-side surface (C or P) and the source-side reflector. The best source-side diffuse reflector is the Lumirror (11) fitted to the polished bottom. In the group (12) - (18), there is a reflector on the cladding as well. The cladding is polished in all cases. There is an air gap between the cladding and the reflector, i.e. total reflection prevails. The best (18) configuration is the sample with 3M unfitted on the cladding and Lumirror 60L fitted on the source side, polished on all surfaces. For the smaller crystal, configuration (19), i.e. Lumirror fitted to a polished source-side surface and 3M with an air gap on the cladding, also provides similarly good light efficiency. For the small size, configuration (20), with an unfitted 3M on a polished cladding on all surfaces and a diffusely reflecting Lumirror on the source side, is also the most favorable.

Sample Reflector Spinning page Coat Measurement (relative) 1 10x10x20 P C 0.79 2 10x10x20 C c 0.89 3 10x10x20 reference P P 1.00 4 10x10x20 c P 1.14 5 3.5x3.5x20 c P 1.20 6 10x10x20 Teflon on the bottom c P 1.55 7 10x10x20 bottom Lu60 C/P P 1.61 8 10x10x20 on the bottom Lu60 or c P 1.67 9 10x10x20 3M on the bottom c P 1.68 10 10x10x20 bottom painted P P 1.84 11 10x10x20 on the bottom Lu60 or P P 1.90 12 10x10x20 paint on the bottom, paint on the coat P P 1.94 13 10x10x20 Teflon on the bottom, Teflon on the jacket P P 2.06 14 10x10x20 3M on the bottom, 3M on the mantle P P 2.38 15 10x10x20 Lu60 on the bottom, 3M on the jacket P P 2.45 16 10x10x20 Teflon on the bottom, 3M on the jacket c P 2.51 17 10x10x20 3M on the bottom, 3M on the mantle c P 2.66 18 10x10x20 on the bottom Lu60 or., on the jacket 3M P P 2.80 19 3.5x3.5x20 on the bottom Lu60 or., on the jacket 3M c P 2.85 20 3.5x3.5x20 Lu60 on the bottom, 3M on the jacket P P 2.97

Table 1

- This is favorable for the design of the 18A crystal matrix, because we can use a relatively thick (188 pm) Lumirror type reflector on the source side that can be easily glued, and a thin (65 pm thick) 3M ESR film on the polished cladding, which results in the smallest possible dead space between the crystal needles. We can also make a so-called "beer compartment" from 3M ESR by mechanical processing or laser, into which the crystal needles can be inserted during matrix preparation.

The scintillator elements 10 are therefore preferably cubic scintillator crystals, all surfaces of which are polished, and in the arrangement of the scintillator elements 10, the scintillator elements 10 are separated from each other by a bright reflector layer with at least 90% reflection, and a diffuser layer with at least 90% reflection is attached to their input end 12.

The reflector layer is preferably an optically mismatched polymer film with at least 95% reflection, e.g. 3M ESR film, and the diffuser layer is preferably a polyethylene terephthalate-based polyester film with at least 95% reflection, e.g. Lumirror E60L film.

The invention is of course not limited to the embodiments described in detail above, but may include further variations and modifications within the scope of protection defined by the claims.

Claims (13)

-19Patent claims 1. A detector module for a positron emission tomography (PET) scanner, which is of a beam splitting design and comprises - an arrangement of elongated scintillator elements (10) arranged parallel to each other, the scintillator elements (10) having an inlet end (12) and an outlet end (14), - an arrangement of light detectors (20) positioned opposite the exit ends (14) of the scintillator elements (10) and detecting light emerging from the scintillator elements (10), said light detectors (20) defining a detection surface (22), characterised in that - it also includes light deflecting means (40) which are assigned to each scintillator element (10) in a predetermined direction and angle from the direction of the scintillator element (10), depending on its position in the arrangement, and thereby directing the light emerging from the scintillator element (10) onto the sensing surface (22). 2. The detector module according to claim 1, characterized in that the light deflecting means (40) comprises a layer (36) with directionally dependent transmission arranged at the exit end (14) of the scintillator element (10). 3. Detector module according to claim 2, characterized in that the directional transmission layer (36) is arranged at an angle to a plane perpendicular to the longitudinal direction of the scintillator element (10). 4. The detector module according to claim 3, characterized in that the directionally dependent transmission layer (36) is formed as a gap in a light deflecting overlay fitted to the exit end (14) of the scintillator element (10), which gap is an air gap or a gap filled with a material with a refractive index different from that of the light deflecting overlay. 5. Detector module according to claim 2, characterized in that the directional transmission layer (36) is formed as a thin film structure, an optical grating, a photonic crystal or a gradient index layer. 6. Detector module according to claim 1, characterized in that the light deflection device (40) is designed as a microlens arrangement. 7. Detector module according to any one of claims 1-6, characterized in that the light deflecting means (40) are formed in a light deflecting unit (44) fitted to the arrangement of the scintillator elements (10). 8. The detector module according to claim 1, characterized in that it also comprises intensity modulation means (46) assigned to each scintillator element (10) for creating a predetermined intensity distribution of the light emerging from the scintillator element (10) in a plane perpendicular to the direction of the light. 9. Detector module according to claim 8, characterized in that the intensity modulation device (46) reduces the intensity of the light emerging from the scintillator element (10) in the form of a light cone inside the light cone and is designed as an optical device that increases it in the outer part of the light cone. 10. The detector module according to claim 9, characterized in that the intensity modulation device (46) is formed as a dielectric thin film system, an optical grating, a photonic crystal or a gradient index layer. 11. The detector module according to any one of claims 1-10, characterized in that the scintillator elements (10) are columnar scintillator crystals, all surfaces of which are polished, and in the arrangement of the scintillator elements (10), the scintillator elements (10) are separated from each other by a bright reflector layer with at least 90% reflection, and a diffuser layer with at least 90% reflection is attached to their input end (12). 12. The detector module according to claim 11, characterized in that the reflector layer is an optically mismatched polymer film with at least 95% reflection, and the diffuser layer is a polyethylene terephthalate-based polyester film with at least 95% reflection. 13. A positron emission tomograph, characterized in that its scintillation detector comprises detector modules according to any one of claims 112. (15 sheets of drawings, 29 figures)