GB2484964A - Asymmetric crosswire readout for scintillator devices using silicon photomultipliers (SPMs) - Google Patents

Abstract

An imaging device comprises plural photodetectors, each connected to a different pair of first and second readout channels; the spatial position of an active photodetector is determined from the corresponding pair of active readout channels. The number of readout channels is less than the number of photodetectors. For each photosensor, its first readout channel is shared (crosswired) with a first plurality of other photodetectors, its second readout channel being shared with a second plurality of other photodetectors. In a first option, the second detector plurality is greater than the first plurality for at least some photodetectors (e.g. forming a rectangular array), with the position of an event at a photodetector determined from the active first and second readout channels, and with event energy and/or timing determined from the active first readout channel. A second option is that at least some of those photodetectors sharing the same first channel are not adjacent. Silicon Photomultipliers (SPMs, SiPM), as claimed, Geiger Avalanche Photodiodes (G-APD) or Micro Pixel Photon Counters (MPPC), may be used. The device has application to scintillator detectors, e.g. used in PET/PEM, CT, SPECT and other medical imaging.

Description

I

Imaging Device and Method The present invention relates to an imaging device and method.

Pixelated scintillator arrays have the potential to dramatically improve the performance of the gamma camera, particularly in PEM [Positron Emission Mammography] systems.

Detection and localisation of gamma events in the scintillator is performed by observing light flash induced by gamma quanta interaction in the scintillator. Traditionally, this 5 performed using one of the following approaches: I. Monolithic scintillators connected optically via continuous light guides to a number of PMTs thus providing light sharing among several detectors for every gamma interaction observed. The localisation of the event is established via calculation of centre of gravity of the light signal distribution over the PMTs. This technique is known as the Anger logic [H. Anger. A new instrument for mapping gamma-ray emitters. Biology and Medicine Quarterly Report UCRL, 1957, 3653: 38 (University of California Radiation Laboratory, Berkeley)]. This technique was first used to build gamma cameras and is still widespread today, particularly in SPECT (Single Photon Emission Tomography) cameras. However, there are some disadvantages associated with such gamma cameras, such as sensitivity to PMT parameter drifts, limited count rate.

2. Another, similar technique, known as block detector can be considered as an adaptation of the Anger logic to PET gamma cameras. It relies on the use of four PMTs optically connected to a light guide. The light guide on another side is interfaced to a pixelated scintillator array. The gamma camera is then assembled using these building blocks allowing faster count rate and reliable localization of gamma interaction points down to the crystal of interaction.

3. The block detector can also be built using APDs (Avalanche Photodiodes) instead of PMTs [H. Anger, referenced above]. However, when using solid state detectors, the main disadvantage of the Anger logic, using analog signals for crystal identification, still remains. This technique requires a lot of calibration data and makes localization sensitive to the drift of the photo-sensor parameters Also 3098572 3-1-rrtbrewer adcBtional problems arise when using solid state photo-detectors instead of PMTs.

APDs have a much higher noise level per area, degrading energy resolution and localization performance.

4. An alternative, simple approach, known as 1:1 coupling relies on separate readout of each scintillator pixel with a separate photodiode. The pixels are optically insulated from each other so gamma interaction signal is seen by only one photo-detector The advantage of pixelated scintillator is that it provides a more precise and reliable localisation of the gamma event interaction point. This technique provides an absolute position based on the particular photodiode channel rather than measurement of relative light output over several PMTs as carried out using the conventional Anger logic approach. The main limitation of 1:1 coupling is the high number of channels to be read out which increases system design complexity and increases overall manufacturing costs.

There have been numerous attempts to readout pixelated scintillator using either multichannel vacuum PMTs or an array of semiconductor photodiodes (of PIN or avalanche type), or a combination of both. Some of these designs are intended to provide DOE (Depth-Of-Interaction) information thus improving the gamma camera performance.

However, there are many limitations associated with these detectors:- * Vacuum PMTs are very bulky and expensive * PIN and avalanche photodiodes are expensive, requiring short interconnects to special low noise electronic amplifiers.

In addition, such photodiodes are plagued by nuclear counter effect resulting from direct detection of the gamma event in the photodiode rather than in the scintillator. This causes erroneous event data, especially if using top (toward the gamma source) placement of the photodiode on the scintillator.

Some multichannel PMT readouts are using crosswire technique for channel reduction. In this case, a number of resistors are used to reduce the channel count [Matrix output device readout system, United States Patent 6747263]. However, the use of the crosswire technique is far simpler for SPMs as it does not require this set of resistors, P 3O9857231-mbrewer SPMs are silicon photon sensitive devices made up of an array of very smaU Geiger-mode avalanche photodiode (APD) cells on a silicon substrate. Typically each cell is connected to one another to form one larger device with one signal output. The entire device size can be as small as lxi mm or much larger. SPMs, also known as SiPM (Silicon Photornultipliers), G-APD (Geiger Avalanche Photodiodes) or MPPC (Micro Pixel Photon Counters), have the potential to become a replacement for PMTs and avalanche photodiode for use as photo detectors in positron emission topography (PET); Single photon emission computed tomography (SPECT), computed tomography (CT), and other radiation detectors. These devices are compact, have high gain, high quantum efficiency (about 20%-70%), which is better than that of traditional PMTs.

Due to their timing performance7 these devices have the potential to be used in time-of-flight PET applications. They are also insensitive to magnetic fields, a quality which makes them ideal for use in an MR environment.

SPMs produce a relatively large charge pulse when struck by a photon of the same amplitude regardless of the energy of the photon. When reading out conventional APDs, noise of the preamplifier significantly degrades timing and amplitude resolution performance for short (shorter than approx 500nS) light pulses. Compared to conventional APDs, SPMs provide much higher output amplitude. This essentially eliminates the impact of preamplifier noise and allows the use of much longer electrical interconnects between the photodiode and following amplifier electronics.

SPMs are practically free from so-called "nuclear counter effect", meaning they are insensitive to direct gamma interaction in the material of the SPM. This is highly advantageous for scintillator based gamma cameras. Additionally, being build on low resistivity, inexpensive silicon wafers using common silicon processing, SPMs are potentially much cheaper to manufacture than competing PIN or APD type photodiodes.

The reader is also referred to the following references for background information: [APD-based PET detector for simultaneous PET/MR imaging. Ronald Grazioso, Nan Zhang, James Corbeil, Matthias Schmand, RaIf Ladebeck, Markus Vester Günter Schnur, Wolfgang Renz and Hubertus Fischer] and [Performance of Philips Gemini TF PET/CT Scanner with Special Consideration for Its Time-of-Flight Imaging Capabilities. Suleman Surti Austin Kuhn Matthew E Werner, Amy E Perkins Jeffrey Kolthammer and Joel S 3og85723-1-mbrewer Karp].

According to a first aspect of the present invention there is provided an imaging device comprising a plurality of photodetectors in a predetermined spatial arrangement. Each photodetector is connected to a different predetermined pair of first and second readout channels so that the spatial position of an active photodetector can be determined from the corresponding pair of active readout channels. The number of readout channels is less than the number of photodetectors. For each photodetector its first readout channel is shared with a first plurality of other photodetectors of the arrangement, and its second readout channel is shared with a second plurality of other photodetectors of the arrangement. A first option is that the second plurality is greater than the first plurality for at least some of the photodetectors, with the position of an event occurring at one of the photodetectors being determined from the active first and second readout channels, and with the energy and/or timing of the event being determined from the active first readout channel. A second option is that at least some of those photodetectors sharing the same first channel are not adjacent in the arrangement. A third option is that the photodetectors are silicon photomultiplier detectors. The first to third options can be used separately, in isolation, or in any combination.

The device may comprise circuitry adapted to determine one or more, preferably all, of the position, energy and timing of the event.

The energy and/or timing of the event may be determined solely from the active first readout channel.

The device may comprise a pixellated scintillator having a plurality of scintillator elements, and wherein each photodetector is arranged in optical communication with a different respective scintillator element.

The second plurality may be greater than the first plurality at least by a factor of two The second plurality may be greater than the first plurality at least by a factor of three.

The second plurality may be greater than the first plurahty at least by a factor of four 30985723 1 mbrewer It may be that aU of those photodetectors sharing the same first channel are not adjacent in the arrangement.

It may be that at least some of those photodetectors sharing the same second channel are not adjacent in the arrangement.

The first plurality may be the same for each photodetector in the arrangement.

The second plurality may be the same for each photodetector in the arrangement.

The device may comprise A first channels and B second channels and A x B photodetectors in the arrangement.

The photodetectors may be arranged in a substantially rectangular grid array.

According to a second aspect of the present invention there is provided an imaging method comprising arranging a plurality of photodetectors in a predetermined spatial arrangement, in which each photodetector is connected to a different predetermined pair of first and second readout channels so that the spatial position of an active photodetector can be determined from the corresponding pair of active readout channels, and in which the number of readout channels is less than the number of photodetectors, and in which, for each photodetector, its first readout channel is shared with a first plurality of other photodetectors of the arrangement, and its second readout channel is shared with a second plurality of other photodetectors of the arrangement. A first option is to arrange for the second plurality to be greater than the first plurality for at least some of the photodetectors, to determine the position of an event occurring at one of the photodetectors from the active first and second readout channels, and to determine the energy and/or timing of the event from the active first readout channel. A second option is to arrange that at least some of those photodetectors sharing the same first channel are not adjacent in the arrangement A third option is to arrange for the photodetectors to be silicon photomultiplier detectors. The first to third options can be used separately, in isolation or in any combination According to a third aspect of the present invention there is provided a silicon photomultiplier imaging device compnsing a plurality of silicon photomultiplier detectors in 3O985723-1-mbrever a predetermined spatial arrangement. Each silicon photomultiplier detectors is connected to a different predetermined pair of first and second readout channels so that the spatial position of an active silicon photomultiplier detectors can be determined from the corresponding pair of active readout channels. The number of readout channels is less than the number of silicon photomultiplier detectors. For each silicon photomultiplier detectors its first readout channel is shared with a first plurality of other silicon photomultiplier detectors of the arrangement, and its second readout channel is shared with a second plurality of other silicon photomultiplier detectors of the arrangement. The second plurality may be greater than the first plurality for at least some of the silicon photomultiplier detectors, with the position of an event occurring at one of the silicon photomultiplier detectors being determined from the active first and second readout channels, and with the energy and/or timing of the event being determined from the active first readout channel. It may be that at least some of those photodetectors sharing the same first channel are not adjacent in the arrangement.

An embodiment of the present invention relates to gamma imaging cameras, particularly to cameras based on Pixelated Scintillators combined with Solid-State Photodiode, such as Silicon Photomultipliers.

According to an embodiment of the present invention there is provided a solid-state gamma imaging array comprising: an X x Y array of 5PM photodetectors arranged in row and columns and optically interfaced to pixelated scintillator based on 1:1 coupling so number of the SPM photodetectors is equal to the number of scintillator elements.

According to an embodiment of the present invention the SPM photodetectors are wired into a scrambled crosswire pattern thus providing information on neighbouring pixels which are wired through different energy channels.

According to an embodiment of the present invention there is provided an asymmetric crosswire wiring scheme of 5PM photodetectors configured using coincidence of signals from A and B sets of channels to identify photodetectors of interaction where the number of electronic readout channels is lower than the number of the photodetectors.

According to an embodiment of the present invention the quantity of A channels used for energy estimation of the event is substantially higher than the quantity of B channels used 30985723 1 mbrewer only for event localisation thus providing optimization of the readout performance against dark noise of the 5PM photodetectors.

According to an embodiment of the present invention there is provided readout electronics providing pixel of gamma quanta interaction localization via coincidence of two channels of the crosswire scheme and estimation of the energy deposited in the pixel of interaction and also in the neighbour pixels via readout of adjacent crosswire energy channels.

Reference will now be made, by way of example, to the accompanying drawings, in which: Figure 1 illustrates a 4x4 crosswire scheme; Figure 2 illustrates an asymmetric crosswire readout scheme; Figure 3 illustrates the channel combinations for the scheme of Figure 2; Figure 4 illustrates an asymmetric scrambled crosswire readout scheme where each pixel is addressed via a unique combination of sixteen A and three B channels; Figure 5 illustrates an asymmetric scrambled crosswire readout scheme where a single pixel/gamma interaction has occurred; and Figure 6 illustrates an example of Asymmetric Scrambled Crosswire Readout.

As explained above, Pixelated Scintillators combined with Solid-State Photodiode readout such as Silicon Photomultiplier Arrays are replacing Vacuum Photornultiplier/ Continuous Scintillator based gamma cameras allowing the ability to build smaller camera heads achieving better spatial resolution and count rate performance However, the readout of individual photodiode elements corresponding to individual scintillator crystal presents a significant challenge because of higher channel count resulting in greater system manufacturing costs.

Relying on the unique properties (very high internal gain and high photo detection efficiency) of the Silicon Photomultipliers (5PM) a solution according to an embodiment of 3O9857231mbrewer the present invention proposes an Asymmetric Crosswire Readout (ACR) that effectively overcomes problems related to large channel count scintillator readout systems.

A method of SPM-based pixelated scintillator readout using Asymmetrical Scrambled Crosswire Readout is also proposed, which allows channel reduction and simplified electrical interconnection whilst limiting the impact of 5PM dark noise on the readout performance and providing the capability to process multi-pixel gamma interaction.

Asymmetric Crosswire Readout (ACR) is based on the use of a reduced channel readout technique known as crosswire technique applied to the SPMs. This technique is used primarily in the readout of charged particle detectors (silicon micro-strip detectors) and gamma imaging.

To explain the principle, first consider an NxN matrix of SPM detectors wired in such a way that: * All anodes of the same column devices are connected to form a set of "X" outputs * All cathodes of the same row devices are connected to form a set "Y" outputs.

An example of such wiring for 4x4 sensor matrix is shown in Figure 1, which constitutes an embodiment of the present invention.

If one of the devices is activated (1fires') in response to light applied to that device, it will generate a current pulse in its corresponding X and Y outputs, allowing that device to be identified by its unique X.Y signature. Such a cross-wire readout technique has not previously been proposed in relation to SPMs.

One limitation of this cross-wring technique is the inability to detect more than one hit at any time along the same row or column of photodetectors. However, an advantage in using the cross-wiring technique is the reduction in readout channels compared to one-to-one readout (one pixel -one channel). In the case of a square matrix of NxN pixels this corresponds to a reduction factor F = N*N I 2*sqrt(N) As a consequence of this channel reduction the interconnection of SPM output signals is simplified For scintillators with low light output and long output pulses such as Bismuth Germanate 3D985723-1 -mbrewer Crystals (BGO) used in PET (Positron Emission Tomography), additional SAM dark noise contribution from row/column pixels can have a significant negative impact on the performance. To overcome this, an Asymmetric Crosswire Readout (ACR) is proposed according to a further embodiment of the present invention.

In this context, asymmetric means the number of X readout channels is different and preferably significantly different from the Y side. Such an approach allows readout optimized for signal/noise performance whilst keeping the total number of readout channels considerably lower than one-to-one.

Alternative techniques {(W012008/107808) Improved Light Detection in a Pixelated PET Detector] do not use 1:1 scintillator-photodetector coupling. Instead they use intentional light sharing combined with analog information localization. However, this complicates the array design and requires analog calibration for correct gamma event localization resulting in a low tolerance of temperature, count-rate and age-related drifts.

Therefore, a further embodiment of the present invention aims to overcome the problems and disadvantages associated with current readout designs using Asymmetric Crosswire Readout (ACR) to provide a vastly improved method for pixelated scintillators readout using SAM arrays.

A yet further embodiment of the present invention is based on Asymmetric Scrambled Crosswire Readout (ASCR) which allows effective and robust processing of multipoint gamma interaction in 1:1 pixelated scintillators -photodetector arrays with a reduced number of electronic channels. This avoids the introduction of both intentional light sharing and analog information (observed signal amplitude) based Focalization as in An embodiment based on ACR is illustrated in Figure 2 which itself is based on the crosswire technique with an increased channel count to reduce noise level. In this example the SAM readout domain will comprise of groups of three 4x4 arrays connected together using the crosswire technique to create a 12 x 4 array (12 A channels and 4 B channels). The pairing of channels for each pixel is illustrated in Figure 3. Each pixel is marked alb where a is the A channel and b" is the B channel connected to that pixel 30985723 1 mbrewer In this case the 12 column channels (A channels), connected each to 4 SPM devices, are used to detect an event (such as a gamma interaction in one of the pixel scintillators) and perform time stamping and energy estimation of the event. Since each channel is only exposed to the noise of 4 pixels, the noise does not have significant impact on the performance. The other 4 channels (B channels, or row readout channels) are affected by the noise from 12 pixels (connected to 12 SPM devices). These will be used mainly to determine which of the 4 actual SPMs in the column has caused the event.

Therefore each photodetector is connected to a different predetermined pair of A and B readout channels so that the spatial position of an active photodetector can be determined from the pair of readout channels that is activated. For each photodetector its readout channel A is shared with a first plurality of other photodetectors of the array (four in the Figure 2 example), and its readout channel B is shared with a second plurality of other photodetectors of the array (twelve in the Figure 2 example). (In a regular grid, the first and second plurality is the same for each photodetector, with the first and second plurality being the B and A channel counts respectively, though this need not be the case in other, non-regular, arrangements.) For an event occurring at one of the photodetectors, the position of the event is determined from readouts from the pair of active readout channels A and B, while the energy and/or timing of the event is determined entirely from the active readout channel A. The second plurality is greater than the first plurality (for each of the photodetectors in this regular grid, though there would be at least some advantage if this were the case for some but not all photodetectors in non-regular arrangements). In this example, the second plurality is greater than the first plurality by a factor of three. In other examples, the factor might be higher (e.g. four) or even lower (e.g. two). The factor need not be an integer number.

Measurements have shown that the energy resolution degradation from 1 pixel to 4 pixels is not significant (change of -0.3% around -20% FWHM at 5llkeV). Hence the building block of three 4x4 arrays can be used for applications using BOO.

Another concern in readout of pixelated scintitlators is the high probability of an event to undergo two point interaction (typically Compton scatter followed by photo-absorption) resulting in energy shanng between two or more pixels Because of that to preserve good photo-peak efficiency of the system it is desirable to sum up the energy from two or more pixels to calculate the event energy In this case one of the pixels triggers the readout by 30985723-1-mbrewer an over-threshold signal. The energies of this pixel along with its neighbouring pixels are simultaneously read to provide an overall energy value.

However, the crosswire readout technique based on the column/row, as presented earlier, implies some limitations on the Neighbour Readout capabilities because the scheme cannot separate energy contributions spread between two or more pixels in the same column.

In order to avoid the above mentioned limitation an Asymmetric Scrambled Crosswire Readout (ASCR) technique is proposed. In this case, instead of wiring column readout lines straight along pixel columns, column readout lines are organized in such a way that neighbouring pixels always belong to a different readout line.

The schematic in Figure 4 demonstrates an example where one pixel fires within the readout domain and highlights how the ASCR technique is deployed.

Figure 4 shows A/B channel numbers wired into an asymmetric scrambled crosswire configuration; in this example 16 channel As and 3 channel Bs are deployed.

Each A channel shares the noise contributon for 3 pixels only and this can be used for precise energy and timing readout. The B channels are used to identify the pixel where the primary gamma interaction has occurred.

By way of example, Figure 5 shows a situation where pixel 10/2 (X channel 10, Y channel 2) has fired due to a Gamma event interaction with this pixel. The pixel itself is localized through observation of S Channel#2 and A Channel#10 firing simultaneously while the energy value of this interaction can be observed from A Channel#1 0 alone. p As a result of pixel 10/2 firIng, neighbouring pixels 5/2, 6/2, 7/2, 9/2, 13/2, 14/2, 15/2, 11/2, 7/2 can be examined, as illustrated in Figure 6, for partial gamma event energy deposit by sampling corresponding A channels 5, 6, 7, 9, 11, 13, 14, 15. Use of scrambled readout in this case allows unambiguous readout of energy deposit in each neighbour pixel 30985723-1-mbrewer It will also be appreciated by the person of skill in the art that various modifications may be made to the above-described embodiments without departing from the scope of the present invention as defined by the appended claims.

For example, although it is stated above that the energy and/or timing of the event is determined entirely from the active readout channel A, due to the smaller number of photodectors sharing that channel, this is not to preclude other embodiments where some account (however small) is taken of the active readout channel B to determine the energy and/or timing of the event. For example, where the A and B channels are combined to determined the energy and/or timing, the proportion of A channel within that combination would preferably be 100%, but may be lower, for example more than 50%, or more than 75%, or more than 90%. This could be summarised by stating that the energy and/or timing of the event is determined primarily from the active first readout channel.

Furthermore, it will be appreciated that although Asymmetric Crosswire Readout (ACR) and Asymmetric Scrambled Crosswire Readout (ASCR) embodiments are described above, a Scrambled Crosswire Readout (SCR) is also possible. In other words, it is not essential to use the scrambled crosswire technique in conjunction with an asymmetric configuration of photodetectors. F 30985723-1-mbrewer

Claims (4)

1. Claims 1. An imaging device comprising a plurality of photodetectors in a predetermined spatial arrangement, wherein each photodetector is connected to a different predetermined pair of first and second readout channels so that the spatial position of an active photodetector can be determined from the corresponding pair of active readout channels, wherein the number of readout channels is less than the number of photodetectors, and wherein for each photodetector its first readout channel is shared with a first plurality of other photodetectors of the arrangement, and its second readout channel is shared with a second plurality of other photodetectors of the arrangement, characterised in that: (a) the second plurality is greater than the first plurality for at least some of the photodetectors, with the position of an event occurring at one of the photodetectors being determined from the active first and second readout channels, and with the energy and/or timing of the event being determined from the active first readout channel; and/or (b) at least some of those photodetectors sharing the same first channel are not adjacent in the arrangement; and/or (c) the photodetectors are silicon photomultiplier detectors.
2. 2. An imaging device as claimed in claim I, comprising circuitry adapted to determine one or more, preferably all, of the position, energy and timing of the event.
3. 3. An imaging device as claimed in claim I or 2, wherein the energy and/or timing of the event is determined solely from the active first readout channel.
4. 4. An imaging device as claimed in claim 1, 2 or 3, comprising a pixellated scintillator having a plurality of scintillator elements, and wherein each photodetector is arranged in optical communication with a different respective scintillator element.An imaging device as claimed in any preceding claim wherein the second plurality is greater than the first plurality at least by a factor of two 30985723-1-mbrewer F 6. An imaging device as claimed in any preceding claim, wherein the second plurality is greater than the first plurality at least by a factor of three.7. An imaging device as claimed in any preceding claim, wherein the second plurality is greater than the first plurality at least by a factor of four.8. An imaging device as claimed in any preceding claim, wherein all of those photodetectors sharing the same first channel are not adjacent in the arrangement.9. An imaging device as claimed in any preceding claim, wherein at least same of those photodetectors sharing the same second channel are not adjacent in the arrangement 10. An imaging device as claimed in any preceding claim, wherein the first plurality is the same for each photodetector in the arrangement.11. An imaging device as claimed in any preceding claim, wherein the second plurality is the same for each photodetector in the arrangement.12. An imaging device as claimed in any preceding claim, comprising A first channels and B second channels and A x B photodetectors in the arrangement.13. An imaging device as claimed in any preceding claim, wherein the photodetectors are arranged in a substantially rectangular grid array.14. An imaging method, comprising arranging a plurality of photodetectors in a predetermined spatial arrangement, in which each photodetector is connected to a different predetermined pair of first and second readout channels so that the spatial position of an active photodetector can be determined from the corresponding pair of active readout channels, and in which the number of readout channels is less than the number of photodetectors, and in which, for each photodetector, its first readout channel is shared with a first plurality of other photodetectors of the arrangement and its second readout channel is shared with a second plurality of other photodetectors of the arrangement characterised by (a) arranging for the second plurality to be greater than the first plurality for at least 3D985723-1 -mbrewer some of the photodetectors, determining the position of an event occurring at one of the photodetectors from the active first and second readout channels, and determining the energy and/or timing of the event from the active first readout channel; and/or (b) arranging that at least some of those photodetectors sharing the same first channel are not adjacent in the arrangement; and/or (c) arranging for the photodetectors to be silicon photomultiplier detectors.3O9857231mbrewer