3D scatter distribution estimation

Abstract

Systems and methods of estimating 3D TOF scatter include acquiring 3D TOF data, determining 2D TOF data from the first TOF data, determining a first estimated scatter based on the second TOF data, reconstructing a first estimated image based on the first estimated scatter and the second TOF data, determining an attenuated unscattered true coincidence based on the first estimated image, determining a second estimated scatter based on the first TOF data and the attenuated unscattered true coincidence, and reconstructing an image of the object based on the first TOF data and the second estimated scatter.

Description

Background Technology

[0001] In traditional positron emission tomography (PET) imaging, a radiopharmaceutical tracer is typically introduced into the patient via radial artery injection. The radioactive decay of the tracer produces positrons, which eventually encounter electrons and are annihilated. This annihilation produces two photons traveling in approximately opposite directions.

[0002] A detector ring surrounding the body detects emitted photons, identifies "coincidences," and reconstructs a PET image based on these coincidences. A coincidence is identified when two detectors positioned on opposite sides of the body detect the arrival of two photons within a specific coincidence time window. Because the two "coincidence" photons propagate in approximately opposite directions, the positions of the two detectors determine the response line (LOR), along which annihilation events may occur.

[0003] A "true" coincidence indicates the detection of two coincident photons arising from a single annihilation event located on the LOR between the two detectors. A "random" coincidence indicates two coincident photons not arising from the same annihilation event. A "scattered" coincidence is a type of true coincidence where the two coincident photons originate from the same annihilation event, but the annihilation event is not located on the LOR between the two detectors because one or both photons interact and scatter within the body or medium.

[0004] Conventional PET scanners detect all matches, regardless of whether they are unscattered true matches, random matches, or scattered true matches. Since only unscattered true matches represent spatial information about the distribution of tracers within the body, random and scattered matches should be resolved before and / or during image reconstruction. Software- and / or hardware-based methods can be used to estimate random matches and subtract them from the detected matches.

[0005] Model-based methods can be used to estimate scattering coincidences. For example, single-scattering simulation (SSS) modeling is currently used to model scattering coincidences with a locus of light (LOR) contained in the direct axial plane (referred to as 2D scattering). Reverse single-slice rebinding can then be used to replicate the 2D scattering to the corresponding tilted plane (i.e., 3D scattering). In the case of PET scanners with short axial ranges, this mapping provides a reasonable estimate of the 3D scattering due to the relatively similar geometric responses in the axial and tilted planes.

[0006] These geometric responses are dissimilar in long-axis field-of-view PET scanners, and the mapping described above does not provide a suitable estimate of 3D scattering. Therefore, estimation of 3D scattering requires accurate modeling of scattering both in the tilted plane and in the direct plane. This modeling can be excessively time-consuming. Even if such modeling is performed, a tail fit is required for the estimated 3D scattering for each tilted plane to account for multiple scattering and out-of-field scattering. In the case of patient imaging, this tail fit may be inaccurate in the tilted plane due to the low number of coincidences caused by high attenuation and reduced scanner efficiency. Attached Figure Description

[0007] Figure 1A and 1B The detection of conformity according to some embodiments is shown.

[0008] Figure 2 A PET detector according to some embodiments is shown.

[0009] Figure 3 This is a block diagram of a conformity detection system according to some embodiments.

[0010] Figure 4 This is a block diagram of a system for reconstructing images from PET data according to some embodiments.

[0011] Figure 5 Includes flowcharts of a process for estimating scattering according to some embodiments.

[0012] Figure 6 The LOR within a multi-ring PET scanner is shown in some embodiments.

[0013] Figure 7 A 2D LOR within a multi-ring PET scanner is shown in some embodiments.

[0014] Figure 8 Low polar angle LOR in a multi-ring PET scanner is shown in some embodiments.

[0015] Figure 9 This is a block diagram of a system for reconstructing images from 3D time-of-flight (TOF) PET data according to some embodiments.

[0016] Figure 10 This is a block diagram of a system for reconstructing images from 3D TOF PET data according to some embodiments.

[0017] Figure 11 This is a block diagram of a PET / CT imaging system according to some embodiments. Detailed Implementation

[0018] The following description is provided to enable anyone in the art to make and use the described embodiments. Various modifications will be apparent to those skilled in the art.

[0019] Some embodiments use 2D scattering estimation to estimate residual 3D TOF scattering, and use residual 3D TOF scattering to estimate 3D scattering within the 3D TOF data. In short, a subset of the 3D TOF data is determined, including the plane associated with a 0-degree polar angle (i.e., the direct axial plane) and an additional tilted plane associated with lower polar angles. 2D scattering estimation is performed (e.g., using single-scattering simulation (SSS) with reverse single-slice reclassification of the tilted planes). An “unbiased” estimated image is reconstructed based on the 3D TOF data, 2D scattering estimation, mean stochastic estimation, attenuation correction factor, and normalization correction factor. The estimated image is then forward-projected to obtain the true data of unattenuated, non-3D TOF scattering.

[0020] The unscattered 3D TOF ground truth data is not corrected for attenuation, and residual 3D TOF scattering is obtained by subtracting the attenuated 3D TOF unscattered ground truth data from the normalized net ground truth data (where net ground truth = (original 3D TOF data - estimated mean random)). The residual 3D TOF scattering is then smoothed using a Gaussian filter or other advanced filtering method and used to reconstruct the image using the original 3D TOF data. Alternatively, 3D scattering can be determined from modeling using the SSS method and fitted to the residual 3D TOF scattering to obtain a 3D scattering estimate for use during image reconstruction.

[0021] Advantageously, the embodiments can leverage the efficiency of model-based 2D scattering estimation, particularly compared to the resource consumption required for model-based 3D scattering estimation. The embodiments also avoid tail fitting of scattering data from high polar planes. Furthermore, the method described herein is based on an efficient and reliable 3D model for true unscattered fits rather than on a 3D scattering model.

[0022] Figure 1A and 1B The detection of conformity according to some embodiments is shown. Figure 1A This is an axial view of the aperture 105 of the scanner 100 and the imaging object 110 disposed therein. The imaging object 110 may include a human body, a phantom, or any other suitable object. Figure 1B yes Figure 1A The image shows a radial view of scanner 100 and object 110. In the example shown, scanner 100 consists of any number (eight in this example) of adjacent and coaxial rings of detectors 150. Each detector 150 may include any number of scintillator crystals and electrical transducers.

[0023] Assume that annihilation events 120, 130, 140, and 142 occur at various locations within object 110. As described above, the injected tracer produces positrons, which are annihilated by electrons to produce two 511 keV gamma photons that travel in approximately opposite directions. Figure 1A and Figure 1A Each annihilation event, as represented in the diagram, results in a detected match. As mentioned above, true matches represent valid image data, while scattered and random matches represent noise.

[0024] A coincidence is detected when a pair of detectors receive two gamma photons within a coincidence time window, as determined based on the calculated arrival times of the two gamma photons at their respective detectors. Event 120 is associated with a true coincidence because event 120 results in the reception of two gamma photons within the coincidence time window, and because the location of annihilation event 120 lies on LOR 125, which connects the locations of the detectors that received the two gamma photons.

[0025] Event 130 is associated with scattering coincidence because even if the two gamma photons generated by event 130 are detected within the coincidence time window, the location of annihilation event 130 is not located on LOR 135 connecting the locations of the two photons. This could be due to a change in orientation of at least one of the two gamma photons within object 110 caused by Compton (i.e., inelastic) or coherent (i.e., elastic) scattering.

[0026] Events 140 and 142 are two separate annihilation events that result in the detection of random coincidences. (As...) Figure 1B As shown, one of the photons generated by event 140 is absorbed in object 110, and one of the photons generated by event 142 escapes detection by any detector 150 of scanner 100. The remaining photons happen to be detected within the coincidence time window, even though no annihilation event occurs on LOR 145 at the location where the coincidence photons were received.

[0027] Since only true unscattered coincidences indicate the location of annihilation events, random coincidences and scattered coincidences are often subtracted from the acquired PET data or otherwise used to correct the acquired PET data during PET image reconstruction.

[0028] As described herein, the direct axial plane is perpendicular to the axis of scanner 100. Therefore, LORs 135 and 145 lie within the direct axial plane of scanner 100 because their associated detectors 150 lie in the same axial plane (i.e., within the same ring of the eight detector rings of scanner 100). Conversely, LOR 125 lies in an inclined plane because its two associated detectors 150 are not located within the same detector ring of scanner 100. The inclined plane of LOR 125 can be considered a “low polar angle” plane due to its small angle of inclination relative to the axis perpendicular to scanner 100.

[0029] Typically, a PET detector includes one or more scintillation elements and one or more electrical transducers. The scintillation element generates photons with energies of several electron volts in response to receiving 511 keV photons generated by an annihilation event. The electrical transducer converts the low-energy photons generated by the scintillation element into electrical signals. According to some embodiments, the electrical transducer may include, for example, a silicon-based photomultiplier (SiPM), a photomultiplier tube (PMT), or a semiconductor-based detector.

[0030] Figure 2 A detector 200 according to some embodiments is shown. The detector 200 consists of eight microblocks, two of which are in the axial direction and four in the meridional direction. In one example, the microblocks comprise a grid of 5×5 lutetium silicate (LSO) scintillation crystals with dimensions of 3.2 mm × 3.2 mm × 20 mm. The microblocks can be coupled to a 4×4 array of SiPMs for receiving photons from them and generating electrical signals based thereon. Thus, the detector 200 comprises 200 crystals in rows of 10 crystals in the axial direction and 20 crystals in the meridional direction. Embodiments are not limited to the above description of the detector 200.

[0031] According to some embodiments, scanner 100 is a long-axis field-of-view scanner, comprising 32 detectors in the axial direction and 38 detectors in the meridional direction. Thus, scanner 100 comprises 60,800 detector crystals, with rows of 80 detector crystals in the axial direction and rows of 760 detector crystals in the meridional direction. Embodiments are not limited to these specifications.

[0032] Figure 3 This is a block diagram of a coincidence detection system 300 according to some embodiments. System 300 includes scintillation units 310, 320, and 330, corresponding electrical transducer units 312, 322, and 332, and corresponding signal processing components 314, 324, and 334. A coincidence detection unit 340 receives signals from each of the signal processing components 314, 324, and 334.

[0033] Each scintillation unit 310, 320, and 330 may include one or more scintillation crystals. For example, each of scintillation units 310, 320, and 330 may include a microblock of 5×5 crystal elements, a macroblock of 2×2 microblocks, or a detector consisting of two macroblocks. Embodiments are not limited to any particular configuration or construction of scintillation units 310, 320, and 330.

[0034] Each electrical transducer unit 312, 322, and 332 may include one or more PMTs, SiPMs, etc. The number of electrical transducers in each unit 312, 322, and 332 may be less than, equal to, or greater than the number of crystal elements in each scintillation unit 310, 320, and 330. According to some embodiments, the electrical transducer unit includes a 4×4 array of SiPMs for each microblock of 5×5 crystal elements in its corresponding scintillation unit.

[0035] Signal processing components 314, 324, and 334 receive electrical signals from corresponding electrical transducer units 312, 322, and 332, and perform signal processing to, for example, determine whether the signal represents a photon detection event, perform signal unpileing through pile-up rejection and / or correction methods, and associate the photon detection event with a specific detector crystal of the scintillation units 310, 320, and 330. Signal processing components 314, 324, and 334 can perform any suitable function and exhibit any suitable implementation.

[0036] The coincidence detection unit 340 receives all photon detection events that have passed energy qualification, referred to as single events, and identifies such event pairs that occur within a coincidence time window. The coincidence detection unit 340 may also include delay logic that delays the apparent arrival time of one event for each comparison before performing coincidence detection. As a result, the delay logic does not detect any actual true coincidences. The “delayed coincidences” detected by the delay logic can be used to estimate mean random coincidences, as described below.

[0037] The coincidence detection unit 340 outputs data specifying each identified event pair and labeling each pair as a true coincidence or a delayed coincidence. For either type of coincidence, the output data also specifies the two detector crystals that received the photon detection event containing the coincidence. In the case of TOF PET imaging, the data for each detected coincidence also includes the difference between the arrival times of the two photons in the coincidence. This difference can be used to more accurately estimate the specific location of the annihilation event corresponding to the occurrence along the LOR.

[0038] A sine graph is a data array that stores the coincidences detected in a single plane over time. A delayed sine graph stores data associated with detected delayed coincidences, while a true sine graph stores data associated with detected true coincidences. The sine graph represents each LOR (Landing of Response) of each detected coincidence as the angle and displacement relative to a center point located on the scanner axis.

[0039] The sinogram includes a row of LORs containing a specific azimuth angle φ. Each of these rows corresponds to a one-dimensional parallel projection of the tracer distribution at different coordinates. The sinogram stores the location of the LOR for each coincidence, such that all LORs passing through a single point in the volume are plotted as a sine curve in the sinogram. The TOF sinogram includes a third dimension specifying the TOF information for each coincidence.

[0040] The plane represented by the sine plot can include a direct axial plane or an inclined plane. "2D" data refers to a set of sine plots for various direct axial planes and may include data from low-polarity inclined planes. As is known in the art, the latter data can be binned into one or more existing sine plots of direct axial planes and / or simulated sine plots of direct axial planes. As is also known in the art, an image can be reconstructed solely from the sine plots of the direct axial planes, while ignoring the inclined planes.

[0041] Figure 4 An imaging system 400 according to some embodiments is illustrated. Each component of system 400 can be implemented by any suitable combination of hardware and software. In some embodiments, one or more components can be implemented by a single software application.

[0042] System 400 includes a detector 410 as part of a scanner and a corresponding scintillator 420. The scintillator 420 can be configured with respect to... Figure 2 The description describes the composition of a single crystal. The embodiments are not limited to scintillator-based detectors. Direct conversion detectors (e.g., CZT and TIBr) may also be used in some embodiments.

[0043] Detector 410 detects gamma photons 435 emitted from volume 430. Systems for facilitating the emission of gamma photons from a volume are known in the art, particularly with respect to PET imaging described herein. As described above, the crystal of scintillator 420 receives gamma photons 435 and emits photons in response. Detector 410 receives photons, and each detector 410 generates an electrical signal based on the energy of the received photons and its own characteristic photoelectric response distribution.

[0044] The detector signal processing unit 440 receives electrical signals generated by each detector 410 and performs signal processing to, for example, determine whether the signal represents a photon detection event, perform signal destacking through stacking rejection, determine the event energy, and determine the event time. The detector signal processing unit 440 can perform any suitable function and exhibit any suitable implementation.

[0045] Within a given time period, the coincidence detection unit 445 receives all photon detection events that have passed energy identification (e.g., between 435 and 585 keV) from all detectors 410 of the scanner. Based on the reception time of each photon detection event, unit 445 identifies pairs of photon detection events received within the coincidence time window and determines that each such pair corresponds to a true coincidence with associated LOR and energy. The coincidence detection unit 445 can also determine a TOF value representing the difference in reception time of the photon detection events for each pair of photon detection events. The coincidence detection unit 445 also uses delay logic to identify delayed coincidences as described above. For each plane, the coincidence detection unit 445 stores a sine graph representing each coincidence detected in that plane. These sine graphs, representing the direct axis and tilt planes, constitute 3D TOF data.

[0046] The random correction unit 450 can estimate the mean random coincidence for each crystal pair. For example, delayed coincidence can be used to estimate the single event rate for each crystal in a PET scanner. The single event rate is the rate at which a crystal detects valid (i.e., energy-qualified) photons during the scanning process. Next, for each crystal pair ( i, j Using a stochastic smoothing model to estimate the mean stochastic : ,in and It is the single-event rate of crystals i and j, and It conforms to the time window. Some techniques further base the estimated mean random on the delayed conformity count. The application is rescaled. The estimated mean is randomly used to correct the 3D TOF data so that the randomly corrected 3D TOF data includes only the net truth (i.e., the scattering coincidence and the unscattered true coincidence).

[0047] The scattering estimation unit 455 estimates 3D scattering based on stochastically corrected 3D TOF data. As will be described in detail below, the estimation may include estimating residual 3D TOF scattering using 2D scattering estimation, and estimating 3D scattering within the stochastically corrected 3D TOF data using residual 3D TOF scattering.

[0048] In some embodiments, the denoising unit 460 employs a Gaussian filter to smooth the estimated 3D scattering. According to other embodiments, the shape of the 3D scattering is modeled directly from the randomly corrected 3D TOF data, and the denoising unit 460 scales this shape based on the estimated 3D scattering. Finally, the reconstruction unit 465 performs a reconstruction algorithm to reconstruct the image based on the denoised estimated 3D scattering and the randomly corrected 3D TOF data output from the random correction unit 450.

[0049] Figure 5 This includes flowcharts of a process 500 for estimating 3D scattering according to some embodiments. Process 500 and other processes described herein can be performed using any suitable combination of hardware and software. The software program code embodying these processes can be stored by any non-transitory tangible medium, including hard disks, volatile or non-volatile random access memory, DVDs, flash drives, and magnetic tapes, and executed by any suitable processing unit, including but not limited to one or more microprocessors, microcontrollers, processing cores, and processor threads. Embodiments are not limited to the examples described below.

[0050] Initially, at S510, the object is scanned using a PET scanner known in the art. According to some embodiments, the object includes phantoms, such as, for example, a uniformly filled water cylinder. A radionuclide tracer is injected into the object prior to scanning. The radionuclide tracer may include any suitable tracer, such as, but not limited to, fluorodeoxyglucose (FDG). The scan may include a conventional static PET scan or a CBM scan, and generates 3D TOF data describing the delayed and true coincidences detected by the PET scanner during the scan as described above.

[0051] 3D TOF data includes sine plots of all planes acquired during the scan. Figure 6 This is a transverse view of a scanner 600 comprising sixteen coaxial detector rings. Dashed lines connect each "upper" detector to each of the "lower" detectors. Each dashed line represents a plane of the LOR, which can be represented by the corresponding sine curve within the 3D TOF data.

[0052] In S520, the 3D TOF data is corrected for random coincidences. In S520, any suitable technique can be used to estimate the mean random coincidence and to correct the 3D TOF data based on the estimated random coincidences. According to some embodiments, an estimated mean random sine wave is generated for each plane of the 3D TOF data, and in S510, the estimated mean random sine wave for the given plane is subtracted from the sine wave obtained for the given plane. S520 produces 3D TOF data representing the net truth (i.e., true coincidences and scattering coincidences).

[0053] In S530, 3D TOF data is reduced to 2D TOF data. Figure 7 The LOR (Line of Response) of the scanner 600 in the direct axial plane is shown. In some embodiments, S530 includes extracting only the TOF (Time of Flight) data (i.e., sine curves) representing the coincidences detected in the direct axial plane. Alternatively, Figure 8 It shows Figure 7 The direct axial plane together with the inclined plane having a low polar angle (i.e., the angle between the axial direction and the vertical direction).

[0054] S530 may include extracting representations in Figure 8 The TOF data detected in the direct axial plane and the tilted plane are consistent. The tilted plane can be selected such that the geometric response does not change significantly from the geometric response of adjacent direct segments, while also preserving sufficient count statistics to achieve reasonable scattering estimation. At this point, the data in the tilted plane can be interpolated to construct a sine plot of the direct axial "inter-plane" known in the art, thereby allowing conventional 2D algorithms to be applied to the 2D TOF data generated on the S530.

[0055] At this point, 2D SSS is performed on the 2D TOF data in S540 to determine the estimated scattering. Typically, according to one implementation of 2D SSS, an initial image is reconstructed from the 2D TOF data using the corresponding attenuation correction factor (e.g., CT-derived) and normalization factor, assuming no scattering. The scattering is estimated based on the initial image and the attenuation correction factor, and the image is reconstructed using the newly estimated scattering, as in the first iteration. This process continues in this manner until the scattering estimation converges.

[0056] Next, at S550, the estimated image is reconstructed based on the scattering estimated at S540 and the randomly corrected 2D TOF data. At S550, any suitable TOF reconstruction algorithm can be used, including but not limited to analytical methods such as filtered back projection (FBP) or inverse discrete Fourier transform (DIFT), which produce an “unbiased” image.

[0057] In S560, attenuated unscattered true coincidences are determined based on the estimated image. S560 may include orthographically projecting the estimated image onto each of several tomographic planes using a tomographic model (e.g., by calculating a line integral) into a set of unscattered true coincidences. Each set may include a sine plot of the unscattered true coincidences. This sine plot is then uncorrected for attenuation to generate attenuated unscattered true coincidences.

[0058] Based on the original 3D TOF data generated in S520, the stochastically corrected TOF data (i.e., net true) generated in S520, and the attenuated unscattered true conformance determined in S560, a 3D scattering estimate is determined in S570. For example, in S550, the stochastically corrected TOF data is normalized according to a normalization factor used to reconstruct the estimated image. The attenuated unscattered true conformance is then subtracted from the normalized stochastically corrected TOF data to produce the 3D scattering estimate.

[0059] In S580, the 3D scattering estimate is denoised. As is known in the art, denoising may include applying a Gaussian filter to smooth the 3D scattering estimate. In some embodiments, denoising includes modeling the shape of the 3D scattering directly from the randomly corrected 3D TOF data, and scaling the shape for each axial plane based on the 3D scattering estimate determined in S570.

[0060] In S590, the image is reconstructed based on the original 3D TOF data, the denoised 3D scattering estimate, and the estimated mean random coincidence, as is known in the art.

[0061] Figure 9 This is a block diagram illustrating process 500 according to some embodiments. As shown, 2D TOF data 910 is obtained from 3D TOF data 905. For clarity, it will be assumed that the 3D TOF data 905 is randomly calibrated, and therefore both the 3D TOF data 905 and the 2D TOF data 910 represent net true data.

[0062] A 2D scattering estimation 915 is performed using the 2D TOF data 910 to generate an estimated scattering 920. The estimated scattering 920 (and based on appropriate attenuation and normalization factors) is then used to apply a 3D TOF reconstruction 925 to the 3D TOF data 905 to generate an estimated image 930. The estimated image 930 is then orthographically projected 935 without attenuation correction to generate an attenuated, unscattered true image 940.

[0063] Normalization 945 is applied to the 3D TOF data 905 to generate a normalized net true 950. 3D TOF residual scattering 955 is calculated as the difference between the normalized net true 950 and the attenuated unscattered true 940. The 3D TOF residual scattering 955 is denoised 960 (e.g., using a Gaussian filter), and the denoised scattering 965 is used to reconstruct the 3D TOF data 905, producing a PET image 975.

[0064] Figure 10 This is a block diagram illustrating process 500 according to some embodiments. Figure 10 Similar to Figure 9In addition to using 3D scattering modeling 1060 to generate modeled scattering 1065 based on 3D TOF data 1005, the 3D TOF residual scattering 1055 generated as described above is denoised 1070 by scaling (i.e., fitting) the modeled scattering 1065 to the 3D TOF residual scattering 1055 for each axial plane. The denoised scattering 1075 is then used to reconstruct a PET image 1085 from the 3D TOF data 1005.

[0065] Figure 11 A PET / CT system 1100 is shown for performing one or more of the procedures described herein. Embodiments are not limited to system 1100.

[0066] System 1100 includes a stand 1110 defining an aperture 1112. As is known in the art, stand 1110 houses a PET imaging assembly for acquiring PET image data and a CT imaging assembly for acquiring CT image data. As is known in the art, the CT imaging assembly may include one or more X-ray tubes and one or more corresponding X-ray detectors.

[0067] PET imaging assemblies can include any number or type of detectors in any configuration known in the art. Typically, a detector includes one or more scintillation elements and one or more electrical transducers. The scintillation element generates photons with energies of several electron volts in response to receiving 511 keV photons generated by an annihilation event. LSO and yttrium lutetium silicate (LYSO) scintillators exhibit suitable stopping capabilities and rapid scintillation decay, and can be used in high count rate scenarios.

[0068] An electrical transducer converts low-energy photons generated by a scintillation element into an electrical signal. According to some embodiments, the electrical transducer may include a SiPM or a photomultiplier tube (PMT). Some embodiments employ a block detector comprising more scintillation elements than electrical transducers. In a block detector, multiple electrical transducers receive scattered low-energy photons resulting from the absorption of one of the photons annihilated at 511 keV. The relative outputs of the transducers are compared to determine the absorption location, which in turn identifies the scintillation element or crystal determined to have received the annihilated photon.

[0069] Injection system 1118 is operable to deliver a calibrated injection of FDG, iodine, or other radiopharmaceuticals to a patient before and / or during a PET scan. In some embodiments, injection system 1118 is integrated into bench 1110. Injection system 1118 may support a wired or wireless communication link with control system 1120 for receiving information on specified dose, injection protocol, and scan delay.

[0070] Bed 1115 and base 1116 are operable to move a patient lying on bed 1115 into and out of port 1112 before, during and after imaging. In some embodiments, bed 1115 is configured to translate over base 1116, and in other embodiments, base 1116 may move with bed 1115 or alternatively move from bed 1115.

[0071] The movement of the patient into and out of port 1112 allows for scanning of the patient using CT and PET imaging elements of gantry 1110. This scanning can be performed based on scanning parameters such as scan range and corresponding scan speed. According to some embodiments, during such scanning, bed 1115 and base 1116 can provide continuous bed movement and / or step-and-shoot motion.

[0072] The control system 1120 may include any general-purpose or special-purpose computing system. Therefore, the control system 1120 includes one or more processing units 1122 and a storage device 1130 for storing program code. The processing unit 1122 is configured to execute processor-executable program code to cause the system 1120 to operate as described herein. The storage device 1130 may include one or more fixed disks, solid-state random access memory, and / or removable media (e.g., thumb drives) mounted in a corresponding interface (e.g., a USB port).

[0073] Storage device 1130 stores the program code of control program 1131. One or more processing units 1122 can execute control program 1131 to control hardware components to inject radiopharmaceuticals into the patient, move the patient through orifice 1112 to the PET detector of bench 1110, and detect coincidence events occurring in the patient, in conjunction with PET system interface 1123, bed interface 1125, and injection interface 1127. Detected events can be stored as PET data 1134 in memory 1130.

[0074] One or more processing units 1122 may also execute control program 1131 to, in conjunction with CT system interface 1124, cause radiation sources within gantry 1110 to emit radiation into the body within aperture 1112 from different projection angles, and control corresponding detectors to acquire two-dimensional CT data. As described above, CT data can be acquired substantially simultaneously with PET data, and can be used for attenuation correction of simultaneously acquired PET data 1134 known in the art.

[0075] Storage device 1130 also includes a scattering estimation program 1132, which can be executed to estimate 3D scattering as described in detail above. Control program 1131 can also be executed to reconstruct PET data 1134 into PET image 1136 based on the estimated 3D scattering using any known or becoming known reconstruction algorithm.

[0076] PET image 1136 can be transmitted to terminal 1140 for display via terminal interface 1126. Terminal 1140 may include a display device and an input device coupled to system 1120. Terminal 1140 may receive user input for controlling the display of data, operation of system 1100, and / or the processing described herein. In some embodiments, terminal 1140 is a separate computing device, such as, but not limited to, a desktop computer, laptop computer, tablet computer, and smartphone.

[0077] Each component of system 1100 may include other elements necessary for its operation, as well as additional elements for providing functionality beyond that described herein. Each functional component described herein may be implemented in computer hardware, program code, and / or one or more computing systems that execute such program code known in the art. Such computing systems may include one or more processing units that execute processor-executable program code stored in a memory system.

[0078] Those skilled in the art will understand that various adaptations and modifications can be configured for the above embodiments without departing from the claims. Therefore, it should be understood that the claims may be implemented in ways different from those specifically described herein.

Claims

1. An imaging system, comprising: Positron emission tomography scanner, used to perform scanning of an object and generate first time-of-flight (TOF) data describing the true coincidences detected in the direct axial plane and the tilt plane during the scan; as well as Processing unit, used for: Second TOF data is determined from first TOF data, which describes true coincidences detected in the direct axial plane during the scan and true coincidences detected in a subset of the inclined plane during the scan. The first estimated scattering is determined based on the second TOF data; The first estimated image is reconstructed based on the first estimated scattering and the second TOF data; The attenuated unscattered true match is determined based on the first estimated image; The second estimated scattering is determined based on the first TOF data and the true match between the attenuated unscattered data; as well as An image of the object is reconstructed based on the first TOF data and the second estimated scattering.

2. The imaging system of claim 1, wherein the subset of the tilt planes includes tilt planes associated with low polar angles.

3. The imaging system of claim 1, wherein determining the first estimated scattering includes applying a 2D single scattering simulation algorithm to the second TOF data.

4. The imaging system according to claim 1, wherein the processing unit is configured to: The estimated mean associated with the first TOF data was determined to be randomly consistent. The determination of the second estimated scattering includes correcting the first TOF data based on the estimated mean random coincidence, normalizing the corrected first TOF data to generate a normalized net true coincidence, generating a decayed unscattered true coincidence for attenuated uncorrected unscattered true coincidence, and subtracting the decayed unscattered true coincidence from the normalized net true coincidence. The reconstruction of the object's image is based on first TOF data, estimated mean random coincidence, and second estimated scattering.

5. The imaging system according to claim 1, wherein the processing unit is configured to: Denoise the second estimated scattering. The reconstruction of the object's image is based on first TOF data and a second estimated scattering with denoised data.

6. The imaging system according to claim 5, wherein, Denoising of the second estimated scattering includes: The third estimated scattering is determined based on the first TOF data, and The third estimated scattering is scaled based on the second estimated scattering.

7. The imaging system of claim 6, wherein determining the third estimated scattering based on the first TOF data comprises applying a 3D single scattering simulation algorithm to the first TOF data.

8. A method for estimating scattering, comprising: Acquire first time-of-flight (TOF) data, which describes the true coincidences detected in the direct axial plane and the tilt plane during the scan of the object; Second TOF data is determined from first TOF data, which describes true coincidences detected only within a subset of the direct axial plane and the tilted planes, each of the tilted planes being associated with a low polar angle; The first estimated scattering is determined based on the second TOF data; The first estimated image is reconstructed based on the first estimated scattering and the second TOF data; The attenuated unscattered true match is determined based on the first estimated image; The second estimated scattering is determined based on the first TOF data and the true match between the attenuated unscattered data; as well as An image of the object is reconstructed based on the first TOF data and the second estimated scattering.

9. The method of claim 8, wherein determining the first estimated scattering comprises applying a 2D single scattering simulation algorithm to the second TOF data.

10. The method of claim 8, further comprising: The estimated mean associated with the first TOF data was determined to be randomly consistent. Determining the second estimated scattering includes correcting the first TOF data based on the estimated mean random coincidence, normalizing the corrected first TOF data to generate a normalized net true coincidence, correcting the attenuated unscattered true coincidence to generate attenuated unscattered true coincidence, and subtracting the attenuated unscattered true coincidence from the normalized net true coincidence. The image of the reconstructed object is based on first TOF data, estimated mean random coincidence, and second estimated scattering.

11. The method of claim 8, further comprising: Denoise the second estimated scattering. The image of the reconstructed object is based on first TOF data and a second estimated scattering with denoised data.

12. The method according to claim 11, wherein, Denoising the second estimated scattering includes: The third estimated scattering is determined based on the first TOF data, and The third estimated scattering is scaled based on the second estimated scattering.

13. The method of claim 12, wherein determining the third estimated scattering based on the first TOF data comprises applying a 3D single scattering simulation algorithm to the first TOF data.

14. A non-transitory computer-readable medium storing processor-executable process steps, which, when executed by a processing unit of a computing system, cause the computing system to: Acquire first time-of-flight (TOF) data, which describes the true coincidences detected in the direct axial plane and the tilt plane during the scan of the object; Second TOF data is determined from first TOF data, which describes true coincidences detected only within a subset of the direct axial plane and the tilted planes, each of the tilted planes being associated with a low polar angle; The first estimated scattering is determined based on the second TOF data; The first estimated image is reconstructed based on the first estimated scattering and the second TOF data; The attenuated unscattered true match is determined based on the first estimated image; The second estimated scattering is determined based on the first TOF data and the true match between the attenuated unscattered data; as well as An image of the object is reconstructed based on the first TOF data and the second estimated scattering.

15. The medium of claim 14, wherein determining the first estimated scattering comprises applying a 2D single scattering simulation algorithm to the second TOF data.

16. The medium of claim 14, when executed by a processing unit of a computing system, wherein the processor-executable process steps cause the computing system to: The estimated mean associated with the first TOF data was determined to be randomly consistent. The determination of the second estimated scattering includes correcting the first TOF data based on the estimated mean random coincidence, normalizing the corrected first TOF data to generate a normalized net true coincidence, generating a decayed unscattered true coincidence for attenuated uncorrected unscattered true coincidence, and subtracting the decayed unscattered true coincidence from the normalized net true coincidence. The reconstruction of the object's image is based on first TOF data, estimated mean random coincidence, and second estimated scattering.

17. The medium of claim 14, when executed by a processing unit of a computing system, wherein the processor-executable process steps cause the computing system to: Denoise the second estimated scattering. The reconstruction of the object's image is based on first TOF data and a second estimated scattering with denoised data.

18. The medium according to claim 17, wherein, Denoising of the second estimated scattering includes: The third estimated scattering is determined based on the first TOF data, and The third estimated scattering is scaled based on the second estimated scattering.

19. The medium of claim 18, wherein determining the third estimated scattering based on the first TOF data comprises applying a 3D single scattering simulation algorithm to the first TOF data.

Images