Hybrid linearization scheme for x-ray CT beam hardening correction

Abstract

To provide a hybrid linearization scheme for x-ray ct beam hardening correction.SOLUTION: Disclosed herein are methods for reducing beam-hardening artifacts in CT imaging using a mapping operator that comprises a hybrid spectral model that incorporates air scan X-ray intensity data acquired at two different effective mean energies. In one variation, the air scan X-ray intensity data acquired during a calibration session is combined with an ideal spectral model for each X-ray detector to derive a hybrid spectral mode. A mapping operator based on the hybrid spectral model is used to correct beam-hardening artifacts in acquired CT projection data. In some variations, the mapping operator is a lookup table of monochromatic (corrected) projection values, and the acquired CT projection data is used to calculate an index of a lookup table entry that contains the corrected projection value that corresponds with the acquired CT projection data.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] Cross - reference to Related Applications This application claims priority to U.S. Provisional Patent Application No. 63 / 006,513, filed Apr. 7, 2020, the disclosure of which is hereby incorporated by reference in its entirety.

Background Art

[0002] An X - ray computed tomography (CT) imaging system includes an array of X - ray detectors positioned opposite an X - ray source and can generate an image of an object positioned between the X - ray source and the detectors based on the attenuation of X - rays passing through the object. The X - ray detectors are positioned at various angular positions relative to the object. A CT image represents the attenuation characteristics of the object and can be determined by measuring the intensity of the X - rays incident on the detectors. Ideally, X - ray attenuation is linear and the X - ray intensity measured by the detectors is linearly related to the attenuation characteristics of the object. However, since the X - ray source of a CT system is typically a polychromatic radiation source, it emits photons of different energies, and photons of different energies attenuate in different ways when passing through the object. In many materials, lower - energy photons are attenuated more than higher - energy photons. Since lower - energy photons are more attenuated in the central part of the object than at the edges, this non - uniform attenuation across different photon energies can cause the edges of the object to appear brighter than the central part of the object (even if it is made of a uniform material). The imaging artifacts resulting from this non - linear attenuation are called beam hardening artifacts. Typically, beam hardening artifacts are corrected by collecting X - ray detector data during a calibration session using several phantoms made of different materials of different thicknesses to generate a mapping function that maps polychromatic projection data to ideal or corrected monochromatic projection values. However, this type of calibration and artifact correction method can be tedious and time - consuming because it requires multiple CT scans of multiple phantoms with different dimensions and / or materials.

[0003] Furthermore, images reconstructed from a CT imaging system in which an X-ray source and an array of X-ray detectors rotate during image acquisition (e.g., the X-ray source and detectors are attached to a rotating gantry) can be particularly sensitive to the different spectral characteristics or responses of individual X-ray detectors.

[0004] Therefore, it is desirable to improve the calibration method of the CT imaging system to help address beam hardening artifacts and / or compensate for the variable spectral responses of individual X-ray detectors within the detector array. SUMMARY OF THE INVENTION MEANS FOR SOLVING THE PROBLEM

[0005] Disclosed herein is a method for reducing beam hardening artifacts in CT imaging using a mapping operator that includes a hybrid spectral model incorporating air scan X-ray intensity data acquired at two different effective average energies. The mapping operator can be calculated for each X-ray detector of a CT imaging system. The mapping operator can be used to transform the acquired CT projection data (calculated based on the output signals from the corresponding X-ray detectors) into corrected projection values. In some variations, the mapping operator can include a look-up table (LUT), and a look-up table index can be calculated using the acquired CT projection data to identify the corrected projection values corresponding to the acquired CT projection data.

[0006] Also disclosed herein is a calibration method without a phantom for reducing beam hardening artifacts. A hybrid spectral model can be generated that maps polychromatic projection data to calculated monochromatic projection values using X-ray intensity data acquired during a calibration session. The hybrid spectral model includes an ideal spectral model combined with air scan X-ray intensity data acquired at two different effective average energies. The hybrid spectral model is used to generate a mapping operator that can be used during a CT scan to convert the acquired CT projection data into corrected projection values that can then be used for CT image reconstruction. During the calibration session, the CT imaging system acquires a first set of X-ray intensity data during a first air scan at a first effective average energy and may acquire a second set of X-ray intensity data during a second air scan at a second effective average energy higher than the first effective average energy. The X-ray intensity data for the first and second air scans can be acquired and stored for each X-ray detector, and in the case of a rotational CT imaging system, the X-ray intensity data can be acquired while the CT imaging system rotates through a plurality of angles. The CT imaging system controller can then calculate a virtual filter for each X-ray detector using the first and second sets of X-ray intensity data. In some variations, the virtual filter can be characterized by a selected filter material (e.g., aluminum) and filter thickness, and the first and second sets of X-ray intensity data can be used to calculate the thickness of the virtual filter for each X-ray detector in the array. The CT imaging system controller can then calculate a mapping operator for each X-ray detector that uses the virtual filter in combination with the ideal spectral model to form a hybrid spectral model. The mapping operator represents the relationship between the polychromatic projection data derived from the acquired X-ray detector intensity data and the corrected projection values. Since the corrected projection values approximate the projection values of a monochromatic X-ray source (i.e., one that emits photons at a single energy level or light of a single wavelength), the corrected projection values may also be referred to as monochromatic projection values and have little or no beam hardening artifacts.

[0007] One variation of a method for reducing beam hardening artifacts in CT imaging may include acquiring CT projection data at each x-ray detector in a CT imaging system, determining a corrected projection value for each of the acquired CT projection data using a mapping operator including a hybrid spectral model incorporating air-scan CT projection data acquired at two different effective average energies, and generating an artifact-corrected CT image by combining the corrected projection values ​​for each of the acquired CT projection data. p The corrected projection value p m The mapping operator may include a virtual filter expressed as a function of s and calculated based on the acquired air-scan x-ray intensity data. The mapping operator may include a look-up table LUT for each x-ray detector in the CT imaging system, each look-up table LUT representing the CT projection data divided by a discretization step size s.

number

number

[0008] In some variations, the look-up table may be a first look-up table LUT_1 for a first CT scan energy level, and the mapping operator may include a second look-up table LUT_2 for a second CT scan energy level, where the second look-up table LUT_2 is CT projection data divided by a discretization step size s'.

Number

[0009] In some variations, multiple hybrid spectral models may be calculated for different CT scan energy levels. In one variation, the hybrid spectral model may be a first hybrid spectral model for a first CT scan energy level, and the mapping operator may include a second hybrid spectral model for a second CT scan energy level. The corrected projection value p for the acquired CT projection data p p m ​​Determining may include identifying the CT scan energy level at which the CT projection data was acquired and calculating a corrected projection value p using a hybrid spectral model corresponding to the identified CT scan energy level. m and calculating.

[0010] In some variations, the air scan x-ray intensity data may be acquired at a first effective average energy and a second effective average energy. For example, the first energy may be 80 kVp, the second energy may be 140 kVp, and the CT projection data acquisition may be at an energy level of 120 kVp.

[0011] One variation of a method for phantomless calibration of a CT imaging system to reduce beam hardening artifacts includes acquiring a first set of x-ray intensity data for each x-ray detector of the CT imaging system during a first air scan at a first effective average energy, acquiring a second set of x-ray intensity data for each x-ray detector of the CT imaging system during a second air scan at a second effective average energy higher than the first effective average energy, using the first and second sets of x-ray intensity data to calculate a virtual filter for each x-ray detector of the CT imaging system, and calculating a mapping operator for each x-ray detector of the CT imaging system, the calculating including a hybrid spectral model in which the mapping operator uses the virtual filter to convert polychromatic projection values to monochromatic projection values. Each virtual filter may be made of a selected material and may have a thickness. Calculating a virtual filter for each x-ray detector may include calculating the thickness of the virtual filter. In some variations, calculating the mapping operator includes calculating, using a hybrid spectral model, a discrete set of k polychromatic projection values each divided by a discretization step size s

Number

Number

[0012] In some variations, calculating the virtual filter thickness for each of the i X-ray detectors involves the X-ray intensity data obtained at a second, higher effective average energy

Number

Number

Number

Number

Number

Number

Number

Number

[0013] Hybrid spectrum model

Number

Number

Number

Number

[0014] In some variations, calculating the mapping operator may include repeatedly calculating the monochromatic projection value (p p ) corresponding to the polychromatic projection value (p m ) based on the hybrid spectrum model,

Number

Number

Number

Number

[0015] In some variants, the CT imaging system may comprise a rotating gantry and an imaging X-ray source attached to the gantry, the X-ray detector may be attached to the gantry on the opposite side of the imaging X-ray source, and acquiring a first set of X-ray intensity data may include rotating the gantry during a first air scan. Acquiring a second set of X-ray intensity data may optionally include rotating the gantry during a second air scan. In some examples, the first energy may be 80 kVp and the second energy may be 140 kVp.

Brief Description of the Drawings

[0016]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5A

Figure 5B

Figure 6A

Figure 6B

Figure 6C

Figure 7A

Figure 7B

Figure 7C

Best Mode for Carrying Out the Invention

[0017] Disclosed herein is a method for reducing beam hardening artifacts in CT imaging using a mapping operator that incorporates air scan x-ray intensity data acquired at two different effective average energies. The air scan x-ray intensity data is acquired during a calibration session and combined with an ideal spectral model for each x-ray detector to derive a hybrid spectral model (i.e., a “hybrid” or combination of the ideal spectral model and empirical data). The combination of the acquired air scan x-ray intensity data and the ideal spectral model can simulate the variability inherent to each of the x-ray detectors, and thus the hybrid spectral model can provide a better approximation of the projection data from the x-ray detectors. Using the hybrid spectral model, a mapping operator can be generated that associates projection data calculated from intensity data from an x-ray detector (which may be affected by beam hardening artifacts from a polychromatic x-ray source) with corrected projection values representative of the detector's response in the absence of beam hardening artifacts. Projection data calculated from the x-ray intensity data may also be referred to as polychromatic projection data. The corrected projection values may also be referred to as monochromatic projection values, since they approximate the projection values of a CT system having a monochromatic x-ray source (presumably without beam hardening artifacts in the CT projection data). In some variations, the mapping operator can be a look-up table (LUT), and the polychromatic projection data can be used to calculate a table index that is used to identify the corresponding monochromatic (i.e., corrected) projection value calculated using the hybrid spectral model. The CT imaging system controller can then combine the corrected projection values for each x-ray detector within the array of x-ray detectors to generate a CT image with little or no beam hardening artifacts.

[0018] In some variations, for each x-ray detector, a virtual filter for that x-ray detector can be calculated using air scan x-ray intensity data acquired at two different effective average energies. The virtual filters represent or approximate the spectral variability in the sensitivity of the x-ray detectors and respond to incident photons that can occur, for example, due to manufacturing variations or inherent structural characteristics. The virtual filters can also represent or approximate the spectral variations of individual x-ray detectors because of their different locations relative to the x-ray source. In a CT imaging system, an array of x-ray detectors is located opposite the x-ray source and can span the expected field of view provided by the radiation beam of the x-ray source (e.g., in the case of a fan-shaped radiation beam, the array of x-ray detectors can be arranged along an arc corresponding to the fan-shaped beam). In some variations, a bowtie filter can be positioned in front of the x-ray source, and since the central portion of the filter is thinner than the peripheral portion of the filter, the spectral characteristics of the radiation incident on the x-ray detectors located in the central portion of the radiation beam can be different from the spectral characteristics of the radiation incident on the x-ray detectors located around the radiation beam. For example, the x-ray intensity data (and corresponding projection data) from the x-ray detectors around the detector array can be more affected by the beam hardening effect from the bowtie filter than the x-ray intensity data (and corresponding projection data) from the x-ray detectors in the center of the array. The virtual filters for each x-ray detector calculated from the calibration air scan can account for these differences, resulting in a hybrid spectral model that better approximates the actual spectrum of the incident radiation and the output of the x-ray detectors than a spectral model without virtual filters.

[0019] "Projection data" as used throughout this document represents the degree to which x-rays are attenuated as they pass through a material (e.g., the material of a scanned object). A projection data value may be a quantity indicative of the strength of attenuation of the material, i.e., how much the energy and / or intensity of x-rays is reduced as they pass through the material. A CT image is a combination (e.g., reconstruction) of projection data collected from x-ray detectors in a detector array. Projection data may be calculated from intensity data acquired by the x-ray detector. For example, intensity data acquired by the x-ray detector in the presence of a scanned object (I) and intensity data acquired by the x-ray detector in the absence of a scanned object (I0) may be used to calculate CT projection data (p) for that x-ray detector.

number

[0020] Acquired CT projection data (or acquired projection data) may refer to projection data calculated based on intensity data measured by an X-ray detector. Because this acquired projection data is calculated based on actual intensity measurements from the X-ray detector, it may be subject to beam hardening effects. Without beam hardening correction, a combination of (e.g., reconstruction from) the projection data acquired across all X-ray detectors of a detector array may result in an image with beam hardening artifacts. In CT imaging systems where the imaging radiation source is a polychromatic X-ray source, the acquired CT projection data may be referred to as polychromatic projection data (p p ) is also sometimes referred to as

[0021] Corrected projection values ​​may refer to projection values ​​calculated based on the hybrid spectral model described herein, where beam hardening artifacts and / or spectral variations of individual x-ray detectors are reduced or eliminated. Corrected projection values ​​approximate projection values ​​of a CT system with a monochromatic x-ray source, and therefore are referred to as monochromatic projection values ​​(p m ) is also sometimes referred to as

[0022] The beam hardening correction and calibration methods disclosed herein are described in the context of a CT imaging system having a rotational polychromatic x-ray source and an array of x-ray detectors that rotate in concert with the x-ray source. However, it should be understood that these methods can be used in non-rotational CT imaging systems, e.g., not only in a linear translation CT imaging system where the x-ray source and detector travel along the object to be scanned, but also in a CT imaging system where the detector is fixed (e.g., a complete 360° ring, a stationary frame, or multiple arcs surrounding the circumference of a gantry), and the x-ray source is rotatable (e.g., mounted on a rotational gantry). The exemplary systems provided herein include a polychromatic x-ray source that emits fan beam radiation, but the same methods can be used in systems having parallel beam and / or cone beam polychromatic x-ray sources. The beam hardening correction and calibration methods described herein can be used in a CT imaging system that is part of a radiation therapy system, as well as in a stand-alone CT imaging system. For example, these methods can be used to calibrate and generate CT images of a radiation therapy system that includes a CT imaging system (e.g., having a rotational imaging radiation source and detector, and a CT imaging system controller), and a therapeutic radiation source that is movable around a patient. The therapeutic radiation source can be mounted on a rotatable and / or extendable C-arm, an articulated robotic arm, and / or a rotational circular gantry. Some variations of the radiation therapy system can optionally include an array of additional x-ray detectors opposite the therapeutic radiation source. An example of a radiation therapy system can include a CT imaging system and a therapeutic radiation source, both mounted on a rotational gantry. The rotational gantry can be configured to rotate rapidly, e.g., at about 20 RPM to about 70 RPM, e.g., at about 60 RPM. In some variations, the radiation therapy system can include a CT imaging system and a therapeutic radiation source, both mounted on a rotational gantry, and optionally one or more arrays of PET detectors that can be mounted on the same rotational gantry.Further details of various radiation therapy systems can be found in U.S. Patent Application No. 15 / 814,222, filed November 15, 2017, which is hereby incorporated by reference in its entirety.

[0023] System Figure 1 is a schematic diagram of one variant of a CT imaging system (100). The CT imaging system (100) can include an X-ray source (102), an array of X-ray detectors (104), and a CT imaging system controller (not shown) that communicates with the X-ray source (102) and the array of X-ray detectors (104). In some variants, the X-ray source (102) can be a polychromatic X-ray source configured to emit X-rays across a spectrum of X-ray energies. The data output from each of the X-ray detectors in the array can represent the intensity values of the X-rays incident on each of the X-ray detectors. The X-ray source (102) and the array of X-ray detectors (104) can be mounted on a rotating gantry configured to rotate around the object to be scanned (106) and acquire X-ray intensity data from various angles. In some variants, the CT imaging system can include a housing surrounding the X-ray source, detector array, and rotating gantry, and the housing can be shaped to define a bore through which a scan platform can move. The scan platform can be moved along the longitudinal central axis of the bore, which can also be the axis of rotation of the rotating gantry.

[0024] The CT imaging system may also include one or more spectral components disposed in the X-ray source beam path. The one or more spectral components may be configured to modify the spectral characteristics of the radiation beam (101) that passes through the scan object (106), which may be a patient. For example, the CT imaging system may include one or more filters, which may be inherent to the X-ray source or separate from or external to the X-ray source. One or more filters and / or collimators disposed on the X-ray detector may optionally be present and may assist in removing scattered radiation and / or selecting specific spectral characteristics. In the variant depicted in FIG. 1, the X-ray source (102) may have one or more inherent filters (108). An optional bowtie filter (110) may be disposed in the radiation beam path (101) external to the X-ray source (102). In some variants, the scan object 106 may be located within the bore of the CT imaging system, and a bore window (not shown) may be present within the radiation beam path (i.e., the path through which the radiation beam exits the bowtie filter). The X-ray spectrum of a polychromatic X-ray source may be represented by the X-ray tube spectrum S(E), and the attenuation of the inherent filter (108) may be represented by exp(-μ int (E)T int ), the attenuation of the bowtie filter (110) may be represented by exp(-μ bow (E)T bow ), the attenuation of the bore window may be represented by exp(-μ bore (E)T bore ). The energy response of the X-ray detectors in the array may be represented by E(1 - exp(-μ det (E)T det ). The coefficient μ is the attenuation coefficient of that particular spectral component, which may vary depending on the material constituting the spectral component or, in the case of the detector, the detector scintillator, and the quantity T represents the thickness of the component (or the path length of the X-rays passing through the component). The attenuation coefficient μ and the thickness T may be known as part of the system design that specifies the material and dimensions (i.e., the overall shape and arrangement) of each component. The X-ray tube spectrum S(E) may also be known, characterized, and / or measured using simulations.

[0025] When there is no object to be scanned (106) in the bore, the model Seff(E) of the effective spectrum of the CT imaging system (100) can be represented by Equation (1). S eff (E)=S(E)exp(-μ int (E)T int )exp(-μ bow (E)T bow )exp(-μ bore (E)T bore )E(1-exp(-μ det (E)T det ))

[0026] As described above, since the relative positions of the respective X-ray detectors with respect to the X-ray source (i.e., the i-th X-ray detector in the array) are different, the effective spectra at the respective X-ray detectors are different

Equation

Equation

[0027] Method A method for correcting beam hardening artifacts may include generating a hybrid spectral model for each X-ray detector that accounts for the variability of the X-ray detector. In some variations, the variability of an individual X-ray detector may be represented by a virtual filter made of a selected material and having a thickness such that the attenuation of X-rays passing through the virtual filter approximates the variability of the spectral response of that X-ray detector. One or more of the parameters of the virtual filter, such as the material and / or thickness of the filter, may be determined using X-ray intensity data acquired during a calibration session. The empirical data acquired during the calibration session may reflect the actual spectral response of the individual X-ray detectors within the detector array and, in combination with an ideal spectral model, may generate a hybrid spectral model that better approximates the actual response of the individual X-ray detectors.

[0028] One variation of the calibration method may include acquiring X-ray intensity data using the X-ray detectors within the detector array at two different effective average energies of the imaging X-ray source. In some variations, the X-ray source and the array of X-ray detectors may rotate around the bore of the system and acquire X-ray intensity data at multiple angles around the bore. For example, the X-ray intensity data may be acquired for each X-ray detector at up to about 1200 locations around the scan region, such as about 4 locations (0°, 90°, 180°, 270°) to about 360 locations, about 370 locations to about 700 locations, about 800 locations to about 1000 locations, about 720 locations, 960 locations, etc., while rotating around the scan region more than once (e.g., rotating around the scan region once to about 200 times). The X-ray intensity data may be acquired with no object in the scan region, i.e., an air scan.

[0029] Next, using the X-ray intensity data acquired during the calibration session, a virtual filter for each individual X-ray detector that models the variability of the X-ray detector can be calculated. The virtual filter, in combination with an ideal spectral model for each individual X-ray detector, can generate a hybrid spectral model for that X-ray detector. Next, using the hybrid spectral model, an individual X-ray detector mapping operator that represents the relationship between the polychromatic projection data (i.e., the acquired CT projection data that may include beam hardening artifacts and variations of individual X-ray detectors) and the corresponding monochromatic projection values (i.e., the corrected projection values with reduced or eliminated beam hardening artifacts and / or X-ray detector variations) can be calculated. The mapping operator can be applied to the CT projection data obtained by each individual X-ray detector during the patient's imaging session to determine the corresponding corrected projection values. Next, using the corrected projection values, a CT image with reduced or removed beam hardening artifacts can be reconstructed.

[0030] One variant of a calibration method for obtaining X-ray intensity data that can be incorporated into a hybrid spectral model that can be used for correction (e.g., reduction) of beam hardening artifacts is shown in FIG. 2. The calibration method (200) includes obtaining a first set of X-ray intensity data (202) for each X-ray detector in the imaging system during an air scan at a first effective average energy, obtaining a second set of X-ray intensity data (204) for each X-ray detector in the imaging system during an air scan at a second effective average energy, using the first and second sets of X-ray intensity data to calculate a virtual filter (206) for each X-ray detector in the imaging system, and calculating a mapping operator (210) for each X-ray detector in the imaging system. The mapping operator may include a hybrid spectral model that uses the virtual filter to convert polychromatic projection values to monochromatic projection values. The acquisition of the X-ray intensity data (202, 204) can be performed as an air scan, i.e., with no phantom in the scan region, but in other variants, the X-ray intensity data can be acquired at two different effective average energies with a phantom in the scan region. The first and second sets of X-ray intensity data can include X-ray intensity data acquired from a plurality of angles around the scan region. For example, in the case of a CT imaging system where the X-ray source and / or X-ray detector is attached to a movable (e.g., rotational) gantry, the X-ray intensity data for individual detectors can be acquired from a plurality of angles and optionally during multiple rotations around the scan region. For example, the acquisition of the X-ray intensity data (202, 204) can occur while the X-ray source and / or X-ray detector rotates around the scan region at least once, and in some variants, the X-ray intensity data can be acquired while the X-ray source and / or X-ray detector rotates around the scan region two or more times. Calculating a mapping operator (208) for each X-ray detector can include calculating different mapping operators for each CT scan energy level used during the imaging session.For example, when it is desired to calibrate a CT imaging system to enable a user to perform imaging scans at a first scan energy level (e.g., 80 kVp) and also at a second scan energy level (e.g., 120 kVp), calculating the mapping operator (208) may include calculating a first mapping operator for the first scan energy level and calculating a second mapping operator for the second scan energy level. The first and second mapping operators for each X-ray detector may be stored in the memory of the CT imaging system controller. During an imaging session, when the user selects a desired scan energy level, the CT imaging system controller may be configured to select from the two mapping operators (i.e., select the mapping operator calculated for that scan energy level) to convert the acquired CT projection data into corrected projection values.

[0031] Method (200) may optionally further include generating (210) a look-up table having k corrected projection values (p m ) for each X-ray detector within the imaging system, each of the k corrected projection values being calculated from a discrete set of k polychromatic projection values, each divided by a discretization step size (s), using the mapping operator. As described above, in a variant where the CT imaging scan can be performed at two different scan energy levels, two different mapping operators and two different look-up tables are generated and stored in the memory of the CT imaging system controller. During an imaging session, when the user selects a desired scan energy level, the CT imaging system controller may be configured to select the look-up table corresponding to the selected scan energy level to convert the acquired CT projection data into corrected projection values.

Number

[0032] As described above, the virtual filter can be characterized by the selected filter material (e.g., aluminum, graphite) and the filter thickness. The attenuation characteristics of the filter material and the thickness of the filter along the x-ray beam (e.g., beam path length) can approximate or represent the variability of the spectral response of the x-ray detector, e.g., the sensitivity variations in those x-ray detectors, e.g., due to manufacturing variations and / or the location of the individual x-ray detectors relative to the x-ray source of the CT imaging system, and respond to incident photons. FIG. 3 depicts one variant of a method for calculating a virtual filter. The method (300) can be implemented by a CT imaging system controller after x-ray intensity data has been acquired (e.g., with respect to steps 202 and 204 as described above and as depicted in FIG. 2). The method (300) determines, for each of the (i) x-ray detectors of the imaging system, the x-ray intensity data measured at a first, lower effective average energy

Number

Number

Number

[0033] The method (300) can then include setting (304) for each of the (i) x-ray detectors, a spectral model of the x-ray intensity at each of the first and second effective average energies to the LHR measured for each of the (i) x-ray detectors. The x-ray intensity data measured by the individual x-ray detectors is an ideal effective spectrum adjusted by a virtual filter having an attenuation coefficient μ vir (E) of the selected virtual filter material.

Number

Number

Number

[0034] Setting the spectral model of the X-ray intensity data at each of the first and second effective average energies to the measured LHR of an individual X-ray detector (304) can be as follows.

Number

Number

Number

[0035] Next, method (300) can include calculating (306) a thickness for an individual X-ray detector within the detector array

Number

Number

Number

Number

Number

[0036] After the virtual filter thickness has been calculated for each individual X-ray detector, the hybrid spectral model for that X-ray detector

Number

Number

[0037] Different hybrid spectral models can be generated for individual X-ray detectors at different CT scan energy levels. The first hybrid spectral model is the first ideal effective spectrum for the first CT scan energy level

Number

Number

[0038] Method for generating a mapping operator FIG. 4 depicts one variant of a method for generating a mapping operator using a hybrid spectral model. Method (400) uses, for each X-ray detector, a virtual filter representing the variability of each X-ray detector, as represented by equation (2) above, to correct the ideal spectral model

Number

Number

Number

Number

Number

Number

Number

[0039] Equation (3) can be rewritten such that the monochromatic projection value p m is represented in terms of the polychromatic projection value p p (i.e., the acquired CT projection data derived from the X-ray intensity data measured by individual X-ray detectors).

[0040] Alternatively or additionally, method (400) may include generating (406) a mapping operator for each X-ray detector that maps the maximum k monochromatic projection values p m to k polychromatic projection values

Number

[0041] In some variations, the k monochromatic projection values p m are each a discrete set of k polychromatic projection values divided by a discretization step size s

Number

Number

[0042] In some variations, the mapping operator (e.g., a lookup table) can be calculated for each desired CT scan energy level based on different hybrid spectral models for each of those CT scan energy levels. In particular, as long as the spectra of the X-ray source at different CT scan energy levels are characterized (e.g., from the technical specifications or simulations provided by the X-ray source manufacturer) or known, additional X-ray intensity data does not need to be collected during the calibration session to calculate the hybrid spectral models for different CT scan energy levels, and two X-ray intensity data acquired during two air scans at two different effective average energies are sufficient to generate any number of hybrid spectral models for calculating the mapping operator for the desired number of CT scan energy levels. Each of the mapping operators for different CT scan energy levels can be stored in the memory of the CT imaging system controller and retrieved during the imaging session to correct beam hardening artifacts based on the scan energy level selected for the imaging session.

[0043] Method for generating a beam hardening correction lookup table As described above, a mapping operator such as a lookup table for an X-ray detector can include a set of mono-color projection values corresponding to a set of multi-color projection values. The set of multi-color projection values can approximate the range of CT projection data that can be acquired by that X-ray detector. During the imaging session, the acquired CT projection data p p is used to generate a lookup table index, and the acquired CT projection data p pIt is possible to identify the monochromatic projection values in the look-up table corresponding thereto. For each X-ray detector, a plurality of look-up tables can be generated, and each look-up table corresponds to each CT scan energy level that can be used during the imaging session.

[0044] Figures 5A-5B illustrate one variant of a method for a beam hardening correction look-up table. Figure 5A shows a discrete set of polychromatic projection values

Number

[0045] FIG. 5B shows a discrete polychromatic projection value within a set of discrete polychromatic projection values corresponding to a selected polychromatic projection value, e.g., a range of polychromatic projection data (e.g., acquired CT projection data).

Number

Number

[0046] In some variations, the objection function may be defined by equation (4).

Number

[0047] Therefore, the first derivative of the above objection function is as follows.

Number

[0048] Updating the monochromatic values may include the following calculations.

Number

[0049] The criterion used to determine whether the intermediate monochromatic values sufficiently correspond to the selected polychromatic projection values (528) may include calculating the value of the objection function defined by equation (4) using the updated intermediate monochromatic projection values. For example, the CT imaging system controller

Number

[0050] Method for correcting beam hardening for image reconstruction As described above, the mapping operator (e.g., look-up table) generated during the calibration session can be used during the imaging session to correct artifacts from the acquired CT projection data that can arise from beam hardening effects and / or detector variability. During the imaging session, the user can select a CT scan energy level that can define the photon energy emitted by the imaging x-ray source to the object being scanned. This selection can also be used by the CT imaging controller system to select a mapping operator (e.g., look-up table) for each x-ray detector. The imaging session can include acquiring x-ray intensity data for each individual x-ray data, calculating CT projection data (i.e., polychromatic projection data) based on the acquired x-ray intensity data, and determining corrected projection values (i.e., monochromatic projection data) using the selected mapping operator for each individual x-ray detector. Combining the corrected projection values across the x-ray detectors in the array can generate (e.g., for use in reconstructing) an image with reduced (or even eliminated) beam hardening artifacts. In some variations, the mapping operator for each individual x-ray detector can be a look-up table, and the acquired CT projection data can be used to identify an entry in the look-up table that includes the corresponding monochromatic projection value. Alternatively or additionally, the mapping operator for each individual x-ray detector can be one or more mapping functions including the hybrid spectral models (e.g., equations (2)-(4)) described herein.

[0051] FIG. 6A shows one variation of a method for correcting beam hardening artifacts for CT imaging. The method (600) includes acquiring (602) CT projection data (i.e., polychromatic projection data p p ) at each individual x-ray detector in the imaging system, and corrected projection values (i.e., monochromatic projection data p mdetermining (604) using a mapping operator and generating (606) an image using the corrected projection values of each X-ray detector in the imaging system. Obtaining CT projection data can include obtaining or measuring X-ray intensity data with individual X-ray detectors and calculating the CT projection data obtained from the measured X-ray intensity data as described above. For example,

Number

[0052] In some variations, determining the corrected projection value may use a mapping operator such as a look-up table. FIG. 6B shows one variation of a method for determining a corrected projection value using a look-up table. The method (610) may include calculating an index value (612) based on CT projection data obtained for each X-ray detector in the imaging system, and using the index value to identify a corrected projection value in a look-up table that includes a plurality of monochromatic projection values calculated using a hybrid spectral model, such as any of those described herein (e.g., mapping polychromatic projection values to monochromatic projection values based on two air scan data sets) (614). The look-up table may be generated using one or more of the methods described above.

[0053] In some variations, determining the corrected projection value may include calculating a look-up table index j from the acquired CT projection data (i.e., polychromatic projection data). FIG. 6C shows one variation of a method for determining a corrected projection value with a look-up table and calculating an index value using the acquired CT projection data to identify a table entry with the corresponding corrected projection value. The method (620) includes calculating an index value (j) by dividing the acquired CT projection data (p p ) of each X-ray detector in the imaging system by a discretization step size (s), i.e., [Number] and identifying the j-th entry in a look-up table that includes up to k corrected projection values calculated from a discrete set of k polychromatic projection values each divided by the discretization step size (s) using a hybrid spectral model, such as any of those described herein (e.g., mapping polychromatic projection values to monochromatic projection values based on two air scan data sets), thereby obtaining the corrected projection value (pm ) to determine (624), and may include.

[0054] Experimental results Experiments were conducted using the calibration method and the hybrid spectral model described herein. The results of this experiment are depicted in FIGS. 7A - 7C. The experiment was performed using a fan - beam imaging X - ray source and an array of X - ray detectors (each having dimensions of 1.024 mm (width)×2.213 mm (height), the source - to - detector distance being 1133.4 mm, and the source - to - isocenter distance being 643 mm), in a CT imaging system. The source and the detector array are located on opposite rotating gantries. The rotating gantry is configured to rotate at a maximum of 60 RPM. Also, the CT imaging system has a bowtie filter in the imaging beam path. This experimental CT imaging system is part of a system that combines imaging and radiation therapy, but similar experiments can be performed on a stand - alone imaging system having the same components as this CT imaging system.

[0055] During the calibration session, air scans were performed at 80 kVp and 140 kVp using the above - described CT imaging system. Using the air - scan data according to the method described herein, a hybrid spectral model with virtual filters was generated. That is, the virtual filters were calculated by calculating the LHR according to the method (300) of generating a hybrid spectral model represented by equations (2) and (3). In this experiment, look - up table mapping operators were generated for each X - ray detector according to the methods (500) and (520) depicted in FIGS. 5A - 5B.

[0056] During the imaging session, while rotating the CT imaging system at 30 RPM, the CATPHAN® 600 uniformity phantom module was scanned at 120 kVp and 150 mA tube current. CT projection data was calculated from the acquired X-ray intensity data. The CT images of FIGS. 7A - 7C were reconstructed using FDK with 3D backprojection. FIG. 7A depicts an image of the acquired CT projection data without correction for beam hardening artifacts, FIG. 7B depicts an image of the acquired CT projection data corrected for beam hardening artifacts using an ideal spectral model without a virtual filter (e.g., Equation (1)), and FIG. 7C shows an image of the acquired CT projection data corrected for beam hardening artifacts using a hybrid spectral model with a virtual filter (e.g., Equations (2) and (3)). In addition, the average CT values of the regions enclosed by the inner circle and the regions enclosed by the outer circle at the same location in all the images are shown in each of FIGS. 7A - 7C. The image of FIG. 7A has distinct ring artifacts (indicated by labeled arrows), a thin shadowed circular band or halo resulting from bowtie filter artifacts, and prominent cupping artifacts appearing as a dark disk-shaped thing surrounding the ring artifacts. The average CT value of the region enclosed by the inner circle (62 HU) is significantly different from the average CT value of the region enclosed by the outer circle (26 HU), with a difference of approximately 36 HU. The image of FIG. 7B also has similar ring artifacts, but the bowtie filter artifacts are reduced and the cupping artifacts are also reduced. The difference between the average CT value of the region enclosed by the inner circle (27 HU) and the average CT value of the region enclosed by the outer circle (16 HU) is 9 HU, which is reduced compared to the image of FIG. 7A. The cupping artifacts, bowtie filter artifacts, and ring artifacts present in FIG. 7A were somewhat reduced in FIG. 7B, but these artifacts were further reduced in FIG. 7C using the beam hardening correction method described herein. As seen in the image of FIG. 7C, the ring artifacts are significantly reduced or eliminated, and there are hardly any visible bowtie filter and cupping artifacts, if any.The difference between the average CT value (20 HU) of the region enclosed by the inner circle and the average CT value (19 HU) of the region enclosed by the outer circle is 1 HU, which is further reduced compared to the image of FIG. 7B. Further, the uniformity of the CT image was improved in FIG. 7C compared to FIG. 7A. Without beam hardening correction, the overall image of FIG. 7A had a CT value uniformity of 36 HU, while with beam hardening correction by the method described herein, the overall image of FIG. 7C had a CT value uniformity of 1 HU. This specification also provides, for example, the following items. (Item 1) A method for reducing beam hardening artifacts in CT imaging, the method comprising: acquiring CT projection data at each X-ray detector in a CT imaging system; determining corrected projection values for each of the acquired CT projection data using a mapping operator that includes a hybrid spectral model incorporating air scan CT projection data acquired at two different effective average energies; generating an artifact-corrected CT image by combining the corrected projection values for each of the acquired CT projection data. (Item 2) The hybrid spectral model represents the acquired CT projection data p p as a function of the corrected projection value p m and includes a virtual filter calculated based on the acquired air scan X-ray intensity data, the method according to Item 1. (Item 3) The mapping operator includes a look-up table LUT for each X-ray detector in the CT imaging system, each look-up table LUT having k corrected projection values p

Number

Number

Number

Number

Number

Number

Number

Number

Number

Number

Number

Number

Number

Number

Number

Number

Number

Number

Number

Claims

【Claim 1】 The invention described in this specification.

Images