Tomography using generalized time-encoded aperture imaging

Abstract

Regarding tomography, more radiation is detected. To detect a large number of radiations and provide better resolution than that provided by a parallel hole collimator, the collimator is replaced by an attenuator having outer and inner edges. Instead of forcing directivity, large holes having different shapes are used to enable detection of more radiation. By operating the attenuator, differences in shadows on the sensor are used as time-coded apertures to reconstruct the radiation source with higher resolution and sensitivity than when a fixed parallel hole collimator is used.

Description

Technical Field

[0001] One aspect of the present invention relates to emission tomography, such as single photon emission computed tomography (SPECT), positron emission tomography (PET), or another type of imaging using a gamma camera.

Background Art

[0002] High-efficiency tomography of gamma rays emitted from radioisotopes is typically performed from a discrete spectrum of greater than about 50 keV to less than 511 keV using a physical collimator. This collimator only allows gamma rays to enter the sensor in a specific direction (i.e., a parallel hole collimator) to generate a projection image (i.e., an image detected by a gamma camera). This process involves a trade-off between the high directional accuracy of the collimator and the sensitivity. The small directional holes of the collimator block a lot of radiation. This results in noisy data due to this low sensitivity, which makes tomography difficult as an inverse problem and causes artifacts. Regarding tomography reconstruction, high directional accuracy is achieved with the collimator at a high cost in terms of sensitivity.

[0003] Medical emission tomography is used for various specific categories of tasks: 1. Detecting lesions ("detection task"), according to which a lesion is an abnormal accumulation that is too high or too low compared to what is predicted based on anatomy, physiology, and the administered radiopharmaceutical, 2. Characterizing lesions or accumulation patterns by descriptive statistics such as average accumulation density. In some cases, the ability to spatially resolve is important (spatial resolution), in other cases, resolving the signal in a noisy background, i.e., contrast resolution, is important, and in still other cases, resolving the signal as a temporal accumulation change ("temporal resolution") is important. Usually, a statistical-based criterion for separating the signal from the noise with a certain level of confidence is used. The fundamental trade-off between direction ("resolution") and the information of presence ("sensitivity") remains.

[0004] This fundamental compromise can be affected by imposing or utilizing certain auxiliary physical conditions for image formation. Contrast and attenuation patterns can be used as sources of information derived from utilizing certain physical conditions for image formation. Edge encoding is presented. Time-dependent variable patterns such as super-resolution time multiplexing methods are being investigated. Compressed sensing is used. If the use of additional information compensates for the loss of directional information and there is more than enough, a further increase in sensitivity can lead to better tomographic imaging for specific clinical tasks. SUMMARY OF THE INVENTION

[0005] As a prelude, the preferred embodiments described below include a method, a system, and a sensor for tomography. To detect more radiation and / or provide better resolution than provided by a parallel hole collimator, the collimator is replaced by an attenuator having outer and inner edges. Instead of forcing directivity (directionality), larger holes having different shapes are used to enable detection of more radiation. By operating the attenuator, differences in shadows on the sensor are used as time-coded apertures to reconstruct the radiation source with higher resolution and sensitivity than when using a stationary parallel hole collimator. The directional information required for a specific task is obtained based on the moving shadows.

[0006] In a first aspect, a tomographic imaging system is provided. The sensor is configured to detect the time, position, and energy of gamma rays. The movable attenuator has one or more inner through-holes (through holes). The drive device is configured to operate the movable attenuator. The image processor is configured to reconstruct the spatial distribution of the radiation detected by the sensor using the movable attenuator that assumes different arrangements (positions / postures) due to the movement by the drive device. The movable attenuator is located between the radiation source and the sensor such that the moving shadow of the through-hole is projected onto the sensor.

[0007] In one form, the sensor is a planar gamma camera. As a further example, the planar gamma camera is connected to a gantry configured to position the planar gamma camera at various positions relative to a radiation source during detection of radiation.

[0008] The movable attenuator has various forms. According to one form, the movable attenuator is an object of lead or tungsten. As a further form, the movable attenuator is operable by translating and / or rotating in three dimensions. In another form, the through-hole is a slit. In other forms, the through-holes have different sizes, shapes, and / or hole angles.

[0009] As a further form, the movable attenuator is a rotatable cylinder, in which case the radiation source or the sensor can be disposed within the rotatable cylinder. The drive device is configured to rotate the rotatable cylinder. The movable attenuator can have various shapes such as a plate.

[0010] In one form, the drive device is configured to rock the movable attenuator and / or rock the movable attenuator about a perpendicular (normal) to the movable attenuator. Other movements may be used, such as depending on the imaging application.

[0011] In one form, the image processor is configured to form a projection from the radiation, and the reconstruction of the spatial distribution is from that projection. As an example, the projection is a virtual parallel hole collimator projection, and the reconstruction is a successive approximation reconstruction (iterative reconstruction) using the virtual parallel hole collimator projection in forward projection and back projection.

[0012] In a second aspect, a method of SPECT is provided. An attenuator having an inner edge is operated between the patient and the sensor. This moving inner edge forms a time-coded aperture on the sensor. Radiation passing from the patient through the attenuator and having different shadows on the sensor is detected by the sensor. The representation of the patient is reconstructed from the radiation detected using the time-coded aperture.

[0013] In one form, the attenuator rotates and / or translates in three dimensions. The shadow varies according to the position on the sensor and / or the rotation of the attenuator.

[0014] In another form, the inner edge forms holes having different shapes, sizes, and / or angles, resulting in corresponding shadows.

[0015] In another form, the holes in the attenuator form edges. The attenuator is operated in three dimensions such that the shape and / or size of the holes in the shadow vary over time.

[0016] In yet another form, the reconstruction is from the edge response of the shadow. As another form of reconstruction, the reconstruction includes constructing projections at different viewing angles for the patient from the detected radiation and based on the time-coded aperture, and reconstructing from the projections.

[0017] In a third aspect, a tomographic imaging system is provided. The radiation blocker has an inner edge that forms a hole passing through the radiation blocker. The sensor is configured to detect radiation passing through the hole when the radiation blocker is in different arrangements relative to the sensor. This different arrangement forms a time-coded aperture for the sensor. The image processor is configured to form virtual projections from different views from the radiation detected by the sensor using the time-coded aperture, and to reconstruct a representation of the patient from the virtual projections.

[0018] In a further form, the holes have different sizes, shapes, and / or angles.

[0019] The present invention is defined by the claims, and nothing in this column should be regarded as limiting the scope of those claims. Further aspects and advantages of the present invention are described below in conjunction with the preferred embodiments and may be claimed later, either independently or in combination.

Brief Description of the Drawings

[0020] The components and the drawings are not necessarily to scale and, rather, are exaggerated in order to illustrate the principles of the present invention. Further, in the drawings, like reference numerals indicate corresponding parts throughout the figures.

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

DETAILED DESCRIPTION OF THE INVENTION

[0021] The non-local pattern of a coded aperture or physical collimator can be used to explore spatial frequencies. In the case of standard SPECT imaging, a regular degenerate pattern (e.g., parallel holes) is used to generate a non-local PSF across the field of view (FOV). The result is a strict quality control regarding the collimator and sensor uniformity, and an aperture smaller than the original resolution is selected to generate a degenerate response that maximizes the signal-to-noise ratio (SNR) without adding computational load. This approach can be abandoned with the use of a very local response and a time-coded varying collimation.

[0022] A tomographic imaging system uses a virtual system resolution restored from generalized time-coded aperture imaging data for imaging discrete or continuous gamma-ray spectra. The concept of time-coded coded apertures is generalized to enable edge-based super-resolution. Edge resolution assuming known attenuators is used to generate a dataset that takes into account super-resolution, distance measurement, and / or tomographic information. The attenuator provides a coded mask.

[0023] Using edge resolution information and the positive semi-definiteness of the problem, tomographic image formation is provided using the operation of a three-dimensional attenuator with sharp edges. The directivity is extracted from the non-local pattern of the point spread function (PSF) across the field of view (FOV) from a coded time-varying aperture pattern based on positive semi-definite data.

[0024] Separate the directivity and holes in the attenuator using the edge response from the time-coded aperture and increase the detected radiation count with long edges to expand the imaging space. The directivity information is extracted more efficiently than with a fixed parallel hole collimator. In the case of near-field tomography, the shape and operation of the attenuation aperture are designed for each application (e.g., a specific operation pattern of the attenuation aperture body based on the purpose of imaging). The edge response using a movable attenuator with an inner edge is, in a sense, an abstract concept of a multi-focus concept paired with time coding.

[0025] FIG. 1 shows an example of a tomographic imaging system. The tomographic imaging system is a medical imaging device such as a SPECT imaging system. This system is an imaging system for imaging a patient 114 on a bed 104. The tomographic imaging system uses edge response with time-coded aperture by operating an aperture with inner through-holes to enhance sensitivity and sharpen resolution.

[0026] The tomographic imaging system includes a gantry 102, a sensor 106, an attenuator 108, a drive device 112, and an image processor 120. Additional, different, or fewer components may be provided. For example, a non-transitory memory for storing detected radiation and / or instructions executed by the processor 120 is provided. As another example, a display for displaying a reconstructed image of the patient 114 is provided. In another example, a separate detection processor for binning detected radiation is provided.

[0027] The gantry 102 is part of a housing. The housing is made of metal, plastic, fiberglass, carbon (e.g., carbon fiber), and / or another material. The housing forms a patient area (e.g., a bore) where the patient is placed during imaging. The bed 104 can move the patient within the patient area to scan different parts of the patient at different times. In other embodiments, a chair or bed without a housing forming a bore is used, such as when the sensor 106 and the attenuator 108 are positioned by one or more robotic arms.

[0028] The gantry 102 is a motor, sensor, and / or track that moves the sensor 106 relative to the patient 114, for example, to capture radiation from different angles and / or positions relative to the patient 114. In an alternative embodiment, the bed 104 moves without a gantry (e.g., the sensor 106 is fixed within the housing). In yet another alternative, both the sensor 106 and the bed 104 are fixed during imaging.

[0029] Sensor 106 is composed of a structure and / or electronics for detecting radiation from radiation within patient 114. Sensor 106 is configured to detect the position, energy, and time of gamma-ray collisions. Sensor 106 is a detector such as a SPECT detector or a gamma camera and does not require a collimator fixed to the sensor. A portion of the radiation passes through hole 202 inside attenuator 108 and / or along the edge around the outside of attenuator 108. Sensor 106 uses attenuator 108 in different known arrangements with respect to sensor 106 to detect the radiation. The different arrangements form a time-coded aperture with respect to sensor 106.

[0030] In one embodiment, sensor 106 is a SPECT sensor having a number of pixelated detector cells. For example, sensor 106 is a gamma camera. The gamma camera includes one or more semiconductor sensors such as a pixelated sensor having detection cells. Sensor 106 consists of room-temperature semiconductor sensors. In another example, it is an array of silicon photomultiplier cells coupled to a scintillator. Sensor 106 forms an array of sensors or pixelated sensor cells. Anode electrodes and cathode electrodes are provided on both sides of sensor 106. These electrodes have the same pitch as the detection cells and are electrically insulated from each other for separate electrical connection to the detection cells of sensor 106.

[0031] Any material is used, such as a scintillator such as NaI, or a direct converter such as CZT, CdTe, TlBr, and / or other similar materials suitable for gamma-ray imaging. Sensor 106 is fabricated from a wafer of any thickness, such as about 5 - 10 mm in the case of CZT. Sensor 106 has a radiation detection surface that is square or has a rectangular shape. Any size is used, such as about 5 cm × 5 cm. Shapes other than rectangular or square, such as triangular or hexagonal, may be used.

[0032] Sensor 106 is adjacent to the patient area, such as being attached to the movable gantry 102. Sensor 106 is designed and configured to detect gamma radiation, such as radiation from patient 114. This gamma camera is a planar camera connected to the gantry 102 or the housing, and the planar gamma camera is positioned at different positions relative to the radiation source to detect radiation. For example, gantry 102 moves sensor 106 laterally or parallel to bed 104 and / or rotates sensor 106 around the longitudinal axis (i.e., around bed 104). In other embodiments, a robotic arm or system is used to move sensor 106. Alternatively, sensor 106 is fixed to housing 102 relative to the patient space.

[0033] Attenuator 108 is made of lead, tungsten, or other materials that block, reflect, or absorb gamma rays from radiation. The gamma rays are redirected or stopped so that the radiation intersecting attenuator 108 does not pass from patient 114 to sensor 106. Attenuator 108 is a radiation blocker.

[0034] Attenuator 108 is a plate or object having a thickness that blocks radiation. Any shape, such as rectangular, hexagonal, or triangular, is provided on the largest surface. This shape forms an outer edge such as a sharp edge. The edge is rounded or shaped according to the amount of rotational and / or translational movement relative to the patient.

[0035] Attenuator 108 also includes one or more inner through-holes 202. Any number of through-holes 202, such as one, two, dozens, or hundreds, are provided. Through-holes 202 form an inner edge. This edge is a sharp edge (e.g., flat / planar), but may have other shapes such as being rounded.

[0036] Through-holes 202 have any shape, size, and / or angle. FIG. 2 shows an example where hole 202 is formed as a slit. Circular, hexagonal, triangular, curved, and / or other shapes of hole 202 may be provided.

[0037] Rather than having holes sized at or below the resolution of sensor 106, the holes are sized to include at least one dimension that exceeds the resolution of sensor 106. For example, slit 202 extends across dozens or hundreds of detector cells along at least one dimension or across two dimensions.

[0038] Hole 202 has an inner edge that is perpendicular to the maximum surface of attenuator 108 (e.g., perpendicular to sensor 106 when the maximum surface is parallel to the detection surface). The edges may be at other angles so as to block or pass radiation directionally. FIG. 1 shows two holes 202 at two different angles (i.e., the inner edges have different angles with respect to the maximum surface of sensor 106 and / or the detection surface). Each portion of the edge formed for a given hole 202 has the same or different angles, such as having a trapezoidal shape in cross-section.

[0039] The holes 202 of the attenuator have the same or different shapes, sizes, edge angles, and / or orientations. For example, all holes 202 have the same size, shape, edge angle, and edge orientation. As another example, each hole 202 has a size, shape, orientation, and / or angle that is different from at least one other hole 202. FIG. 2 shows three holes 202 each having a different size and / or shape. One hole 202 is tilted in the plane of attenuator 108 with respect to the other holes 202, showing a different orientation. FIG. 3 shows an example where the holes 202 are concentric rings or slits having different widths. FIG. 4 shows an example where the holes 202 are non-concentric rings or slits having different widths and different center positions at the maximum surface. FIG. 6 shows an example where the holes 202 are rectangular slits and each hole 202 overlaps or intersects. Different sizes, orientations, and / or thicknesses can be used.

[0040] By providing an inner edge, many opportunities are provided to depict the directivity from the radiation source to the sensor 106. By providing different sizes, shapes, and / or angles, much information regarding the depiction is provided. The larger and / or more numerous the holes 202 are, the more radiation can be collected, increasing the sensitivity.

[0041] In principle, any three-dimensional (3D) collimation can be used. Different shapes, sizes, edge angles, orientations, and / or combinations of the holes 202 provide different processing (i.e., reconstruction) efficiencies. For example, the holes 202 of the attenuator 108 in FIG. 6 are an efficient pattern for mapping to a rectangular form factor having the same aspect ratio as the sensor 106. Any 3D collimator shape can be created by stacking 2D shapes. For example, a rectangular hole collimator can be created by plates having slits (a "slit - slat") orthogonal to each other, and instead of moving the sensor having the collimator to obtain different field of view angles, a movement pattern can be created on the detection plane of the sensor just by moving the "slit" in front of the "slat".

[0042] FIG. 8 shows another embodiment of the attenuator 108. This attenuator 108 is a hollow cylinder having slits as the holes 202. Although shown as parallel slits of the same size and shape, holes of various orientations, sizes, shapes, or edge angles may be used. A hollow shape other than a cylinder, such as a cube shape, may be used. The sensor 106 or the radiation source (e.g., patient 114) is disposed inside the cylindrical attenuator 108. The cylinder 108 rotates around the radiation source or the sensor 106 to provide relative movement with respect to the sensor 106. When the sensor 106 is inside the attenuator 108, an external (ecto) tomography with a small installation area is provided, and it becomes possible to robotically control the operation of the sensor 106 and the attenuator 108 around the patient or fix them to the patient without a large housing and bore (i.e., the patient sits on a chair or lies on a bed).

[0043] Referring again to FIG. 1, the drive device 112 is a motor such as a servo, an electric motor, a pneumatic compressor, an actuator, a hydraulic pump, or other motor that can apply force to the damper 108. A gear mechanism, pulley, guide, clutch, rack and pinion, tube, and / or other mechanism transmits the force from the motor, e.g., the rotational force from an electric motor, to the damper 108. The drive device 112 is constituted by control and mechanical connection for operating the movable damper 108. Since the drive device 112 encodes the opening, it is an encoding drive device. The drive device 112 positions the damper 108 in a known arrangement and / or in a known operation. This encoding provides the positional relationship between the damper 108 and the sensor 106.

[0044] In one embodiment, the drive device 112 is part of a robotic arm. The robotic arm positions and operates the damper 108. In other embodiments, the damper 108 is connected to a guide or gear mechanism and operates repetitively under control. The drive device 112 is configured to operate the damper 108 between the sensor 106 and the patient 114 (i.e., the emitter), together with the inner edge and possibly the outer edge from the hole 202.

[0045] The drive device 112 operates the damper 108 with any degree of freedom, e.g., 1 to 6 degrees of freedom. In one embodiment, the operation is three-dimensional, such as three rotational degrees of freedom, three translational degrees of freedom, or three rotational degrees of freedom and three translational degrees of freedom respectively.

[0046] Some examples are shown in FIG. 2. The shadow 200 of the damper 108 on the sensor 106 is shown. Three holes 202 and the outer edge result in the shadow 200 on the sensor 106 in this example. If the damper 108 is larger than the sensor 106, the shadow will include the holes 202 (i.e., the inner edges) and none or all of the outer edges.

[0047] The shadow 200 shown in the upper left figure is such that the maximum surface of the attenuator 108 is parallel to the sensor surface of the sensor 106 (that is, the perpendiculars of both are parallel to each other and perpendicular to the plane of the paper). The attenuator 108 can be rotated around the perpendicular of the center of the attenuator 108 or shifted (offset) to another position. The upper right figure shows the rotation or change in orientation around the center of the attenuator 108 and the shadow 200, as well as a slight translation to the left. The lower left figure shows the shadow 200 translated to a position that is different in the vertical dimension and slightly different in the horizontal dimension. The lower right figure shows the shadow 200 resulting from tilting the attenuator 108 around the horizontal axis on the plane of the paper such that one edge approaches the sensor 106 more than the opposite edge. The shadow 200 resulting from tilting around this horizontal axis has a slit narrowed by the hole 202 as seen from the perpendicular to the sensor surface. Other operations or combinations of operations may be provided.

[0048] Figure 8 shows the rotational operation of the cylinder formed by the attenuator 108. The drive device 112 rotates the attenuator 108 around the central axis or the offset axis to move the shadow from the hole 202 relative to the sensor 106. The sensor 106 may be made to perform operations such as translation along the axis of the cylinder and / or tilting around an axis perpendicular to the axis of the cylinder.

[0049] Figure 7 shows the operation of tilting the attenuator 108. The attenuator 108 with the hole 202 is tilted due to the cosine alpha effect. The attenuator 108 rotates around a central axis parallel to the maximum surface of the attenuator. Other rotation axes may be used. The attenuator 108 may be rotated to provide rocking. The attenuator 108 may be periodically rotated by further rocking in the opposite direction.

[0050] Figure 5 shows the rocking of the attenuator 108. The attenuator 108 is rocked about the perpendicular of the attenuator 108. This rocking is along another axis that may or may not be parallel to the perpendicular. The attenuator 108, as a plate having the hole 202, is rocked such that the perpendicular vector precesses (slides) with a certain opening angle and frequency ω.

[0051] Other operations and / or combinations of operations may be provided. The operations are performed with any frequency, speed, and / or range. The operations can be varied, such as by changing the speed, frequency, or range. The operations may be continuous or may operate in a step function. For example, the attenuator 108 is held in a predetermined arrangement for a period, such as 5 minutes, and then moved to another arrangement where the attenuator 108 is held for the same or a different period. Any number of steps or discrete holding arrangements, such as 2, dozens, or hundreds, can be used. Any holding time, such as seconds or minutes, can be used. The operations are used to generate a strong locally varying pattern that explores the spatial frequency.

[0052] The image processor 120 of FIG. 1 is a general-purpose processor, an artificial intelligence processor or accelerator, a tensor processor, a digital signal processor, a graphics processing unit, an application-specific integrated circuit, a field-programmable gate array, a digital circuit, an analog circuit, a combination thereof, or other currently known or future-developed devices for processing radiation information and / or reconstructing an image based on the detected radiation (e.g., the position, energy, and / or time of the incident radiation). The image processor 120 is one device, multiple devices, or a network. In the case of two or more devices, parallel processing or sequential partitioning processing may be used. Each of the devices constituting the image processor 120 performs a separate function. For example, one processor controls the operation of the attenuator 108, and another processor forms a projection from the detected radiation (i.e., position, energy, and / or time) and the time-coded aperture and performs reconstruction from the projection. In one embodiment, the image processor 120 is the control processor or other processor of a medical imaging system. In other embodiments, the image processor 120 is a separate workstation, server, or part of a computer.

[0053] The image processor 120 operates according to stored instructions to execute various processes described herein, such as processes 1106, 1108, and / or 1110 of the method of FIG. 11. The image processor 120 is configured by software, firmware, and / or hardware to execute each process.

[0054] The image processor 120 is configured to reconstruct an object from the detected radiation. The position, time, and energy of the radiation are used to reconstruct the object. The spatial distribution of a radiation source, such as a radiopharmaceutical within the patient 114, is reconstructed. The image processor 120 reconstructs the spatial distribution of the radiation source detected by the sensor 106 with the movable attenuator 108 that assumes different arrangements with the operation of the driving device 112.

[0055] Since the attenuator 108 operates between the radiation source and the sensor 106 to produce a local pattern (i.e., shadow 200) at each time, this time-coded aperture is used to determine the directivity in the reconstruction. The moving shadow 200 of the hole 202 projected onto the sensor 106 is used in the reconstruction.

[0056] FIG. 9 shows an example based on the outer edge of the attenuator 108. The same principle applies to the inner edge. By having multiple edges, more information is obtained about a given arrangement of the attenuator 108 with respect to the sensor 106.

[0057] The radiation source 900 emits radiation at different times. The attenuator 108 is in different arrangements with respect to the sensor 106 at those different times. Due to the different arrangements of the attenuator 108, the position of the radiation source 900 is resolved based on the position of the edge of the attenuator 108 and the position of the detection on the sensor 106. Triangulation and time-coded aperture provide the position of the radiation source 900 in three dimensions. The shadow determined by the shape and moment of the attenuator 108 (assuming a known deformation of a rigid or non-rigid body) represents the directivity.

[0058] Since the detection position by the sensor 106 and the arrangement of the attenuator 108 are known, triangulation is possible. Even if the position of the attenuator 108 is unknown, if its shape and operation are known, the set of its shadows 200 is sufficient to determine the position of the radiation source 900. Since a large number of radiations are generated from various radiation sources, tomography is used instead of trying to resolve each radiation source individually. The statistics can be constrained by the optimization for reconstructing the radiation source 900 by tomography. The optimization includes the aforementioned auxiliary conditions, solves the maximum information quantity question (MIQ), and in this case, the directivity is extracted from the non-local pattern of the PSF over the entire field-of-view coding time-varying aperture pattern based on the positive semi-definite detection event data.

[0059] The image processor 120 is configured to perform reconstruction. The detected radiation has time and position. Energy information is included, for example, for filtering the radiation into a specific energy or energy range. The optimization is applied to minimize the difference between the detected radiation and the distribution of the radiation sources that would have caused the radiation. The optimization includes forward and inverse projections between the radiation or sensor space and the object or image space. The projection includes a system model. The system model includes the time-coded aperture (e.g., the position and orientation of the attenuator 108 with respect to the sensor 106). The physical characteristics of the local pattern (shadow 200) are part of the system model. The point spread function determined by the time-coded aperture is used in the reconstruction. By iteratively optimizing, the difference is minimized until the reconstruction by the pixels or voxels of the object results. Various optimizations can be used, such as the various optimizations used in SPECT.

[0060] The reconstruction may be performed in other ways. In one embodiment, the image processor 120 forms virtual projections from separate views. The representation of the patient 114 (i.e., the distribution of the radiation sources in the patient 114) is reconstructed from the virtual projections. The virtual projections are formed from the radiation detected by the sensor 106 using time-encoded apertures. The radiation data is rebinned, resampled, and / or collected for different angles or views to form the projections. The projections are formed from the radiation, and the reconstruction of the spatial distribution is from the projections.

[0061] The virtual PSF is restored by measuring the edge response (shadow 200) of the attenuator 108 on the sensor 106. The time encoding is extended from small movements in front of the sensor 106 to any movement of multiple apertures (holes 202 of one or more shapes and / or sizes), which generates different patterns on the sensor 106 for each stationary view. For example, k distinct patterns can be generated by tilting, rotating, and / or shifting a plate with slits in front of the sensor 106. If the movement is known, it is also known that the patient 114 has not changed during the dwell time (at the sampling time determined by quasi-steady state, required contrast, noise, etc.), and the virtual PSF is constructed. The virtual PSF mimics a non-local good-resolution PSF. The virtual PSF is based on the following inversion: JPEG2025519689000002.jpg15134where Sigma1 (σ1) and Sigma2 (σ2) define Gaussian curves, ERF is the error function based on the integral of the Gaussian, and Mu (μ) is the linear attenuation coefficient. FIG. 10 shows an example plot of the virtual PSF. The PSF models a sharp edge to determine directivity.

[0062] In one embodiment, the projection is formed as a virtual parallel hole collimator projection. The detected radiation and time-encoded apertures are used to determine a projection having a common directivity for each of the different views with respect to patient 114. Adding a convolution PSF of a Heaviside function (a Gaussian having σ1, σ2 in the example of FIG. 10) results in the restoration of a quasi-planar projection of a regular parallel hole collimator at the field of view angle. This same process is repeated at different field of view angles until the tomography conditions of Orlov, Tuy, and Nyquist are met. A projection view set is restored that is as if obtained from a very high-resolution collimator having a much larger aperture. Since the directivity is based on the edge response, the resolution is increased. Since the aperture (hole 202) is larger than a parallel hole collimator, greater sensitivity is provided. Projections suitable for tomography other than parallel hole collimator projections can be formed. The field of view angle requirements can be relaxed while still meeting the tomography conditions.

[0063] Image processor 120 uses the projections for successive approximation reconstruction. For example, the virtual parallel hole collimator projections are reconstructed using forward and back projections. Any of the currently known or later developed emission tomography or SPECT reconstructions can be used to perform the reconstruction from the projections. The projections represent the directional counts from a gamma camera provided as if a collimator were used. SPECT successive approximation reconstruction is performed using a virtual PSF.

[0064] This example of reconstruction uses a two-step process. First, a virtual projection (e.g., a virtual parallel hole collimator projection image) is restored. Second, reconstruction is performed on that virtual projection. In the first step, the generated local pattern can be restored to virtual non-local image formation. Artificial intelligence (AI) (e.g., a machine learning model such as a deep learning model) can be used to form projections from the radiation data. A time-encoded aperture is used as an input to the AI with the radiation data, or the AI is trained with a known time-encoded aperture and thus only the radiation data is input. The AI outputs one or more projections. Convolutional integration can be used to model the complex decision-making process of an observer, thereby optimizing the image formation design.

[0065] Figure 11 shows an example of a flowchart of a method of emission tomography (e.g., SPECT). A time-encoded aperture is formed by operating an attenuator and / or a sensor relative to each other. By providing holes in the attenuator, a lot of information is provided for edge response measurement. Since the direction is not very limited, the reconstruction uses a step-by-step approach in which a virtual projection is formed, and then an object or an image is reconstructed from the virtual projection. In another reconstruction approach, optimization using a time-encoded aperture as part of a system model is used.

[0066] This method is performed by an emission tomography imaging device or system such as an SPECT system. A motor operates the attenuator, and the sensor detects radiation using the attenuator at different positions relative to the sensor. The image processor performs reconstruction from the detected radiation and the time-encoded aperture (e.g., a pattern from the shadow of the attenuator). The image processor generates an image of the reconstructed object, and this image is shown by a display device or screen. Other components can be used to execute and / or assist in any of the processes.

[0067] Each process is executed in the order shown (i.e., top to bottom, or in numerical order) or in another order. Processes 1102 and 1104 are repeatedly executed any number of cycles, such as by operating the attenuator and then performing detection, and then operating and performing detection again. It is also possible to operate the attenuator in process 1102 after first executing process 1104, and vice versa.

[0068] Additional processes, different processes, or fewer processes may also be provided. For example, processes such as moving the sensor, positioning the patient, and / or performing three-dimensional rendering from the reconstructed volume distribution may be included. As another example, the display process 1112 is not executed, and for example, the image or reconstruction is stored in a radiology report, an electronic patient medical record, or an image management system (PACS) or transferred for storage.

[0069] In process 1102, an attenuator (e.g., an attenuator or a radiation blocker) is operated. The attenuator has an inner edge. The attenuator is operated between the patient and the sensor. With or without an outer edge, the inner edge forms a time-coded aperture on the sensor due to the operation. The pattern or shadow from the attenuator is shifted two or more times to provide a time-coded aperture. By this operation, the inner edge and, if any, the outer edge from the hole will have different positions, shapes, sizes, orientations, and / or angles (see Figure 2).

[0070] The operation is a rotation and / or translation along one or more (e.g., three) dimensions. The operation by three-dimensional translation or rotation broadens the diversity of the shadows. Operating in both rotation and translation for the three dimensions further broadens the diversity. Shifting or changing one or more holes adds diversity to the shadows. Moving the sensor also leads to an increase in diversity. The shape and / or size of the holes in the shadows on the sensor are different over time.

[0071] The movement can be by translation and / or rotation in a plane parallel to the detection surface of the sensor. The movement can be a translation, tilt, or sway that deviates from the plane. In another embodiment, the movement is the rotation of a damper around the sensor.

[0072] In process 1104, the sensor detects radiation from the patient. Some of the radiation passes through the sensor. Other radiation is blocked by the attenuator. Other radiation passes through the attenuator, for example through the outer or through holes, and reaches the sensor. The attenuator projects a shadow onto the sensor based on the radiation.

[0073] The detection is performed using an attenuator in a predetermined arrangement. Next, the detection is performed using an attenuator in a different arrangement. Any period is used for the detection. The energy, time, and position on the sensor of each detected radiation are used to identify the detection event. Counts at different positions on the sensor are performed for a given attenuator arrangement when many radiations occur.

[0074] Since the detected radiation is only the incident radiation passing through the attenuator, different shadows result from the attenuator at different positions. The detected events reflect these shadows and enable an edge response. Many events are detected by the inner edge, and further edges are provided to determine the directivity by the edge response. For each position of the aperture formed by the attenuator, a time-coded aperture and the corresponding detection event are recorded. The movement of the damper results in different shadows on the sensor based on the detection by the sensor. The orientation and / or position of the damper with respect to the sensor vary over time according to the movement in process 1102.

[0075] In process 1106, the image processor reconstructs the patient's representation from the detected radiation using a time-coded aperture. The edge response of the shadow is used to determine or limit the directivity in the optimization. The edge is limited to an angle within a certain range of angles. By having an edge at different positions relative to the sensor at each time, the edge response indicates the position of the radiation source based on the detected position on the sensor. The representation is reconstructed from the edge response of the shadow of the attenuator on the sensor.

[0076] In one embodiment, optimization is used when the shadow or time-coded aperture is part of the system model. The events detected at different times and the time-coded aperture are used to solve for the spatial distribution of the radiation source in the patient.

[0077] In another embodiment, a two-stage approach represented by processes 1108 and 1110 is used. In process 1108, projections at different field-of-view angles with respect to the patient are constructed. The projections are constructed from the detected radiation based on the time-coded aperture. For example, a virtual PSF is created by integration using a Gaussian model. The virtual PSFs from different directions are determined by ray tracing and accurate recording of signal changes under known non-negativity conditions. In process 1110, the projections are used as samples or binned counts from the sensors in different views. Iterative optimization is performed with the projections as input. The iterative optimization solves for the spatial distribution of the radiation source from the projections based on the radiation from the radiation source.

[0078] In process 1112, an image is generated from the reconstruction. The spatial distribution represents the distribution of the radiation source in a plane or volume. The representation is reformatted to be displayed as an image on a two-dimensional display. In the case of a two-dimensional representation, the reformatting scan is converted to the resolution and / or dynamic range of the display. In the case of a three-dimensional representation, the voxels are rendered into a two-dimensional image using, for example, volume or surface rendering.

[0079] The resulting image is displayed on the display screen. The physician can view the location and / or intensity of the uptake of the radiopharmaceutical by the patient's tissue. Hot spots or locations represent the function of the tissue, enabling the physician to identify dysfunctions or areas of insufficiency that may be the subject of diagnosis and / or treatment.

[0080] Instead of using a sensor with a fixed collimator, a combination of which is moved to detect from different angles, the relative movement of the attenuator with respect to the sensor is provided, adding a change in pattern. The attenuator projects a shadow on the sensor. By operating the attenuator in a known manner (e.g., translation, rotation, and / or deformation), the shadow of the attenuator on the sensor changes, generating a changing shadow pattern on the sensor. By including a plurality of holes and / or edges of different known sizes, shapes, and angles, much of the radiation passes through the attenuator with additional information arising from the edges. The radiation pattern is estimated by viewing from different directions towards the radiation source and collecting information from different shadows. To solve for the distribution of radiators within the radiation source (patient), more signals and resultant information are used with known aperture encoding. Instead of observing through a collimator that only looks at what is in front and rotating the sensor and collimator together to capture more viewing angles, the sensor remains stationary and captures information by the operation of the attenuator.

[0081] Although the present invention has been described with reference to various embodiments, many changes and modifications can be made without departing from the scope of the present invention. Therefore, the above detailed description is to be construed as illustrative rather than limiting the present invention, and it is to be understood that it is the claims that are intended to define the spirit and scope of the present invention, including all equivalents.

Claims

**Claim 1** A positron emission tomography (PET) system comprising: a sensor configured to detect the position, energy, and time of gamma-ray collisions; a movable attenuator having one or more inner through-holes and movable relative to the sensor; a drive device configured to operate the movable attenuator; an image processor configured to reconstruct the spatial distribution of radiation detected by the sensor using the movable attenuator that takes different positions by the operation of the drive device; and wherein the movable attenuator is positioned between the radiation source and the sensor such that the moving shadow of the through-hole is projected onto the sensor. **Claim 2** The PET system according to claim 1, wherein the sensor includes a planar gamma camera. **Claim 3** The PET system according to claim 2, wherein the planar gamma camera is connected to a gantry configured to position the planar gamma camera at different positions relative to the radiation source for detecting the radiation. **Claim 4** The PET system according to claim 1, wherein the movable attenuator includes a lead or tungsten object. **Claim 5** The PET system according to claim 1, wherein the movable attenuator is operable in three-dimensional translation and / or rotation. **Claim 6** The PET system according to claim 1, wherein the through-hole includes a slit. **Claim 7** The PET system according to claim 1, wherein the through-hole includes different sizes, shapes, and / or hole angles. **Claim 8** The PET system according to claim 1, wherein the movable attenuator includes a rotatable cylinder, and the radiation source or the sensor is positionable inside the rotatable cylinder; and the drive device is configured to rotate the rotatable cylinder. **Claim 9** The PET system according to claim 1, wherein the drive device is configured to rock the movable attenuator. **Claim 10** The PET system according to claim 1, wherein the drive device is configured to swing the movable attenuator about a perpendicular line to the movable attenuator. **Claim 11** The PET system according to claim 1, wherein the image processor is configured to form a projection from the radiation, and the reconstruction of the spatial distribution is from the projection. **Claim 12** The radiation tomography system according to claim 11, wherein the projection is a virtual parallel hole collimator projection, and the reconstruction is a sequential approximation reconstruction using the virtual parallel hole collimator projection in forward projection and back projection.

13. A method of single photon emission computed tomography (SPECT), comprising: operating an attenuator having an inner edge between a patient and a sensor, the operating inner edge forming a time-coded aperture on the sensor; detecting, by the sensor, radiation from the patient passing through the attenuator that provides different shadows on the sensor by the time-coded aperture; reconstructing a representation of the patient from the detected radiation using the time-coded aperture.

14. Operating the attenuator includes rotating and / or translating in three dimensions, The method according to claim 13, wherein detecting the radiation includes detecting using the different shadows that vary depending on a position and / or rotation of the attenuator on the sensor.

15. The method according to claim 13, wherein operating the attenuator includes operating such that the inner edge forms holes having different shapes, sizes, and / or angles.

16. The holes of the attenuator form the inner edge, The method according to claim 13, wherein operating the attenuator includes operating in three dimensions such that the shape and / or size of the holes in the shadow vary over time.

17. The method according to claim 13, wherein reconstructing the representation includes restoring a virtual point spread function from an edge response of the shadow.

18. The method according to claim 13, wherein reconstructing the representation includes constructing projections at different field of view angles for the patient based on the time-coded aperture from the detected radiation and reconstructing from the projections.

19. A radiation tomography system, comprising: a radiation blocker having an inner edge; The inner edge forms a hole passing through the radiation blocker; a sensor configured to detect radiation passing through the hole; The radiation blocker takes different positions with respect to the sensor, and the different positions form a time-coded aperture with respect to the sensor. An imaging processor configured to form virtual projections from different views of the radiation detected by the sensor using the time-coded aperture and to reconstruct a representation of a patient from the virtual projections, and a computed tomography system including the same.

20. The computed tomography system according to claim 19, wherein the holes have different sizes, shapes, and / or angles.

Images