Multi-spectral zonal reconstruction for spect

Abstract

For SPECT reconstruction, zonal reconstruction is provided intra-modally. Rather than or in addition to using CT for structure, SPECT data from one energy may be used to provide structural information or zones for another energy. The zonal reconstruction may be combined with model-based multi-energy image formation. Multi-spectral, zonal reconstruction is used as an intra-modal imaging approach.

Description

MULTI-SPECTRAL ZONAL RECONSTRUCTION FOR SPECTBACKGROUND

[0001] The present embodiments relate to Single Photon Computed Tomography (SPECT). The reconstruction of a SPECT image is often difficult because the data is characterized by small signal rates and low signal-to- noise ratio. For SPECT imaging, the count rate is limited by the amount of a radionuclide (i.e., radioactive substance or radiotracer) that can be administered without harming the patient. Some radionuclides, typically alpha and beta emitting isotopes and their progeny (e.g., Tb-161 , Ho-166, Lu-177, I- 131 , Ac-225, Pb208, Pb212, At211 , 1-123, Ga-67 and Y-90) have complicated gamma energy spectra, including discrete or continuous energy spectra due to bremsstrahlung. Typically, the most prominent emission peak with the highest photon yields per decay is used in acquisition and subsequent reconstruction for which the image formation is optimized. As an example, PET is designed for 511 keV gamma from positron annihilation, and in SPECT different collimators are used for different emission energy ranges. For radionuclides with complicated spectra, this limits the efficiency and applicability of the imaging as emissions from other energies suitable for imaging are essentially discarded.

[0002] Multi-energy reconstruction may be used to benefit from the complicated energy spectra. Image blurring may still result due to degradation of resolution in the image formation process. A SPECT image does not necessarily provide structural information. Thus, a SPECT image is often evaluated with the help of an adjacent structural image. Computed tomography (CT) may be used for the structural image. In multi-modality imaging, the CT data may be used as part of the SPECT reconstruction, such as by reconstructing separately for different types of tissue or zones. The zonal reconstruction may improve accuracy or resolution. When zonal reconstruction is used for radionuclides with complicated energy spectra from alpha and beta emitting isotopes and their progeny (e.g., 1-123 and Lu-177), due to the inaccurate image formation model, enhanced resolution does not necessarily lead to improved quantitative accuracy. In fact, resolution enhancement may reduce the quantitative accuracy.

[0003] Zonal reconstruction using CT in combination with multiple energies has been provided (see U.S. Patent No. 10,395,353). This requires the separate CT imaging to identify zones.SUMMARY

[0004] By way of introduction, the preferred embodiments described below include methods, systems, instructions, and computer readable storage media for SPECT reconstruction. Zonal reconstruction is provided intra-modally.The zonal reconstruction may be combined with model-based multi-energy image formation. Rather than or in addition to using CT for structure, SPECT data from one energy may be used to provide structural information or zones for another energy. Multi-spectral, zonal reconstruction is used as an intra- modal imaging approach. A hybrid energy detector may be used to detect emissions at multiple energies, allowing for solving of the image formation for multiple energy ranges, assisting lower energy image formation with higher energy formation.

[0005] In a first aspect, a method is provided for SPECT reconstruction. SPECT emissions in a patient at first and second energies are detected. An image object is zonally reconstructed from the detected SPECT emissions. Zones of the zonal reconstruction for the second energy are determined from the SPECT emissions at the first energy. An image is generated from the image object.

[0006] In one approach, the radionuclide is Lu-177, 1-131 , TI-201 , Tb161 , or progeny of Ac225, but other radionuclides may be used. In one approach, the emissions from such radionuclides may be detected with a multi-camera SPECT system where a first camera of the multi-camera SPECT system is configured to detect at a lower energy than a second camera of the multicamera SPECT system. In another approach, the SPECT emissions at the first energy are detected with physical collimation, and the SPECT emissions at the second energy are detected with Compton scattering detection.

[0007] In another approach, the zonal reconstruction includes generating a zone map including the zones from a first reconstruction of the SPECT emissions at the first energy, forward projecting the SPECT emissions at the second energy separately for each of the zones, weighted summing of theforward projections, obtaining an update from a comparison of a result of the weighted summing to an object model, back projecting the update, and updating an estimate from the back projected update.

[0008] As another approach, zonal reconstructing includes generating a zone map of the zones from the SPECT emissions at the first energy. The first energy is lower than the second energy. For example, the SPECT emissions at the first energy spatially stabilize the zonal reconstruction from the SPECT emissions at the second energy where the second energy is a higher energy than the first energy.

[0009] According to another approach, the zonally reconstruction includes modeling scatter for the second energy. The second energy is higher than the first energy. The SPECT emissions at the second energy are segmented by the zones determined from the first energy. The zones are determined with a greater spatial resolution than an image object of the SPECT emissions at the second energy.

[0010] In a further approach, an initial reconstruction of the SPECT emissions at the first energy is performed. The zones are segmented from the initial reconstruction for the zonal reconstruction of the SPECT emissions at the second energy.

[0011] As yet another approach, the zonal reconstruction includes iterative reconstruction with reconstruction for the SPECT emissions at both the first and second energies in sequence within each iteration. For example, the iterative reconstruction includes a loop sequencing from the first energy to the second energy in each iteration. The first energy is lower than the second energy, and each iteration has an objective function including the SPECT emissions at both the first and second energies.

[0012] The image may be generated using various information. For example, the image is generated from the image object of the SPECT emissions at the second energy. As another example, the image is generated from the image object zonally reconstructed from the detected SPECT emissions at the second energy combined with another image object reconstructed from the detected SPECT emissions at the first energy. In yet another example, the image is generated from the image object zonallyreconstructed from the detected SPECT emissions at the second energy in combination with reconstructed from detected SPECT emissions at the first energy.

[0013] In a second aspect, a method is provided for SPECT reconstruction. SPECT emissions from a patient are detected. The SPECT emissions are from a radionuclide with multiple energies. An image object is reconstructed with multi-spectral zonal reconstruction. The multi-spectral zonal reconstruction includes the multiple energies and zones from a first of the multiple energies. An image is generated from the image object.

[0014] In an approach, reconstructing the image object includes iterative reconstruction where each iteration includes low energy to high energy successive projection operations.

[0015] As another approach, the zones are from the first energy. The first energy is a lower energy than a second energy. The multi-spectral zonal reconstruction reconstructs for the second energy with the zones from the first energy.

[0016] In a third aspect, a medical imaging system is provided for SPECT intra-modal zonal reconstruction. A detector arrangement is configured to detect emissions from the patient, the emissions are of different energies from a radiotracer. A processor is configured to reconstruct an object representing the patient from the detected emissions for at least two energy windows of the distributed energies. The reconstruction for a lower of the at least two energy windows is segmented, and the reconstruction for a higher of the at least two energy windows is zonal based on the segmentation from the reconstruction for the lower of the at least two energy windows. A display is configured to display an image of the reconstructed object.

[0017] According to one approach, the detector arrangement is a detector configured to detect with the at least two energy windows. In another approach, the detector arrangement is a physically collimated detector and a Compton scattering detector.

[0018] The present invention is defined by the following claims, and nothing in this section should be taken as a limitation on those claims. Further aspects and advantages of the invention are discussed below inconjunction with the preferred embodiments and may be later claimed independently or in combination.BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The components and the figures are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention. Moreover, in the figures, like reference numerals designate corresponding parts throughout the different views.

[0020] Figure 1 is a flow chart diagram of one embodiment of a method for SPECT reconstruction;

[0021] Figure 2 illustrates examples of different detector arrangements for detection of emissions at different energies;

[0022] Figure 3 illustrates an iterative zonal reconstruction according to one embodiment;

[0023] Figure 4 illustrates a sequential loop for multi-spectral reconstruction according to one embodiment; and

[0024] Figure 5 is a block diagram of one embodiment of a system for SPECT reconstruction for radionuclides with complicated spectra or spectra with multiple peaks.DETAILED DESCRIPTION OF THE DRAWINGS AND PRESENTLY PREFERRED EMBODIMENTS

[0025] Intra-modality SPECT zonal reconstruction is provided. Multienergy or multi-spectral and zone techniques used together improve both the quantitative accuracy and quality of SPECT images. Extending SPECT processes to include multi-energy tracers instead of a single energy tracer may not be strait forward. Rather than requiring extra-model information, zone information may be obtained from SPECT emissions at one energy for use in zonal reconstruction for another energy.

[0026] Image formation was designed for specifically a gamma emission range or a specific energy (e.g., 511 k eV- PET). By deploying a hybrid image formation spectral detector (see fig 11. US published patent application US 2022-0330909) a combination of multi-spectral and zone techniques may be provided. This covers a gap in a multi-modal approach using EMI fromCT / MR as the energy is used as a modality differentiator instead of CT / MR (e,g., a Y90 emits 511 keV and bremsstrahlung and Lutetium177 compound may have contamination with Lu176 and that too emits 511 keV, so a system can image the 511 keV image formation (i.e., PET) at very low count rate, and the Lu177 or Y90). Not only is one isotope that emits a complex spectrum, 100% pure isotope is not typically generated, which in addition makes a more complicated spectra as the spectral finger print depends on the production mechanism. This combination multi-energy and zone approach attempts to make sense of the data by comprehensively solving the image formation equation for multiple image formation’s optimized from the energy ranges, and assisting the lower energy image formation with the higher energy image formation, that typically doesn’t have good spatial resolution of yield.

[0027] A combination of zonal reconstruction from multiple energy emissions with multi-spectral reconstruction simultaneously addresses three important problems. One is the inaccurate model of image formation process in iterative reconstruction, especially for radionuclides with complicated energy spectra (e.g., Lu-177, 1-123, Ga-67, and Y-90). Another is the image blurring due to degradation of resolution in the image formation process. By combining zonal reconstruction and model-based multi-energy image formation model, SPECT image reconstruction is improved. Zonal reconstruction provides more accurate segmentation of critical organs, and improved quantitative accuracy in the critical organs provides more accurate dosimetry. The third is use of intra-modal emissions at one energy, such as a lower energy with greater spatial resolution, to provide the zones. The requirement for a CT scan is removed. Resolution and quantitative accuracy are improved without requiring CT or other extra-modal information for spatial structure. Lower resolution from collimation (e.g., electric) at higher energies may be improved.

[0028] In reconstructing for multiple energy windows, separate projections for model-based scatter correction are provided for each of the different zones. This combination for iterative reconstruction is processing intensive. To reduce the processing, the forward projection for the model-based scatter correction may be handled for the total image rather than zones. The zone-based scatter correction may be used in earlier iterations and total image object-based scatter correction may be used in later iterations. This different strategy for zonal forward projection in different iterations achieves both high accuracy and fast reconstruction. In alternative approaches, other multienergy reconstruction approaches are used.

[0029] Single photon image reconstruction is provided with combined image formation methods operating at different emission energies. Multiple image formation models are combined in one joint reconstruction as an extension of both extra-modal imaging (the extension is using intra modal for zones) and multi-emission imaging. This combination, for example, enables Compton reconstruction using zonal reconstruction of low energy emissions at better quality than possible without use of, e.g., electron kinematics.

[0030] Figure 1 shows one embodiment of a method for SPECT reconstruction. Multi-energy image formation (e.g., model-based multi-energy image formation) is combined with zonal reconstruction. Alternatively, multiple energies are used, one for segmentation and the other for zonal reconstruction. The segmentation of zones relies on SPECT data at one or more energies rather than extra modal information (e.g., ratherthan CT). Resolution enhancement and improved quantitative accuracy may be simultaneously achieved without requiring CT imaging.

[0031] The method of Figure 1 is implemented using the system of Figure 5, a processor, a computer, a SPECT imager, and / or another device. For example, a SPECT imager performs act 100. A computer (e.g., server, workstation, or processor) performs acts 110-120, such as a computer of the SPECT imager.

[0032] The method is performed in the order shown (numerical or top-to- bottom), but other orders may be used. Acts 112-116 are performed simultaneously, as part of act 110, or in any sequence.

[0033] Additional, different, or fewer acts may be provided. For example, an act for acquiring CT data for a mu-map or attenuation is provided. As another example, acts for modeling attenuation as well as or instead of scatter in the zonal reconstruction are provided. In another example, motion correction is performed. As another example, the generation of the image inact 120 is not performed, instead saving the image in memory and / or transmitting over a computer network.

[0034] In act 100, SPECT data is obtained. SPECT scanning is performed on a patient. In alternative, or additional, embodiments, other functional imaging is performed, such as PET and / or Compton scattering. The SPECT data is measurements of single photon emissions from a patient.

[0035] The SPECT data is obtained from scanning, from data transfer, or from memory. A SPECT system provides the SPECT data directly by scanning or indirectly by transfer or loading.

[0036] The activity concentration in a patient having received a radiotracer or radiotracers may be determined as part of reconstruction by a SPECT system. After ingesting or injecting the radiotracer or tracers into the patient, the patient is positioned relative to a SPECT detector, and / or the SPECT detector is positioned relative to the patient. Emissions from the radiotracer or tracers within the patient are detected over time. The lateral position of a line or cone relative to the detector may be determined. The SPECT detector may be rotated or moved relative to the patient, allowing detection of emissions from different angles and / or locations in the patient.

[0037] The emissions are from a radionuclide with multiple energies. The emissions are at different energies. Energies at two, three, or more levels or windows are detected. The energies are for chosen ranges whether from a continuous energy spectrum, from different major peaks, and / or from different minor peaks. In one embodiment, the emissions are generated by two or more radiotracers, such as where impurity exists or by design. Each radiotracer causes emissions at a different energy, such as using Tc-99m M I Bl and 1-123 MIBG for cardiac imaging. Any combination of two or more radiotracers may be used for a given scan of a patient (i.e. , at a same time). In another embodiment, a radionuclide with different emission energies is used. For example, 1-123, Lu-177 or ln-111 is used. Lu-177 emits with energy peaks at 113kv and 208kv. Other peaks may not be included or may be included within the energy ranges set around the peaks being used. In yet another embodiment, Y-90 is used. The energy spectra of Y-90 are generally continuous rather than having specific peaks. The broad spectra may beapproximated into quasi emission lines. Any two or more portions of the spectra may be used for multi-energy reconstruction. In other approaches, I- 131 , Y90, Sm153, Re186, W188 / Re188, Ho166, Lu177, Cu67, 1125, TI-201 , Tb161 , or progeny of Ac225, Ar211 , Ra223, Tb149 . . . are used as radionuclides with multiple energy emissions. Other isotopes or combinations of isotopes (radionuclides or combinations) may be used.

[0038] Raw SPECT data or preprocessed data is provided for reconstruction. The reconstruction may use a system matrix or projection operators to describe the properties of the SPECT imaging system to iteratively improve a data model of an image object representing the SPECT data. The image object may then be displayed using volume rendering or other imaging techniques.

[0039] The image object, which is defined in an object space, is a reconstruction of the SPECT data measured in a data space. The object space is the space in which the result of the image reconstruction is defined and which corresponds, for example, to the 3D volume (i.e. , field-of-view or “FOV”) that is scanned.

[0040] The SPECT detector or gamma camera may be capable of detecting the different energies. For example, a detector has an operational range of energies that includes multiple energy peaks of a radionuclide. For example, a Cadmium zinc Telluride (CZT) detector has a sufficient thickness to detect over a range of energies. For example, the CZT or other type of detector detects over 40-1000, 40-3000, or 30-3000 keV ranges. Other ranges may be possible.

[0041] In other approaches, the detector arrangement provides for detection at different energies. There are broad classes of emission tomography, separated by primary emission energy. Figure 2 shows four examples. Below or at about 400 keV, 511 keV, or another level keV, the detector 200 operates with physical collimation from a collimator 202, such as a parallel hole collimator. “About” is used for + / -10%. Other values are possible based on material and state of the art. Rather than rely on any form of coincidence, individual emissions progressing through the collimator 202 at a given direction are detected by the detector 200. The collimator 202 in frontof the detector 200 limits the direction of photons detected by the SPECT detector 200, so each detected emission is associated with an energy and line or cone of possible locations from which the emission occurred. For about 400 keV or 500 keV to about 600 keV or 700 keV, an object 204 with a known edge or pattern relative to the detector 200 forms an encoded aperture by moving the object 204 relative to the detector 200. For other approaches at about 500-800 keV, electronic collimation is used. Coincidence processing, such as in positron emission tomography, is provided. For example, a ring 206 of detectors detects a pair of emissions in coincidence with each other. For about 800 keV, 900 keV, or higher, Compton scattering detection may be used. In one example, coincidence is provided by a catcher detector 210 detecting scatter from an event detected by the scatter detector 208.Compton scattering detection may alternatively use electron tracking in a solid detector for both scatter and catcher detection.

[0042] In other alternatives, a hybrid detector arrangement is used. For example, a multi-camera SPECT system is used. One camera is configured to detect at one energy (e.g., lower energy (LE) window), and the other camera is configured to detect at a different energy (e.g., higher energy window). For example, one head or camera is optimized for, e.g., LE SPECT, and the other head or camera is optimized for mid-energy (ME) SPECT. The gammas from the higher peak become a scatter correction only problem, while the image of the LE SPECT gets converted to become spatial or structure information to improve the ME SPECT.

[0043] As another example, a physical collimator 202 is used for one or more lower energy windows, and Compton scattering detection is used for one or more higher energy windows. Separate detectors 200, 208, 210 may be used. Alternatively, the scatter and / or catcher detectors 208, 210 may be used with a collimator to detect the lower energy events. An intrinsic single layer Compton imaging or multi-layer detector may also detect lower energy events.

[0044] In yet other approaches, a three-dimensional tile-able gamma ray detector, such as disclosed in US 2022 / 0354443A, is used to detect in multiple energy ranges. The sensor layout for a direct converter detector,such as disclosed in US 2022 / 0342091 A1 , may be used. A multi-modal Compton and single photon emission computed tomography medical imaging system, such as disclosed in US 2022 / 0330909A1 , may be used for detecting emissions in different energy windows.

[0045] In act 110, an image processor zonally reconstructs an image object from the detected emissions. Any reconstruction in SPECT using zones for different anatomy or spatial structure may be used. Examples include the extra-modal zonal reconstructions disclosed in US Patent Nos. 8,577,103; 8,675,936; or 9,171 ,353. Instead of using CT or other “extra” (not SPECT) data for segmentation to identify zones, SPECT data (detected emissions) at one or more energies are used for segmentation. For example, lower energy emissions may have a greater spatial resolution, so reconstruction of the detected SPECT emissions at the lower energy window is used instead of CT to identify the zones.

[0046] In one implementation, the zonal reconstruction uses emissions at one energy (i.e., energy window or range) to reconstruct an image object from emissions at a different energy. The detected emissions at two energies are used in the zonal reconstruction, which is multi-energy or multi-spectral reconstruction in that sense, but without using the emissions at the energy for identifying the zones in the projections or optimization. For example, the detected emissions for a 511 keV emission window are used to generate a zone map for tumors to improve Lu 177 imaging with zonal reconstruction. The energies used for the zone map are not used in the reconstruction of the image object other than the use of zones. Separate reconstruction is used where the initial reconstruction is provided for segmenting.

[0047] In act 112, in other approaches, the SPECT data at different energies are used together so that the resulting image object is matched to the SPECT data at the different energies. The emissions at the energy used to identify the zone are also used in the reconstruction of the image object as part of the projection and optimization of the zonal reconstruction. Examples of multi-spectral or multi-energy reconstruction are in US Patent Nos. 10,126,439 (reconstruction with multiple photopeaks in quantitative SPECT); or 10,395,353 (multi-modality multi-energy SPECT reconstruction). Zonalreconstruction is used in combination with the multi-energy or spectral reconstruction but with detected photon emissions being used to define the zones instead of or in addition to another modality. The image processor reconstructs the image object with multi-spectral zonal reconstruction. The multi-spectral zonal reconstruction includes the multiple energies and zones from one or more of the multiple energies.

[0048] In act 114, for zonal reconstruction, the image processor determines the zones from the SPECT data at an energy. A zone map of the zones is determined from the emissions at the energy, such as a lower energy window for the radionuclide. The SPECT data at the energy for determining the zones spatially stabilizes the zonal reconstruction from the SPECT data at one or more energies, such as higher energy windows. For example, lower energy photons of an isotope are used to make an image to “stabilize” the higher energy of the same isotope. The low energy peak is imaged, and the resulting image is segmented to identify zones. The higher energy photon emissions are treated as a scattering correction problem for reconstruction. This reconstruction of the higher energy gammas uses the zones or segments from the lower peak energy at higher resolution, improving the spatial resolution of the higher energy collimation (physical or electric) of the higher energy emissions.

[0049] The zones represent locations within the examined object and are derived from SPECT data at one or more energies. The spatial resolution for some energies may be greater than others, so that better spatial resolution is used for zonal reconstruction of the other energies. The zonal reconstruction may improve the image quality and / or reduce the acquisition time of the SPECT imaging process by considering the zonal information in the reconstruction.

[0050] The SPECT emissions at the energy or energies are reconstructed into an image object. This image object is then segmented, such as with thresholding, pattern matching, random walker, model fitting, and / or artificial intelligence. The segmentation separates regions or zones, which can then be used in zonal reconstruction. Each zone is a three-dimensional (3D) region of similar anatomy. For example, bone tissue is segmented from non-bone tissue. The reconstructed image object may be a full or final image object from reconstruction or may be an image object generated in an iteration during reconstruction.

[0051] In the multi-energy zonal reconstruction, a SPECT image of an examined object is reconstructed by considering the spatial or spatial- temporal structure of the object when approximating the SPECT image according to the acquired SPECT data. The structure of the object allows separating the object into multiple zones. Each organ or type of tissue is assigned to a separate zone. The volume within each of those zones is treated separately and equally in the reconstruction. Rather than equal treatment, the independence of the zones may be used for different treatment for different zones. Different amounts of signal are allocated to the zones according to the zone's contribution to the functional feature observed.

[0052] Zonal reconstruction may impose a separation in anatomical zones of the reconstructed image object, but the zones do not modify the merit function of the applied reconstruction algorithm. Figure 3 shows an example. The different zones 330 of the zone map 300 are separately forward projected 320 from the estimate 310 and renormalized as part of the iterative reconstruction. The result of multi-modal reconstructions may be increased resolution as compared to reconstruction with the functional information without zones 330, even with attenuation correction.

[0053] The reconstruction includes use of multiple energy windows. For example, one of the multi-energy models in reconstruction disclosed in U.S. Published Patent Application No. 2017 / 0086757 is used. Since the various image degrading effects (e.g., scatter, attenuation, and / or collimator-detector response function) are different for different energy ranges, the forward projectors 320 of the image formation process for photons at different energy ranges are modeled separately. In one embodiment, scatter, attenuation, and collimator-response functions are modeled separately for each of the different emission energies, emission energy ranges, and / or acquisition energy windows. In act 116, the multi-energy or multi-spectral zonal reconstruction uses a model of scatter to correct for scatter. Similarly, attenuation and / or collimator-detector response function are modeled. One model handles thescatter, attenuation, and / or collimator-response function differently for different energies, providing separate models for separate energy windows.

[0054] Any type of scatter model may be used. Model-based scatter estimation is provided by modeling the physics of scatter in the patient. A Monte-Carlo simulation or other simulation may be used. Other physics or types of modeling of scatter may be used. The scatter may be modeled differently for different energies. Photons with different energies may scatter differently.

[0055] Any type of attenuation model may be used. For example, attenuation coefficients as a function of three-dimensional location in the patient are estimated from anatomical information provided by computed tomography (CT). The attenuation as emitted photons travel through tissue of the patient is modeled using the measured attenuation coefficients. Different energies attenuate differently, which may be modeled as different attenuation coefficients for different energies or a different scaling factor for the different energies.

[0056] Any type of collimator-detector response function model may be used. In one embodiment, point response functions are measured for the specific collimator and detector or for a class (i.e., type of collimator-detector pair). A Monte-Carlo or other simulation may be used. The point response function varies as a function of energy level. Other collimator-detector response functions may be used.

[0057] The image formation models for different energies are used for separate reconstructions at the different energies. The resulting image objects are then combined 332. Rather than totally separate reconstruction at each energy and post-reconstruction combination, the combination for the multiple photopeaks may be performed within or as part of reconstruction. Reconstruction is performed iteratively, so the combination for the multiple photopeaks is performed within the iteration loop of the reconstruction, such as combining back projected feedback 350 of the different photopeaks for updating 360 the volume based on comparison 340. An update is obtained from a comparison 340 of a result of the weighted summing 332 to SPECT data. By combining feedback 350 from the different photopeaks within thereconstruction, one image object is reconstructed for quantitative SPECT. Reconstruction using photon counts from multiple photopeaks in a combined way may increase the signal-to-noise ratio and improve image quality and quantitative accuracy for SPECT imaging.

[0058] For each photopeak, the image volume is projected 320 and back projected 350 with photopeak specific system matrix or projection operators (e.g., projection operators modeling attenuation correction, scatter correction, point response function, and / or sensitivity). The residuals, negradients (for conjugate-gradient method), or an analog resulting from back projection 350 of the multiple photopeaks are combined. The image volume is updated by adding the conjugate gradient resulting from the combined negradients. For the update 360, the conjugate gradient is multiplied by an optimal step size based on the combined negradients. This reconstruction scheme combines multiple photopeaks in one image volume for quantitative SPECT.Alternatively, the combination is of image objects from different energies postreconstruction.

[0059] The reconstruction includes forward projections 320 for zones 330. For each zone 330, forward projections 320 are performed for different energies using different image formation models. Part of the image formation models includes scatter correction in act 116. Any scatter correction may be used, such as energy window-based scatter correction. In one embodiment, model-based scatter correction is used. The scatter correction is performed as part of the forward projection 320 from the image or object space to the data space. The scatter correction model is used in applying the image formation process to the activity distribution. The resulting projection data model has reduced scatter.

[0060] For model-based scatter correction, a scatter response function (SRF) is combined with the activity distribution of the patient to form a modelbased scatter source. The SRF is represented by scatter kernels. The scatter kernels for the given SPECT system are used. The interaction of scatter resulting from different sources with a detector and collimator are simulated. Monte Carlo or other stochastic simulation may be used. The simulation is performed for all systems of a given type, such as all SPECTsystems using a same combination of collimator and detector. The simulation is for that combination, such as based at least in part of the size, shape, and / or material characteristics of the collimator and detector. The simulation is not performed by the SPECT system, but by a computer, workstation, or server. Alternatively, the SPECT system performs the simulation. The results of the simulation are scatter kernels for the collimator and detector combination. The scatter kernels model the common physics in the image formation process for scatter.

[0061] The simulation is for a given radiotracer. The simulation provides for the source or sources to emit at the energy level for the selected photon energy. The simulation provides scatter kernels for different energy levels for the multi-energy image formation model.

[0062] The scatter kernels may be adapted to a specific SPECT system using measured sensitivity for that specific SPECT system. A measured sensitivity is used to normalize across the kernels, adapting the scatter kernels and resulting SRF to the specific collimator-detector combination. The sensitivity is measured at the energy level for the primary photons. Different sensitivities are provided for different energy levels or energy windows. The normalization is by the sensitivities for the primary photons or non-scattered energies from the radiotracer emissions at the different energies.

[0063] The scatter kernels are convolved with the activity distribution to create the model-based scatter source. The model-based scatter source is used to model detection of scatter by the SPECT system. The model-based scatter source generated from the scatter kernels is forward projected. This detection is modeled as the forward radiation transfer to create the model of scatter in the data space.

[0064] The forward projection of the activity distribution from the multienergy image formation is combined with the model of scatter to reduce the scatter in the resulting projection data model for each zone. The image object or activity distribution for a given zone (i e. , zonal object) is convolved with the scatter kernels and also forward projected using the multi-energy formation model.

[0065] The image object for forward projecting is at a given resolution. The resolution may be a resolution of the SPECT data of one energy to increase SPECT resolution for another energy. In alternative embodiments, other resolutions are used. For modeling scatter, a different resolution is used. The image object for the zone is resampled, such as down-sampled or up-sampled. The scatter kernels are based on the energy resolution of the SPECT system, fitting the physics of the different energy windows. The resampling matches the image object resolution to the energy and spatial resolution of the SPECT system as represented by the scatter kernels. Different resampling of reconstructed images is provided for different models in multi-energy modeling.

[0066] The resampling matches the image object resolution to the SPECT system resolution rather than the resolution used for forward projection of the zonal image object. The zonal objects are used at one resolution for forward projecting the zonal objects and at one or more other resolutions for modeling scatter. Resampling is provided for each of the energies being used in the multi-energy reconstruction. In the image formation modeling process, to best incorporate spatial information, the voxel size of reconstructed images is set to be the same as the voxel size of input zone map. In the model of image formation process for primary photons, the voxel size of reconstructed images is not changed. In the model of scattered photons, however, to facilitate necessary convolution with scatter kernels, the reconstructed images (e.g., zonal image objects) are resampled so that the voxel size of resampled images is the same as the voxel size of scatter kernels.

[0067] Figure 4 shows an example implementation of multi-energy zonal reconstruction. This example includes or does not include the scatter correction of act 116. A loop of three operations 400, 410, 420 for three energy windows is provided. The loop may include two, three, four, or more operations. The loop represents iterative reconstruction with reconstruction for the photon emissions at the different (e.g., three) energies E1 , E2, E3 in sequence within each iteration. In the iterative reconstruction, each iteration includes low energy to high energy successive projection (forward and backward) operations 400, 410, 420. The image object of an operation (e.g.,400) from one energy may be used to form zones for zonal-based forward projection in other operations (e.g., 410, 420) of other energies in the loop. Each iteration of the loop has an objective function including the photon emissions at the different energies. For example, the objective function is given by:O»)2= Xi=i« (fMj)’ wherebyrepresents the Mighell modified Chi-squared metric (L2 norm for analytical simplicity and assuming Poisson noise, yet any L1 or L2 appropriate for respective problem might be used) for the image formation model IMt, containing the appropriate image formation model as a system matrix Hi. The minimization of theyields the tomographic reconstruction solving the respective inverse problem (i.e., inversion of generalized attenuated Radon transform). The data model d1^ = HIF1IFfor the specific Image formation (IF) with its system matrix HIF, allows a joint estimation with weights to allow for tailoring based on the information content obtained. That can be done by the designer or by a artificial intelligence (Al) approach. E.g., at the high energy, the spatial resolution may be very poor, but high statistics, while at the low energy with better spatial resolution the statistics is poor. In which case, one would use the lower energy to bound the region and high energy to estimate the uptake within the boundary. Cross terms can further be introduced to tailor the minimization efficiency, yet not to bias the result.

[0068] For each iteration through the loop, the same operations 400, 410, 420 are repeated. In other approaches, the operations may change for different iterations. For example, after a given number of iterations or other change criterion is meet, zones are not updated. The same zones continue to be used for later iterations.

[0069] The iterations continue until a stop criterion or criteria are meet. Once complete, the image object from the last iteration is used for imaging.

[0070] In act 120 of Figure 1 , the image processor generates an image from the image object. The output of the reconstruction is used for imaging. The reconstruction outputs an image object or volume representing thepatient from a last iteration. This final image object is used for generating the image.

[0071] The image object is a three-dimensional representation of the detected emissions of the patient. The image object is rendered or otherwise used to generate an image. For example, a multi-planar reconstruction or single slice image of a plane is generated. The intersection of one or more planes with the image object is visualized. As another example, a surface or projection rendering is performed for three-dimensional imaging. Other imaging may be used.

[0072] One image is generated. Alternatively, a sequence of images is generated. For example, image objects from different time periods are used to generate a sequence of images representing the patient over time.

[0073] The image of the functional information from the zonal reconstruction is displayed alone. Alternatively, an anatomical image is displayed with the functional image. For example, the functional image is overlaid on a CT image. The overlay may be colored for display on a gray scale CT image. Other combinations may be used.

[0074] For quantitative SPECT, the image may be an alphanumeric text of a specific uptake value for a location. A graph, chart, or other representation of uptake at multiple locations may be output. The spatial image representing distribution of uptake may use color or brightness modulation to represent a level of uptake by location.

[0075] The image object used for imaging is reconstructed from emissions at one or more energies. For example, the image object is zonally reconstructed from emissions at one energy window based on zones determined from reconstruction of emissions at a different energy window. As another example, multiple image objects are separately zonally reconstructed from the different energy windows. The image object used for imaging is a combination (e.g., average or weighted average) of the image objects from the different energies (i.e. , post reconstruction combination). In yet another example, the image object is zonally reconstructed from the detected emissions at the different energy windows. The combination is within thereconstruction. The emissions at the different energies are combined in the reconstruction, such as with a shared objective function.

[0076] Figure 5 shows one embodiment of a medical imaging system 500 for SPECT intra-modal zonal reconstruction. The system 500 may implement multi-energy zonal reconstruction where the zones are intra-modal (i.e. , from the SPECT emissions). The method of Figures 1 , 3, and / or 4 or another method is implemented.

[0077] The system 500 is a SPECT imaging system or scanner and includes a detector arrangement 510, reconstruction processor 520, a memory 530, and a display 540. Additional, different, or fewer components may be provided. For example, a PET or Compton imaging system is provided instead of the SPECT imaging system. In one embodiment, the reconstruction processor 520, memory 530, and / or display 540 are part of the SPECT imaging system. In alternative embodiments, the reconstruction processor 520, memory 530, and / or display 540 are provided as a workstation, server, or computer separate from the detector arrangement 510. The memory 530 is part of a computer or workstation with the reconstruction processor 520 or is a remote database, such as a picture archiving and communications system (PACS).

[0078] The detector arrangement 510 includes one or more detectors for detecting emitted radiation from within the patient. For SPECT, a gamma camera is used to detect. The detector detects photon emissions. The photon is emitted from a tracer or radiopharmaceutical. The detector detects the photon. A given detector may detect a sequence of events from the same or different locations of the patient.

[0079] The tracer includes a radionuclide with a complex energy spectrum. Multiple energy peaks or a region of substantially continuous energy are provided. A combination of radionuclides may be provided to generate the emissions at different energies. The radionuclide emits energies at or near the different energy peaks or within a continuous energy region.

[0080] The detector arrangement 510 includes one or more detectors for detecting in two or more different energy windows. The detector arrangements of Figure 2 or other detector arrangements may be used. Forexample, one detector detects emissions at different energy windows, such as using a CZT detector. As another example, a combination of detectors for physical collimation and Compton scattering is provided. In yet another example, a multi-camera system has one detector with physical collimation for a relatively lower energy range (e.g., LE detector) and another detector with electrical or physical collimation for detecting at a relatively higher energy range (e.g., ME detector).

[0081] The reconstruction processor 520 is a general processor, central processing unit, control processor, graphics processor, digital signal processor, application specific integrated circuit, field programmable gate array, artificial intelligence processor, digital circuit, analog circuit, timing circuit, combinations thereof, or other now known or later developed device for reconstructing a patient volume from detected emissions. The reconstruction processor 520 is a single device or multiple devices operating in serial, parallel, or separately. The reconstruction processor 520 is specifically designed or provided for reconstruction but may be a main or general processor of a computer, such as a laptop or desktop computer, or may be a processor for handling tasks in a larger system. The reconstruction processor 520 may perform other functions than zonal reconstruction.

[0082] The reconstruction processor 520 is configurable. The reconstruction processor 520 is configured by software, firmware and / or hardware. Different software, firmware, and / or instructions are loaded or stored in memory 530 for configuring the reconstruction processor 520.

[0083] The reconstruction processor 520 is configured to reconstruct an object representing the patient from the detected emissions. The reconstruction may be performed for at least two energy windows of the distributed energies of the radionuclide. The reconstruction also includes zonal reconstruction where the zones are intra-modal (i.e. , from detected emissions). Detected emissions at one or more energies are reconstructed or projected to image space for spatial segmenting. For example, the emissions at a lower energy window are used to form an image object or patient representation for segmenting the zones. These zones are then used for zonal reconstruction. The emissions from the higher energy window orwindows are zonally reconstructed using the zones from the lower energy emissions. In a further embodiment, the reconstruction is multi-spectral or multi-energy, so an image object is zonally reconstructed from the emissions of both or multiple energy windows, including or not the energy window used to identify the zones.

[0084] In one embodiment using a scatter model, the reconstruction processor 520 is configured to forward project a zonal image at a first resolution with multi-energy projectors and to model scatter with the zonal image at a second resolution different than the first resolution. The zonal images are resampled (i.e., sampled differently) for the primary multi-energy projection and the model of scatter. For example, the resolution of the zonal images is a resolution of the image object from reconstruction of the lower energy emissions, and the resolution for modeling scatter is a system or scatter kernel resolution.

[0085] The reconstruction processor 520 may be configured to alter the reconstruction. The reconstruction is iterative. A different reconstruction process may be used for later iterations than for earlier iterations.

[0086] The memory 530 is a random-access memory, graphics processing memory, video random access memory, system memory, cache memory, hard drive, optical media, magnetic media, flash drive, buffer, database, combinations thereof, or other now known or later developed memory device for storing data. The memory 530 stores detected emissions (e.g., PET, Compton, or SPECT detected event data), zone information, segmentation information, energy information, and / or reconstruction information. The memory 530 stores data as processed, such as storing an updated image object, zonal image objects, renormalization coefficients, scatter kernels, projection operators or system matrix, zonal data models, combined data models, zone functions, resampled image objects, and / or other information.

[0087] The memory 530 or other memory is a non-transitory computer readable storage medium storing data representing instructions executable by the programmed reconstruction processor 520 for SPECT reconstruction. The instructions for implementing the processes, methods and / or techniques discussed herein are provided on computer-readable storage media ormemories, such as a cache, buffer, RAM, removable media, hard drive, or other computer readable storage media. Computer readable storage media include various types of volatile and nonvolatile storage media. The functions, acts or tasks illustrated in the figures or described herein are executed in response to one or more sets of instructions stored in or on computer readable storage media. The functions, acts or tasks are independent of the particular type of instructions set, storage media, processor or processing strategy and may be performed by software, hardware, integrated circuits, firmware, micro code and the like, operating alone, or in combination. Likewise, processing strategies may include multiprocessing, multitasking, parallel processing, and the like.

[0088] In one embodiment, the instructions are stored on a removable media device for reading by local or remote systems. In other embodiments, the instructions are stored in a remote location for transfer through a computer network or over telephone lines. In yet other embodiments, the instructions are stored within a given computer, CPU, GPU, or system.

[0089] The display 540 is a monitor, LCD, plasma, touch screen, printer, or another device for displaying an image for viewing by a user. The display 540 shows one or more images representing function, such as uptake or activity concentration. The image is a quantitative or qualitative SPECT image of the reconstructed object. The image may be a volume rendering, a multi-planar reconstruction, a cross-section, and / or another image from a final image object. The image represents a distribution of the radionuclide in the patient based on detected emissions from the SPECT system 500.

[0090] While the invention has been described above by reference to various embodiments, it should be understood that many changes and modifications can be made without departing from the scope of the invention. It is therefore intended that the foregoing detailed description be regarded as illustrative rather than limiting, and that it be understood that it is the following claims, including all equivalents, that are intended to define the spirit and scope of this invention.

Claims

I (WE) CLAIM:1 . A method for single photon emission computed tomography (SPECT) reconstruction, the method comprising: detecting SPECT emissions from a patient, the SPECT emissions at first and second energies; zonally reconstructing an image object from the detected SPECT emissions, zones of the zonal reconstruction for the second energy being determined from the SPECT emissions at the first energy; and generating an image from the image object.

2. The method of claim 1 wherein detecting comprises detecting the SPECT emissions from a radionuclide where the radionuclide comprises Lu- 177, 1-131 , TI-201 , Tb161 , or progeny of Ac225.

3. The method of claim 1 wherein detecting comprises detecting with a multi-camera SPECT system where a first camera of the multi-camera SPECT system is configured to detect at a lower energy than a second camera of the multi-camera SPECT system.

4. The method of claim 1 wherein detecting comprises detecting the SPECT emissions at the first energy with physical collimation and detecting the SPECT emissions at the second energy with Compton scattering detection.

5. The method of claim 1 wherein zonally reconstructing comprises generating a zone map including the zones from a first reconstruction of the SPECT emissions at the first energy, forward projecting the SPECT emissions at the second energy separately for each of the zones, weighted summing of the forward projections, obtaining an update from a comparison of a result of the weighted summing to the SPECT emissions, back projecting the update, and updating an estimate from the back projected update.

6. The method of claim 1 wherein zonally reconstructing comprises generating a zone map of the zones from the SPECT emissions at the first energy, the first energy being lower than the second energy.

7. The method of claim 6 wherein the SPECT emissions at the first energy spatially stabilize the zonal reconstruction from the SPECT emissions at the second energy, the second energy being a higher energy than the first energy.

8. The method of claim 1 wherein zonally reconstructing comprises modeling scatter for the second energy, the second energy higher than the first energy, with the SPECT emissions at the second energy being segmented by the zones determined from the first energy, the zones determined with a greater spatial resolution than an image object of the SPECT emissions at the second energy.

9. The method of claim 1 further comprising initial reconstruction of the SPECT emissions at the first energy, and then segmenting the zones from the initial reconstruction for the zonal reconstruction of the SPECT emissions at the second energy.

10. The method of claim 1 wherein zonally reconstructing comprises iterative reconstruction with reconstruction for the SPECT emissions at both the first and second energies in sequence within each iteration.11 . The method of claim 10 wherein the iterative reconstruction includes a loop sequencing from the first energy to the second energy in each iteration, the first energy being lower than the second energy, each iteration having an objective function including the SPECT emissions at both the first and second energies.

12. The method of claim 1 wherein generating the image comprises generating the image from the image object of the SPECT emissions at the second energy.

13. The method of claim 1 wherein generating the image comprises generating the image from the image object zonally reconstructed from the detected SPECT emissions at the second energy combined with another image object reconstructed from the detected SPECT emissions at the first energy.

14. The method of claim 1 wherein generating the image comprises generating the image from the image object zonally reconstructed from the detected SPECT emissions at the second energy in combination with reconstructed from detected SPECT emissions at the first energy.

15. A method for single photon emission computed tomography (SPECT) reconstruction, the method comprising: detecting SPECT emissions from a patient, the SPECT emissions being from a radionuclide with multiple energies; reconstructing an image object with multi-spectral zonal reconstruction, the multi-spectral zonal reconstruction including the multiple energies and zones from a first of the multiple energies; and generating an image from the image object.

16. The method of claim 15 wherein reconstructing the image object comprises iterative reconstruction where each iteration includes low energy to high energy successive projection operations.

17. The method of claim 15 wherein reconstructing comprises using the zones from the first energy, the first energy being a lower energy than a second energy where the multi-spectral zonal reconstruction reconstructs for the second energy with the zones from the first energy.

18. A medical imaging system for single photon emission computed tomography (SPECT) intra-modal zonal reconstruction, the medical imaging system comprising:a detector arrangement configured to detect emissions from the patient, the emissions of different energies from a radiotracer; a processor configured to reconstruct an object representing the patient from the detected emissions for at least two energy windows of the distributed energies, wherein the reconstruction for a lower of the at least two energy windows is segmented, and wherein the reconstruction for a higher of the at least two energy windows is zonal based on the segmentation from the reconstruction for the lower of the at least two energy windows; and a display configured to display an image of the reconstructed object.

19. The medical imaging system of claim 18 wherein the detector arrangement comprises a detector configured to detect with the at least two energy windows.

20. The medical imaging system of claim 18 wherein the detector arrangement comprises a physically collimated detector and a Compton scattering detector.

Images