Sub-pixel temporal skew correction for positron emission tomography (PET)

Abstract

The present invention relates to a calibration method for a gamma-ray detector (100), the gamma-ray detector comprising: a pixelated scintillator array (110) for emitting scintillation photons at photon conversion locations in response to incident gamma rays (90); and a pixelated photodetector array (120) for determining the spatial intensity distribution of the scintillation photons. The present invention is based on the concept that the use of optical light sharing of scintillation photons, which are emitted in one element (i.e., one scintillator pixel of the scintillator array (110)) and distributed over a plurality of photodetector pixels (122) of the pixelated photodetector array (120), allows obtaining an estimate of the temporal skew between adjacent photodetector pixels (122). The present invention also relates to a calibration module (200) for the gamma-ray detector (100), the calibration module comprising a recorder (210) and a processing module (220) for performing the functions of the above method. Furthermore, the present invention also relates to a gamma ray detector (100) and a medical imaging device (50) comprising the gamma ray detector (100).

Description

Technical Field

[0001] The present invention relates to a calibration method for a gamma ray detector, a corresponding calibration module for a gamma ray detector, a gamma ray detector and a medical imaging device. Background Art

[0002] Positron emission tomography (PET) is a commonly used medical imaging technique. The technique is based on the detection of gamma-ray photons emitted after a radioactive tracer substance is administered to a patient (e.g., by injection into a peripheral vein). The nuclei of the radioactive isotope of the radioactive tracer substance emit positrons (positively charged electrons), which collide with electrons in the patient's tissue. This process converts mass into energy in the form of two emitted gamma-ray photons. The two gamma-ray photons are directed in opposite directions and are detected by radiation detectors surrounding the imaging object (patient) as substantially simultaneous radiation detection events, which define a line of response (LOR) between them.

[0003] A disadvantage of conventional PET is that the exact location of the annihilation event cannot be determined, which reduces the resolution of the constructed patient image. In time-of-flight (TOF) PET, a small time difference (or no time difference) between two essentially simultaneous detection events is used to further locate the positron along the LOR. In this context, the concept of TOF simply means that for each electron-positron annihilation event, the exact time at which each of the coincident photons is detected is noted to calculate the time difference. Since the closer photon will arrive at its detector first, the difference in arrival time helps to more accurately determine the location of the annihilation event along the LOR. In order to provide very accurate localization operations, gamma rays should be detected with sub-nanosecond time resolution. However, the spatial variation of the inherent time delay across the detector arrays typically used for PET applications (also known as "skew") will produce substantial errors if not corrected.

[0004] Conventional gamma-ray detectors typically include a scintillation detector or scintillator (e.g., a scintillator crystal or an array of scintillators) coupled to a photodetector (e.g., an array of photosensitive elements). The scintillator scintillates (i.e., emits) a flash of light in response to an incoming particle (e.g., a high-energy gamma photon or positron, etc.). The emitted photons are captured by the photodetector and then read out by dedicated readout electronics based on the time and position at which the scintillation photons were captured. The temporal and spatial position of the incident gamma photon in the scintillator can be determined. In this context, skew is an individual fingerprint for these detectors that depends on the intrinsic properties of the scintillator crystal and its manufacturing process, but also on external factors such as supply voltage and temperature variations. In addition, the main contribution to the skew comes from the electronic channels connected to the array of photosensitive elements. Wave impedance and cable length cause time delays, and without additional buffers with corresponding individual time delays, clock distribution along the individual channels cannot be achieved.

[0005] Even though TOF PET can reduce acquisition time and improve the effective sensitivity of the PET acquisition by calculating the time difference between two coincident photons, for a system with a coincidence timing resolution of 200 ps (FWHM), an uncorrected time skew of 20 ps to 40 ps can degrade system performance.

[0006] Therefore, there is a need for temporal deskew correction for TOF PET applications in order to provide more accurate medical imaging devices (eg, PET / CT or PET / MR) with improved temporal resolution.

[0007] US2016 / 0187497 A1 discloses a calibration method based on the following operations: determining a set of coincidentally emitted scintillation photons based on a recorded spatial intensity distribution of scintillation photons emitted by a scintillator array, determining a center of gravity position and cumulative energy for the set of coincidentally emitted scintillation photons, and performing cluster analysis based on the determined center of gravity position and cumulative energy. When imaging is performed using this method, improved resolution is achieved. Summary of the Invention

[0008] The present invention aims to provide an alternative calibration method and calibration module for a gamma ray detector to provide fast and accurate calibration of the gamma ray detector.Another object of the present invention is to provide a gamma ray detector, a medical imaging device and a computer program.

[0009] In a first aspect of the present invention, a calibration method for a gamma-ray detector is provided, the gamma-ray detector comprising: a pixelated scintillator array having a plurality of scintillator pixels, the pixelated scintillator array being configured to emit scintillation photons at photon conversion locations in response to incident gamma rays, and

[0010] a pixelated photodetector array, i.e., a pixelated PD array, having a plurality of PD pixels coupled to the pixelated scintillator array, the pixelated PD array being used to determine the spatial intensity distribution of the scintillation photons,

[0011] The PD pixel is subdivided into a plurality of PD sub-pixels.

[0012] The calibration method comprises the following steps:

[0013] enabling a first PD subpixel coupled to a first scintillator pixel of the pixelated scintillator array,

[0014] enabling a second PD pixel coupled to a second scintillator pixel of the pixelated scintillator array, wherein the enabled second PD pixel is positioned adjacent to the PD pixel to which the enabled first PD sub-pixel belongs,

[0015] recording, by the enabled first PD sub-pixel, scintillation photons emitted at a photon conversion location located in the first scintillator pixel to obtain a first PD sub-pixel detection signal at a first time point,

[0016] recording, by the enabled second PD pixel, shared scintillation photons converted from the first scintillator pixel and travelling into the second scintillator pixel to obtain a second PD pixel detection signal at a second time point,

[0017] estimating a first time skew between the first time point and the second time point, and

[0018] The first time skew is corrected.

[0019] In another aspect of the present invention, a calibration module for a gamma-ray detector is provided, the gamma-ray detector comprising:

[0020] a pixelated scintillator array having a plurality of scintillator pixels configured to emit scintillation photons at photon conversion locations in response to incident gamma rays,

[0021] a pixelated PD array having a plurality of PD pixels coupled to the pixelated scintillator array, and the pixelated PD array is configured to determine a spatial intensity distribution of the scintillation photons,

[0022] The PD pixel is subdivided into a plurality of PD sub-pixels.

[0023] The calibration module includes:

[0024] Logger, which is configured to:

[0025] recording, by an enabled first PD sub-pixel, scintillation photons emitted at a photon conversion location in a first scintillator pixel of the pixelated scintillator array to obtain a first PD sub-pixel detection signal at a first point in time, the enabled first PD sub-pixel being coupled to the first scintillator pixel, and

[0026] recording, by an enabled second PD pixel, shared scintillation photons resulting from photon conversion in the first scintillator pixel and traveling into a second scintillator pixel of the pixelated scintillator array to obtain a second PD pixel detection signal at a second point in time, the enabled second PD pixel being coupled to the second scintillator pixel and positioned adjacent to the PD pixel to which the enabled first PD sub-pixel belongs, and

[0027] A processing module configured to:

[0028] estimating a first time skew between the first time point and the second time point, and

[0029] The first time skew is corrected.

[0030] In another aspect of the present invention, a gamma-ray detector is proposed, which includes: a pixelated scintillator array having a plurality of scintillator pixels, and the pixelated scintillator array is configured to emit scintillation photons at photon conversion positions in response to incident gamma rays; a pixelated PD array having a plurality of PD pixels coupled to the pixelated scintillator array, and the pixelated PD array is configured to determine the spatial intensity distribution of the scintillation photons, wherein the PD pixels are subdivided into a plurality of PD sub-pixels; and a calibration module according to the above aspects of the present invention.

[0031] In another aspect of the present invention, a medical imaging device is provided, the medical imaging device including the gamma-ray detector disclosed herein. In yet another aspect of the present invention, a computer program including a program code module is provided, and when the computer program is executed on a processor of a calibration module according to the above aspect, on a processor of a gamma-ray detector according to the above aspect, or on a processor of a medical imaging device according to the above aspect, the program code module is used to cause the calibration module, the gamma-ray detector, or the medical imaging device to perform the steps of the calibration method described herein. In another aspect of the present invention, a non-transitory computer-readable recording medium is provided, in which a computer program product is stored, and when the computer program product is executed by a processor, the calibration method disclosed herein is performed.

[0032] Preferred embodiments of the invention are defined in the dependent claims. It should be understood that the claimed calibration module, gamma ray detector, medical imaging device, computer program and medium have similar and / or identical preferred embodiments as the claimed calibration method and the embodiments defined in the dependent claims.

[0033] The present invention is based on the idea of ​​using the concept of light sharing between neighboring scintillator pixels (ie scintillator elements of a pixelated scintillator array) to estimate the coincident timing and skew (ie temporal skew) between them by exploiting their optical light sharing properties.

[0034] For the purpose described, gamma rays hit an array of scintillation pixels at a certain position (i.e., a photon conversion position) and cause a light flash (i.e., a scintillation photon) to be emitted in the scintillator. A PD array (which includes a photosensitive element) is coupled to the scintillator array so that the scintillation photons emitted in one element of the scintillator array (i.e., one scintillator pixel) are distributed over a plurality of PD pixels of the pixelated PD array, i.e., light is shared between adjacent scintillator pixels and adjacent PD pixels. Preferably, each PD pixel is subdivided into a plurality of PD sub-pixels. The pixelated PD array allows recording the spatial intensity distribution of the emitted scintillation photons. Gamma absorption occurs at a photon conversion position located in a first scintillator pixel of the pixelated scintillator array. This first scintillator pixel is coupled to an enabled first PD sub-pixel of the pixelated PD array. Due to light sharing, a coincidence event is generated and read out in an adjacent second PD pixel, which is coupled to a second scintillator pixel of the pixelated scintillator array.

[0035] The operation of recording scintillation photons emitted at a photon conversion position located in a first scintillator pixel by an enabled first PD sub-pixel allows a first PD sub-pixel detection signal at a first point in time to be obtained. The operation of recording shared scintillation photons resulting from photon conversion in the first scintillator pixel and travelling into the second scintillator pixel by an enabled second PD pixel allows a second PD pixel detection signal at a second point in time to be obtained, the enabled second PD pixel being coupled to the second scintillator pixel. Thus, two different detection signals (a first PD sub-pixel detection signal and a second PD pixel detection signal) are obtained at typically two slightly different points in time and are used for temporal skew estimation. Therefore, the present invention proposes a calibration routine based on local light sharing between adjacent scintillator crystals.

[0036] In the described context, "enable" means turning on the corresponding PD pixel or PD sub-pixel (by means of a corresponding trigger as explained below). Since each PD pixel is preferably subdivided into a plurality of PD pixels, the PD pixel will be (fully) enabled if all PD sub-pixels of the PD pixel are enabled. Therefore, the present invention is based on estimating the time skew between the detection event at an enabled first PD sub-pixel and the detection event at a fully enabled second PD pixel, the fully enabled second PD pixel being positioned adjacent to the (first) PD pixel to which the first PD sub-pixel belongs. In other words, the fully enabled PD pixel acts as a "reference PD pixel" for the PD sub-pixel. Since the term "time skew" refers to the time difference between two channels (the PD sub-pixel channel and the reference PD pixel channel), it can also be referred to as a "time offset". However, the term "time offset" generally refers to any fixed reference. In the current context, the corresponding PD pixel acts as a reference for the corresponding PD sub-pixel, but is not itself a fixed reference. Therefore, the time difference between the channels is called "time skew" rather than "time offset".

[0037] Preferably, the step of correcting the first time skew comprises delaying the first PD sub-pixel detection signal and / or the second PD pixel detection signal to reduce the time skew. This means that the first PD sub-pixel detection signal is preferably delayed relative to the second PD pixel detection signal, or vice versa. In this context, the time skew is typically in the range of tens of picoseconds. Therefore, a precise electronic delay unit is required. Such a delay unit can be implemented by additional buffers (actually combined with RC components) and additional tracks (such as metal traces in silicon designs, which also create RC components).

[0038] According to another embodiment, the pixelated PD array is connected to a tunable delay unit array of tunable PD sub-pixel delay units and tunable PD pixel delay units. The calibration method may include the following steps: setting the delay time for a first tunable PD sub-pixel delay unit connected to the first PD sub-pixel and the delay time for a second tunable PD pixel delay unit connected to the second PD pixel to correct the first time skew. Preferably, each PD pixel and each PD sub-pixel is connected to a tunable PD pixel delay unit and a tunable PD sub-pixel delay unit, respectively. Therefore, the calibration method may further include the following steps: setting the delay time for all tunable PD pixel delay units and the delay time for all tunable PD sub-pixel delay units, respectively. The delay units may be any controllable delay units known in the art of processing electrical signals, and these delay units may be controlled by any type of suitable processor. These delay units may achieve precise delays within a certain picosecond range.

[0039] As mentioned above, the skew shift depends primarily on the intrinsic properties of the scintillator pixels, which can be determined by their manufacturing process. Furthermore, if the gamma-ray detector is operated in a magnetic field, the skew shift can also depend on environmental data, such as temperature, the supply voltage of the corresponding PD pixels and PD subpixels, or the magnetic field. It is known in the art that magnetic fields can affect PET image resolution.

[0040] In order to take these situations into account, the first time skew can also be corrected based on a time skew model, which relates the environmental data to a time offset for correcting the first time skew. This is of particular interest for PET systems operating in a magnetic field or air-cooled PET systems with potential temperature changes exceeding 10K, as this can lead to large drifts of the time skew (tens of ps). Therefore, in order to significantly improve the timing performance, it is advantageous to correct the time skew not only based on delaying the PD sub-pixel detection signal relative to the PD pixel detection signal, but also based on a time skew model that is preferentially based on a plurality of environmental data measured by one or more sensors. Therefore, the calibration method includes directly correcting temperature drifts (as well as voltage drifts, etc.). The model can be a linear model or a non-linear model and can also be based on a reference table of environmental data and reference time skews.

[0041] Enabling one PD sub-pixel (i.e., the first PD sub-pixel) and one PD pixel (i.e., the second PD pixel) allows estimating only the temporal skew shift between them. Typically, pixelated PD arrays known in the prior art include many PD pixels, which are preferably subdivided into a plurality of PD sub-pixels. Therefore, estimating only the temporal skew shift between the first PD sub-pixel and the second PD pixel is insufficient; rather, more temporal skew shifts must be estimated to correct for all of these temporal skew shifts.

[0042] For this purpose, the PD pixels are preferably connected to corresponding PD pixel triggers, and the PD sub-pixels are preferably connected to corresponding PD sub-pixel triggers. The triggers are configured to enable or disable the corresponding PD pixels and PD sub-pixels.

[0043] According to an embodiment, the calibration method may further comprise the following steps: enabling the plurality of PD sub-pixels and the plurality of PD pixels by means of corresponding PD sub-pixel triggers of the plurality of PD sub-pixels and corresponding PD pixel triggers of the plurality of PD pixels to form a predetermined pattern of enabled PD sub-pixels and PD pixels and disabled PD sub-pixels and PD pixels. This allows obtaining a plurality of time skews between the corresponding PD sub-pixels and the corresponding PD pixels, all of which are then used for the correction process as described above. The predetermined pattern may be programmed by applying a corresponding enable matrix to the PD pixel array. Preferably, each PD pixel is connected to a corresponding PD pixel trigger, and each PD sub-pixel is connected to a corresponding PD sub-pixel trigger to independently enable or disable each PD pixel and PD sub-pixel. If the pattern is stored in an FPGA connected to the PD pixel trigger and the PD sub-pixel trigger (or generated by the FPGA), the predetermined pattern may be further changed within a few milliseconds.

[0044] Switching between several predetermined patterns of enabled and disabled PD subpixels and PD pixels allows for addressing light sharing issues between a variety of different PD subpixels and PD subpixels. Consequently, a range of temporal skews can be estimated and used in the correction process discussed above. This further improves the calibration method and results in smaller temporal skews for all PD subpixels in the gamma-ray detector.

[0045] According to another embodiment, the calibration method further comprises the following steps:

[0046] enabling a third PD pixel coupled to a third scintillator pixel of the pixelated scintillator array, wherein the enabled third PD pixel is positioned adjacent to the PD pixel to which the enabled first PD sub-pixel belongs,

[0047] recording, by the enabled third PD pixel, shared scintillation photons converted from the first scintillator pixel and travelling into the third scintillator pixel to obtain a third PD pixel detection signal at a third time point,

[0048] estimating a second time skew between the third time point and the first time point, and

[0049] The first time skew and the second time skew are averaged to obtain an average time skew.

[0050] This has the following technical effect: not only does one PD pixel (i.e., the second PD pixel) serve as a reference PD pixel for the first PD sub-pixel, but another PD pixel (i.e., the third PD pixel) also serves as a reference PD pixel for the first PD sub-pixel. Therefore, before correcting two (or more) temporal skew shifts, the two (or more) temporal skew shifts can be estimated and averaged. In the context described, the second PD pixel and the third PD pixel are positioned adjacent to the (first) PD pixel to which the first PD sub-pixel belongs. Therefore, the temporal skew shifts can be averaged by using, for example, a horizontal reference PD pixel and a vertical reference PD pixel. It is also possible to choose to correct the first temporal skew shift and the second temporal skew shift separately before calculating the average of the corresponding corrected temporal skew shifts.

[0051] According to an embodiment, the calibration method may further include the following steps:

[0052] An incident gamma ray is emitted by a point source remote from the gamma ray detector and / or, if the pixelated scintillator array includes a radionuclide, by a radionuclide, wherein the scintillation photons are emitted at a photon conversion location in response to the incident gamma ray.

[0053] The advantage of using a point source (e.g., a Na-22 positron source) is that a large number of photon conversion count statistics are obtained, which can speed up the calibration process. However, the calibration process does not require coincidence events created by an external positron emitter placed and adjusted in the field of view (FOV) of the gamma ray detector. If the scintillator array exhibits a radioactive material such as Lu-176 (which is commonly used as a scintillator for gamma ray detectors), it is sufficient to calibrate the gamma ray detector using inherent self-emission. Another option is to even use cosmic gamma rays. The advantage of a background scan (without an external point source) is that it can be performed without user intervention whenever the PET system is not actively recording patient data. Therefore, in a clinical environment, it is preferred to perform the calibration method several times a day between medical examinations of the patient.

[0054] According to an embodiment, the calibration method may further include the following steps:

[0055] comparing the estimated first time skew, second time skew and / or average time skew with a reference table of time skews, and

[0056] If the estimated time skew is outside the acceptance window, a warning is generated.

[0057] If the gamma-ray detector's temporal skew is too great, accurate medical imaging of the patient cannot be achieved. Therefore, a warning is generated, indicating to medical personnel in a clinical setting that the PET system may be producing erroneous imaging results. This improves medical examinations of patients, particularly if calibration methods are routinely performed to permanently monitor the function of the gamma-ray detector. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter. In the accompanying drawings:

[0059] Figure 1 A schematic diagram of a medical imaging device according to the present invention is shown;

[0060] Figure 2 shows a schematic diagram of a gamma ray detector according to the present invention;

[0061] Figure 3 shows a flow chart illustrating a calibration method according to the present invention;

[0062] Figure 4 shows a schematic diagram of a calibration module according to the present invention;

[0063] Figure 5 A schematic diagram showing the electrical connections of the PD pixel array to the delay unit and the trigger;

[0064] Figure 6 shows a block diagram illustrating a programmable time delay of a PD pixel subdivided into several PD sub-pixels;

[0065] Figure 7 Schematic diagram showing top views of eight different reference patterns of enabled PD pixels and PD sub-pixels and disabled PD pixels and PD sub-pixels;

[0066] Figure 8 Illustrations showing flood maps of different enabled and disabled 2×2 PD pixel matrices;

[0067] Figure 9 shows energy and timing histograms for PD sub-pixel tilt calibration using lutetium background scanning;

[0068] Figure 10 shows a color map of the PD sub-pixel temporal skew estimation results from a Lu background scan for a 2×2 PD pixel matrix before (left) and after (right) correction;

[0069] Figure 11shows the energy and timing histograms for PD sub-pixel tilt calibration using a Na-22 point source in singles mode;

[0070] Figure 12 shows a color map of PD sub-pixel temporal skew estimation results from a Na-22 scan in single shot mode for a 2×2 PD pixel matrix before (left) and after (right) correction;

[0071] Figure 13 a schematic diagram showing a top view of different coincident test patterns of enabled and disabled PD sub-pixels;

[0072] Figure 14 shows energy and timing histograms for PD subpixel tilt calibration using a Na-22 point source in coincidence mode;

[0073] Figure 15 shows a color map of PD subpixel temporal skew estimation results from a Na-22 scan in coincidence mode for a 2×2 PD pixel matrix before (left) and after (right) correction;

[0074] Figure 16 shows a schematic diagram of a sensor patch with autonomous time skew correction, wherein the control loop is implemented within the patch FPGA;

[0075] Figure 17 A diagram illustrating the effect of temperature on time ramp is shown. DETAILED DESCRIPTION

[0076] Figure 1 A schematic diagram of a medical imaging device 50 according to the present invention is shown. The medical imaging device 50 may be, for example, a PET device, a PET / CT device, a PET / MR device, or a SPECT / PET / CT device. The illustrated medical imaging device 50 includes a gamma ray detector 100, which is preferably calibrated using a calibration method or calibration module according to the present invention.

[0077] The illustrated medical imaging device 50 further comprises a tunable patient support 55 for supporting a patient 60 being treated with the aid of the medical imaging device 50. A PET device typically detects particles, in particular gamma rays, emitted within the patient 60 to be imaged. For example, the patient 60 may be administered a radioactive tracer substance, and the gamma ray detector 100 calibrated according to the present invention may be used to determine the spatial location of the tracer substance within the patient 60. Thus, a PET device typically comprises, for example, a PET device. Figure 1The ring-shaped gamma-ray detector 100 is shown for detecting two simultaneously impinging gamma rays.

[0078] exist Figure 1 This is more clearly illustrated in the portal of FIG. The portal shows a cutout of a ring-shaped gamma-ray detector 100 and the head of a patient 60, where the radiotracer material is localized after being administered to the patient 60. In PET imaging, the radiopharmaceutical administered to the patient 60 emits positrons, each of which in turn annihilates with an electron of the surrounding imaging subject in an electron-positron annihilation event to produce two oppositely directed 511 keV gamma rays 90. In conventional PET imaging, these two oppositely directed gamma rays 90 are detected by the gamma-ray detector 100 surrounding the imaging subject as substantially simultaneous radiation detection events that define a line of response (LOR) therebetween. In time-of-flight (TOF) PET, the small time difference (or lack of time difference) between these two substantially simultaneous detection events is used to further localize the positrons along the LOR. Compared to PET without time-of-flight application, the difference in arrival times helps to more precisely localize annihilation events along the LOR, since the closer photons will reach its detector first.

[0079] Figure 1 The entrance in FIG. 1 also illustrates that the gamma ray detector 100 includes a pixelated scintillator array 110 having a plurality of scintillator pixels for emitting scintillation photons at photon conversion locations in response to incident gamma rays 90. Furthermore, the ring-shaped gamma ray detector 100 also includes a pixelated photodetector (PD) array 120 having a plurality of PD pixels coupled to the pixelated scintillator array 110 for determining a spatial intensity distribution of the scintillation photons. The photodetector pixels may be photomultiplier tubes or silicon avalanche photodiodes (SPADs).

[0080] Even if current TOF PET methods are found to provide adequate (e.g., sub-nanosecond) temporal resolution for a given radiation detector pixel at a given time, spatial variations across the detector array (also known as "skew") can introduce substantial errors. Skew is an inherent fingerprint of the individual scintillator pixels of the scintillator array 110, but also depends on external factors such as supply voltage and temperature variations. This can substantially reduce the effective temporal resolution of the PET detector array.

[0081] The following figures explain how the aforementioned limitations can be overcome through a fast and efficient calibration process to identify and correct PD sub-pixel tilt of a pixelated photodetector array.

[0082] Figure 2 FIG. 1 is a schematic diagram of a gamma-ray detector 100 according to the present invention. Figure 2 A shows a schematic diagram of a side view of the gamma ray detector 100. The gamma ray detector 100 includes a pixelated scintillator array 110 having a plurality of scintillator pixels. Figure 2 This is clearly illustrated in Figure B. Figure 2 B shows a schematic diagram of a top view of the pixelated scintillator array 110 . The gamma-ray detector 100 further includes a pixelated PD array 120 having a plurality of PD pixels 122 . Figure 2 C shows a schematic diagram of a top view of the pixelated PD array 120. Each PD pixel can be subdivided into four corresponding PD sub-pixels. However, it should be understood that this is only exemplary; the PD pixel can also be subdivided into more or fewer PD sub-pixels.

[0083] exist Figure 2 As can be seen in FIG. 1A , the PD pixels 122 of the pixelated PD array 120 are coupled to the pixelated scintillator array 110. The coupling can be achieved by any technique known in the art, in particular by coupling based on a glue layer (not shown), in particular a light-guiding glue, between the pixelated scintillator array 110 and the pixelated PD array 120. The coupling allows determining the spatial intensity distribution of scintillation photons emitted at photon conversion locations in the plurality of scintillator pixels in response to the incident gamma ray 90.

[0084] The calibration method according to the present invention may include the steps of emitting the incident gamma rays 90 through a point source 150 (e.g., a Na-22 positron source) remote from the gamma ray detector 100 and / or, if the pixelated scintillator array 110 includes a radionuclide (e.g., lutetium (Lu-176)), emitting the incident gamma rays through the radionuclide. Thus, for a calibration procedure with a point source 150, the point source may be placed in the middle of the ring-shaped gamma ray detector 100 (see Figure 2 ), during medical imaging, a patient is positioned within the ring-shaped gamma ray detector 100. An advantage of performing the calibration procedure using a radioactive scintillator material (e.g., Lu-176) is that the external point source 150 does not need to be positioned and adjusted for the calibration procedure.

[0085] The calibration method according to the present invention comprises the following steps: enabling a first PD sub-pixel 124a, the first PD sub-pixel 124a being coupled to a first scintillator pixel 112a of the pixelated scintillator array 110. In addition, the calibration method according to the present invention further comprises the following steps: enabling a second PD pixel 122b, the second PD pixel 122b being coupled to a second scintillator pixel 112b of the pixelated scintillator array 110, wherein the enabled second PD pixel 122b is positioned adjacent to the PD pixel to which the enabled first PD sub-pixel 124a belongs. In said context, the expression "adjacent" means that these PD pixels are positioned next to each other in the plane of the pixelated PD array 120. It should be noted that only in Figure 2 In A, enabled (on) PD pixels and PD sub-pixels are illustrated for simplicity.

[0086] The photon conversion in the first scintillator pixel 112a not only causes scintillation photons to be detected by the first PD sub-pixel 124a, but also causes shared scintillation photons to travel into the second scintillator pixel 112b. The fully enabled second PD pixel 122b then detects these shared scintillation photons to obtain a second PD pixel detection signal, while simultaneously obtaining a first PD sub-pixel detection signal based on the scintillation photons recorded by the enabled first PD sub-pixel 124a. It should be noted that in the context described, "fully enabled" means that the second PD pixel 122b (which is subdivided into four PD sub-pixels (see Figure 2 C)) is fully enabled, ie, all four PD sub-pixels of the second PD pixel are enabled.

[0087] In addition, the calibration method further comprises the following steps: estimating a first time skew between a first time point at which the first PD sub-pixel detection signal is obtained and a second time point at which the second PD pixel detection signal is obtained. This first time skew is derived from the "skew" mentioned above. Therefore, the intrinsic fingerprint properties of the first scintillator pixel 112a and the second scintillator pixel 112b give rise to different time points at which the enabled first PD sub-pixel 124a and the enabled second PD pixel 122b detect the scintillation photons. This time skew is a problem for TOF PET, where a high temporal resolution is required to determine the position of the annihilation event along the LOR (see Figure 1 For further explanation). For this purpose, the calibration method further comprises the step of correcting the first time skew. This allows reducing the time skew, ie the time difference between the first time point and the second time point.

[0088] Figure 3A flow chart illustrating the discussed calibration method is shown. In step S10, a first PD sub-pixel 124a is enabled, the first PD sub-pixel 124a being coupled to a first scintillator pixel 112a of the pixelated scintillator array 110. In step S20, a second PD pixel 122b is enabled, the second PD pixel 122b being coupled to a second scintillator pixel 112b of the pixelated scintillator array, wherein the enabled second PD pixel 122b is positioned adjacent to the PD pixel to which the enabled first PD sub-pixel 124a belongs (see Figure 2 ). In step S30, the scintillation photons emitted at the photon conversion position positioned in the first scintillator pixel 112a are recorded by the enabled first PD sub-pixel 124a to obtain a first PD sub-pixel detection signal. In step S40, the shared scintillation photons obtained from the photon conversion in the first scintillator pixel 112a and traveling to the second scintillator pixel 112b are recorded by the enabled second PD pixel 122b to obtain a second PD pixel detection signal. In step S50, the first time skew of the recorded photons of the enabled first PD sub-pixel 124a to the recorded photons of the enabled second PD pixel 122b is estimated as the time skew between the first time point of obtaining / deriving the first PD sub-pixel detection signal and the second time point of obtaining / deriving the second PD pixel detection signal. In step S60, the first time skew is corrected to reduce the time skew (time offset) between these signals. Refer to the following Figure 5 A more detailed description is given of this step S60.

[0089] Therefore, in summary, light sharing with neighboring scintillator pixels is an essential feature of the present invention, as it allows estimating the timing resolution of the coincidence and the skew shift between them simply by exploiting their optical light sharing properties. For this reason, light sharing with neighboring scintillator crystals is preferably not less than 10% to ensure appropriate timing correlation values.

[0090] Figure 4 A schematic diagram of a calibration module for a gamma ray detector according to the present invention is shown. The gamma ray detector is preferably the same as that already referenced. Figure 2 The gamma-ray detector discussed above is identical and thus comprises a pixelated scintillator array having a plurality of scintillator pixels, the pixelated scintillator array being configured to emit scintillation photons at photon conversion locations in response to incident gamma rays. The gamma-ray detector preferably further comprises a pixelated PD array having a plurality of PD pixels coupled to the pixelated scintillator array, and the pixelated PD array being configured to determine a spatial intensity distribution of the scintillation photons, wherein the PD pixels are subdivided into a plurality of PD sub-pixels (e.g., each having four PD sub-pixels).

[0091] like Figure 4The calibration module shown includes a recorder 210 and a processing module 220. The recorder 210 records the scintillation photons emitted at the photon conversion position in the first scintillator pixel of the pixelated scintillator array through the enabled first PD sub-pixel to obtain a first PD sub-pixel detection signal at a first time point, and the enabled first PD sub-pixel is coupled to the first scintillator pixel. In addition, the recorder records the shared scintillation photons obtained from the photon conversion in the first scintillator pixel and traveling to the second scintillator pixel of the pixelated scintillator array through the enabled second PD pixel to obtain a second PD pixel detection signal at a second time point, and the enabled second PD pixel is coupled to the second scintillator pixel and is positioned adjacent to the PD pixel to which the enabled first PD sub-pixel belongs. Based on the detection signal, the processing module 220 estimates a first time skew between the first time point and the second time point and corrects the first time skew.

[0092] Figure 5 Schematic diagram showing the electrical connection of the PD pixel array to the delay unit and the trigger. Figure 2-4 As explained above, the first PD sub-pixel 124a and the second PD pixel 122b are enabled. Therefore, the first PD sub-pixel 124a can be electrically connected to the first PD sub-pixel trigger 144a, and the second PD pixel 122b can be connected to the second PD pixel trigger 142b. It should be understood that Figure 5 The diagrams shown are for illustrative purposes only. Figure 5 In the embodiment of the present invention, only one PD sub-pixel is connected to the PD sub-pixel trigger, and only one PD pixel is connected to the PD pixel trigger. Preferably, all PD pixels and PD sub-pixels of the pixelated PD array are connected to corresponding PD pixel triggers or PD sub-pixel triggers, each of which is configured to enable and disable the corresponding PD pixel and PD sub-pixel.

[0093] exist Figure 5 1 and 2. It is also exemplarily shown that the first PD sub-pixel 124a can be electrically connected to the first tunable PD delay unit 134a, and the second PD sub-pixel 122b can be electrically connected to the second tunable PD delay unit 132b. It should be understood that Figure 5 The diagram shown is merely exemplary, and all PD pixels and PD sub-pixels are preferably connected to corresponding delay units.

[0094] Correcting the first time skew may include delaying the first PD sub-pixel detection signal and / or the second PD detection signal to reduce the first time skew. This may be accomplished by setting a delay time for a first tunable PD sub-pixel delay unit 134a connected to the first PD sub-pixel 124a and a delay time for a second tunable PD pixel delay unit 132b connected to the second PD pixel 122b.

[0095] References so far Figure 1-5 All methods and processes discussed are based on the following assumptions: partially enabling one PD pixel (only one of the four PD sub-pixels is enabled) and fully enabling the neighboring PD pixels for said PD pixel (all four PD sub-pixels are enabled). It should be understood that it is the minimum possible number of enabled PD pixels and enabled PD sub-pixels that explains the basic feature of the present invention of light sharing with neighboring PD pixels. In general, as hereinafter referred to Figure 6 As explained, more than one PD sub-pixel is enabled.

[0096] Figure 6 A block diagram illustrating the programmable time delay of a PD pixel that is subdivided into several PD sub-pixels is shown. Each PD pixel is subdivided into a plurality of PD sub-pixels, which in turn are Figure 6 The left column of FIG. 1 includes multiple single-photon avalanche diodes (SPADs). Thus, each PD subpixel preferably includes an array of SPADs, where multiple SPADs in a row can form a SPAD row. The SPAD row delay unit 136 and the PD subpixel delay unit 134 are preferably independently controlled. The programmable delay unit for the SPAD row (left), the programmable delay unit for the PD subpixel (center), and the programmable delay unit for the PD pixel (i.e., the tunable PD pixel delay unit 132) (right) create the basis for multi-layer skew correction for the PET gamma-ray detector.

[0097] Figure 7 Schematic diagram showing a top view of eight different reference patterns for enabled and disabled PD pixels and PD subpixels. These reference patterns can be used to estimate individual PD subpixel temporal skew for an example matrix of a 2×2 pixelated scintillator array (not shown out of plane) and a pixelated PD array 120 with a 4×4 PD subpixel area. The coding indicates whether an individual PD pixel 122 or an individual PD subpixel 124 is enabled or disabled (bright = enabled / on; dark = disabled / off).

[0098] This is explained exemplarily with reference to the reference pattern in the upper left, where the first PD sub-pixel 124a belonging to the PD pixel in the lower right is enabled. In addition, the second PD pixel 122b in the lower left is fully enabled (all four PD sub-pixels are enabled). This allows for an exemplified estimation of a first time skew (a time skew between a first point in time at which a first PD sub-pixel detection signal (of the first PD sub-pixel 124a) is obtained and a second point in time at which a second PD pixel detection signal (of the second PD pixel 122b) is obtained). Thus, as Figure 7 The reference pattern shown in the upper left of allows estimation of the temporal skew of the first PD sub-pixel 124a with reference to the second PD pixel 122b (thereby acting as a reference pixel).

[0099] In addition to this first PD sub-pixel 124a and this second PD pixel 122b, the reference pattern in the upper left corner also includes an enabled third PD pixel 122c, which is coupled to a third scintillator pixel of the pixelated scintillator array 110 (not shown out of plane), wherein the enabled third PD pixel 122c is positioned adjacent to the PD pixel to which the enabled first PD sub-pixel 124a belongs. The calibration method discussed with reference to the previous figures may also include the following steps: recording, by the enabled third PD pixel 122c, shared scintillation photons converted from the first scintillator pixel 112a and traveling into the third scintillator pixel. This allows the third PD pixel detection signal at the third time point to be obtained. Therefore, the second time skew can be estimated as the time difference between the third time point and the first time point. This allows the first time skew and the second time skew to be averaged to obtain an average time skew. Therefore, it is preferred to average the temporal skew estimation values ​​for the first PD sub-pixel 124a by using the horizontal reference (second PD pixel 122b) and the vertical reference (third PD pixel 122c), which is the basis for the temporal skew correction.

[0100] all in all, Figure 7 The reference pattern on the upper left of FIG allows estimation of the average temporal skew for the first PD sub-pixel 124 a. As another PD sub-pixel (i.e., the fourth PD sub-pixel 124 d is also enabled in the reference pattern), the reference pattern allows estimation of not only the temporal deviation for the first PD sub-pixel 124 a but also (through the same process as discussed above) the temporal deviation for the fourth PD sub-pixel 124 d.

[0101] Also like Figure 7As shown, eight different patterns can be used to generate temporal skew estimates for both horizontal and vertical neighbor PD pixels. Thus, these eight patterns can be used to extract 16 temporal skew estimates (one for each PD sub-pixel). Since only the temporal skew differences of the four PD sub-pixels to the reference PD pixel are of interest, the mean-free horizontal and vertical estimates can be averaged. Additionally, poor fits can be detected by applying a limit on the maximum permissible difference.

[0102] The test pattern can be programmed by applying the corresponding enable matrix to the individual PD pixels and PD sub-pixels. If the test pattern is stored in the local sensor tile FPGA (or generated by the local sensor tile FPGA), such changes can be implemented in a few milliseconds. It should be understood that, in general, the PD sub-pixels 124 can cover any fraction of the corresponding PD pixels 122.

[0103] Figure 8 Illustrations of flood maps for different enabled and disabled 2×2 PD pixel matrices are shown, with the square root of the counts plotted to visualize dynamic compression. The left image shows a flood map with all four PD pixels enabled. This configuration is used for normal PET operation. The middle image shows Figure 7 Flood map of one of the eight test patterns shown, where only two of the four PD pixels (1 / 4 of the area) are enabled (here, top left and bottom right). The distortion of the pixels in the middle image is due to the lower detection intensity of these partially enabled PD pixels compared to the fully enabled PD pixels in the flood map on the left. The right image shows a flood map of essentially the same test pattern as the middle image, with the difference that the events in the fully enabled PD pixels (top right and bottom left) are masked out. Therefore, only two regions of interest are retained for further PD sub-pixel skew shift estimation to extract only events due to light sharing.

[0104] Typically, the calibration process does not require coincident events created by positron emitters. A "single" gamma (e.g., self-emission from Lu-176 in the case of lutetium-based scintillators, or generated by an external single gamma emitter (CS-137 with 661 keV)) as well as cosmic gamma rays are sufficient. The advantage of background scanning is that it can be performed without user intervention whenever the PET system is not actively recording patient data. Additionally, data processing and updating of PD sub-pixel skew estimation results can also be achieved by decentralized processing on a local FPGA at the sensor tile or module level.

[0105] Figure 9The left panel shows the PD subpixel of interest (e.g., Figure 2 and Figure 6 The energy histogram of the first PD sub-pixel 124a in FIG1 and the energy histogram of the first PD sub-pixel 124b in FIG1 due to the adjacent fully enabled PD pixels (eg, Figure 2 and Figure 6 The corresponding energy histogram due to light sharing on the second PD pixel 122b in the image is shown. The histogram of the PD subpixel is energy-gated, as only the appropriate portion of the lutetium background spectrum is selected. The right figure shows the corresponding timing histogram of the gamma event after energy clipping, which provides an estimate of the time skew of the PD subpixel relative to the reference pixel. As can be seen, a time skew of approximately -0.6ns (-600ps) is recorded.

[0106] Figure 10 Color maps of PD subpixel temporal skew estimation results based on a lutetium background scan for a 2×2 PD pixel matrix before (left) and after (right) correction are shown. These values ​​(e.g., "13 ps" for the fourth PD subpixel 124 d) represent the temporal skew shift of the PD subpixel to a reference PD pixel. Furthermore, these temporal skew shift estimation results are averaged using a horizontal reference (e.g., the third PD pixel 122 c for the fourth PD subpixel 124 d) and a vertical reference (e.g., the second PD pixel 122 b for the fourth PD subpixel 124 d). The value at the center of the corresponding PD pixel (e.g., "11 ps" for the second PD pixel 122 b) is the standard deviation obtained by averaging the corresponding temporal skew shifts of all four PD subpixels of the corresponding PD pixel. Furthermore, all four temporal skew shifts of the PD subpixels of the corresponding PD pixel cancel each other out, i.e., the sum of the four PD subpixel skew shifts of the corresponding PD pixel is (almost) zero.

[0107] Figure 11 Energy and timing histograms are shown for PD subpixel skew calibration using a Na-22 point source in singles mode. In this context, "singles mode" means emitting only one gamma photon. The following example shows that PD subpixel skew and PD pixel skew can also be estimated by decentralized individual processing of the reference point source at the tile level. This has the advantage of speeding up the calibration process due to faster counting statistics, but on the other hand (as already mentioned above) requires manual work to place the gamma point source.

[0108] Figure 12Color maps of PD subpixel temporal skew estimation results based on Na-22 scans in singles mode for a 2×2 PD pixel matrix before (left) and after (right) correction are shown. The example with a Na-22 point source is mainly for illustration, showing that subpixel skew calibration based on 511 keV single gamma photons emitted by the Na-22 point source gives essentially the same results as the lutetium-based estimation results when the corresponding energy is selected.

[0109] The difference from Na-22 in singles mode is that the calibration source must be placed in a predefined location (e.g., the field of view (FOV) of the gamma-ray detector) and system-based processing of coincidence events is required to determine the PD pixel and PD sub-pixel time skew estimation results for the coincidence-based calibration method.

[0110] Figure 13 Schematic diagram showing top views of different coincidence test patterns for enabled and disabled PD sub-pixels. Since the Na-22 point source is operated in coincidence mode, the patterns must be permuted to form 4×4=16 pattern combinations, and according to Figure 13 The leftmost test pattern in FIG1 has the first PD subpixel 124a, the second PD subpixel 124c, the third PD subpixel 124d, and the fourth PD subpixel 124d enabled. Thus, a complete set of 16 test pattern combinations can be formed, with one pattern formed for each overlapping side. In this context, "coincidence mode" means detecting and processing two overlapping 511keVs. The resulting energy peak for 511keV has a gain of 1 / 4 of the energy peak if only one PD subpixel is selected compared to a fully enabled PD pixel. Since this process does not require any light sharing between adjacent scintillator crystals here, it can be considered the "gold standard."

[0111] Figure 14 The energy histogram and timing histogram for the sub-pixel tilt calibration using a Na-22 point source in coincident mode are shown. The above energy histogram and timing histogram for the PD sub-pixel test pattern can be obtained in coincident mode using only one of the four PD sub-pixels in each PD pixel being enabled (see Figure 13 ). Exemplary curves may pertain to an enabled first PD sub-pixel 124a and an enabled second PD sub-pixel.

[0112] Figure 15Color maps of PD subpixel temporal skew estimation results from a Na-22 scan in coincidence mode for a 2×2 PD pixel matrix are shown before (left) and after (right) correction. The coincidence measurement results can be used to acquire a reference global temporal skew map, including module skew, patch skew, PD pixel skew, and PD subpixel skew. A mapping matrix between PD subpixel skew and PD pixel skew estimated from the coincidence mode and lutetium background scans can be generated to scale subsequent measurements based on the lutetium background.

[0113] Figure 16 A schematic diagram of a sensor patch 300 with autonomous temporal skew correction is shown, wherein the control loop is implemented within the patch FPGA. In this context, the sensor patch 300 includes several sensor dies on a printed circuit board with an FPGA and other infrastructure on the back side of the printed circuit board. Each sensor die has four PD pixels, and each PD pixel has four PD sub-pixels, as shown in the figure (e.g., Figure 2 C) shown.

[0114] The present invention may include a feedback loop to identify and correct for tile-based time offsets. This can be Figure 16 As can be seen in FIG, sensors 230 and 240 are used to provide environmental data. PD sub-pixel time skew also depends on external factors, such as supply voltage and temperature variations. Therefore, the calibration method may further comprise the step of correcting the time skew by setting a delay time based on said environmental data (e.g., supply voltage or temperature). This may be accomplished by reading the environmental data obtained from sensors 230 and 240 and applying a model-based time skew correction based on these environmental data. Thus, the control loop may also identify and correct for tile-based time offsets, which are typically temperature-dependent and voltage-dependent and can originate from the reference block's buffer 250.

[0115] Figure 17 A diagram illustrating the effect of temperature on time skew is shown. The left graph illustrates how increasing the temperature from 31°C to 54°C shifts the time skew from approximately 900 ps to a larger value of approximately 1080 ps. The right graph illustrates that a linear model with a 7.7 ps / K value in this example can be used to correct for time skew. For many applications, it is important to use these models to correct for time skew. In particular, air-cooled PET systems with potential variations exceeding 10 K can experience a 77 ps time skew shift within a ring-shaped gamma-ray detector, which can significantly degrade timing performance.

[0116] Although the present invention has been illustrated and described in detail in the drawings and the foregoing description, such illustration and description should be considered illustrative or exemplary rather than restrictive; the present invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention by studying the drawings, the disclosure, and the claims.

[0117] In the claims, the word "comprising" does not exclude other elements or steps, and the word "a" or "an" does not exclude a plurality. A single element or other unit may fulfill the functions of several items recited in the claims. The fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

[0118] The computer program may be stored / distributed on suitable non-transitory media, such as optical storage media or solid-state media supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

[0119] Any reference signs in the claims should not be construed as limiting the scope.

Claims

1. A calibration method for a gamma ray detector (100), the gamma ray detector comprising: a pixelated scintillator array (110) having a plurality of scintillator pixels for emitting scintillation photons at photon conversion locations in response to incident gamma rays (90), and a pixelated photodetector array, i.e., a pixelated PD array, having a plurality of PD pixels (122) coupled to the pixelated scintillator array (110), the pixelated PD array being used to determine the spatial intensity distribution of the scintillation photons, The PD pixel (122) is subdivided into a plurality of PD sub-pixels (124). The calibration method comprises the following steps: enabling a first PD subpixel (124a) coupled to a first scintillator pixel (112a) of the pixelated scintillator array (110), enabling a second PD pixel (122b) coupled to a second scintillator pixel (112b) of the pixelated scintillator array, wherein the enabled second PD pixel (122b) is positioned adjacent to the PD pixel to which the enabled first PD sub-pixel (124a) belongs, recording, by the enabled first PD sub-pixel (124a), scintillation photons emitted at a photon conversion location located in the first scintillator pixel (112a) to obtain a first PD sub-pixel detection signal at a first time point, The enabled second PD pixel (122b) records the shared scintillation photons converted from the first scintillator pixel (112a) and traveling to the second scintillator pixel (112b) to obtain a second PD pixel detection signal at a second time point, estimating a first time skew between the first time point and the second time point, and The first time skew is corrected.

2. The calibration method according to claim 1, in, Correcting the first time skew includes delaying the first PD sub-pixel detection signal and / or the second PD pixel detection signal to reduce the first time skew.

3. The calibration method according to claim 2, in, The pixelated PD array is connected to a tunable delay unit array of tunable PD sub-pixel delay units (134) and tunable PD pixel delay units (132), and The calibration method includes the following steps: setting a delay time for a first tunable PD sub-pixel delay unit (134a) connected to the first PD sub-pixel (124a) and a delay time for a second tunable PD pixel delay unit (132b) connected to the second PD pixel (122b) to correct the first time skew.

4. The calibration method according to claim 3, The following steps are also included: Environmental data is read from at least one sensor to correct the first time skew by setting the delay time based on the environmental data.

5. The calibration method according to claim 4, in, The environmental data includes one or more of the following: temperature, a supply voltage of the second PD pixel (122b) or the first PD sub-pixel (124a), or a magnetic field, and The first time skew is corrected based on a time skew model, which associates the environmental data with a time offset for correcting the first time skew.

6. The calibration method according to any one of the preceding claims, in, The PD pixels are connected to corresponding PD pixel triggers (142), and the PD sub-pixels are connected to corresponding PD sub-pixel triggers (144), Among them, the calibration method also includes the following steps: enabling the several PD sub-pixels and the several PD pixels through the corresponding PD sub-pixel triggers of the several PD sub-pixels and the corresponding PD pixel triggers of the several PD pixels to form a predetermined pattern of enabled PD sub-pixels and PD pixels and disabled PD sub-pixels and PD pixels.

7. The calibration method according to claim 6, Also includes: Switching between a plurality of predetermined patterns of enabled PD sub-pixels (124) and PD pixels (122) and disabled PD sub-pixels and PD pixels, The steps according to any one of claims 1 to 5 are performed for each predetermined pattern.

8. The calibration method according to claim 7, The following steps are also included: enabling a third PD pixel (122c) coupled to a third scintillator pixel of the pixelated scintillator array (110), wherein the enabled third PD pixel (122c) is positioned adjacent to the PD pixel to which the enabled first PD sub-pixel (124a) belongs, The enabled third PD pixel (122c) records the shared scintillation photons converted from the first scintillator pixel and traveling to the third scintillator pixel (112b) to obtain a third PD pixel detection signal at a third time point, estimating a second time skew between the third time point and the first time point, and The first time skew and the second time skew are averaged to obtain an average time skew.

9. The calibration method according to any one of claims 1 to 5 and 7 to 8, The following steps are also included: An incident gamma ray (90) is emitted by a point source (150) remote from the gamma ray detector (100), and / or if the pixelated scintillator array (110) includes a radionuclide, by a radionuclide, wherein the scintillation photon is emitted at a photon conversion location in response to the incident gamma ray (90).

10. The calibration method according to claim 8, The following steps are also included: comparing the estimated first time skew, second time skew and / or average time skew with a reference table of time skews, and If the estimated time skew is outside the acceptance window, a warning is generated.

11. A calibration module for a gamma ray detector (100), the gamma ray detector comprising: a pixelated scintillator array (110) having a plurality of scintillator pixels configured to emit scintillation photons at photon conversion locations in response to incident gamma rays (90), a pixelated PD array having a plurality of PD pixels (122) coupled to the pixelated scintillator array (110), and the pixelated PD array is configured to determine a spatial intensity distribution of the scintillation photons, The PD pixel (122) is subdivided into a plurality of PD sub-pixels (124). The calibration module includes: A recorder (210) configured to: recording scintillation photons emitted at a photon conversion location in a first scintillator pixel (112a) of the pixelated scintillator array (110) by an enabled first PD sub-pixel (124a), the enabled first PD sub-pixel being coupled to the first scintillator pixel (112a), to obtain a first PD sub-pixel detection signal at a first point in time, and recording, by an enabled second PD pixel (122b), a shared scintillation photon converted from a photon in the first scintillator pixel (112a) and travelling into a second scintillator pixel (112b) of the pixelated scintillator array (110), to obtain a second PD pixel detection signal at a second point in time, the enabled second PD pixel being coupled to the second scintillator pixel (112b) and positioned adjacent to the PD pixel to which the enabled first PD sub-pixel (124a) belongs, and A processing module (220) configured to: estimating a first time skew between the first time point and the second time point, and The first time skew is corrected.

12. A gamma ray detector (100), comprising: a pixelated scintillator array (110) having a plurality of scintillator pixels configured to emit scintillation photons at photon conversion locations in response to incident gamma rays (90), a pixelated PD array having a plurality of PD pixels (122) coupled to the pixelated scintillator array and configured to determine a spatial intensity distribution of the scintillation photons, wherein the PD pixels (122) are subdivided into a plurality of PD sub-pixels (124), and The calibration module according to claim 11.

13. A medical imaging device (50) comprising the gamma ray detector (100) according to claim 12.

14. A computer program product comprising a program code module, which, when the computer program product is executed on a processor of a calibration module according to claim 11, on a processor of a gamma-ray detector (100) according to claim 12, or on a processor of a medical imaging device (50) according to claim 13, causes the calibration module, the gamma-ray detector (100) or the medical imaging device (50) to perform the steps of the calibration method according to any one of claims 1 to 10.

Images