Apparatus and method for adaptive fluence modulation for multi-source static computed tomography

Abstract

A multi-source static computed tomography (CT) imaging system includes a gantry having a bore and defining a radial plane extending transversely to a central axis, a plurality of source modules for directing X-ray beams toward a subject and a plurality of beam modulators positioned in front of the plurality of source modules. Each beam modulator can modify the X-ray beam from one or more corresponding source modules. The system further includes a plurality of detector modules arranged in diametric opposition to and on opposite sides of the radial plane of one of the plurality of source modules, and a control system configured to control the plurality of source modules to perform an acquisition sequence to acquire CT image data from the plurality of detector modules. The plurality of source modules, plurality of beam modulators and control system can be configured to perform adaptive fluence modulation.

Description

APPARATUS AND METHOD FOR ADAPTIVE FLUENCE MODULATION FOR MULTI-SOURCE STATIC COMPUTED TOMOGRAPHYCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application is based on, claims priority to, and incorporates herein by reference in its entirety U. S. Serial No. 63 / 714,833 filed October 31, 2024, and entitled “Structures Source filters to Modulate X-Ray Fluence Fields in Multi-Source Static Computed Tomography."STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[0002] This invention was made with government support under award number W81XWH-18-9-0004-11 (agreement 2023a002848) awarded by the U. S. Medical Research and Material Command and award number 1 AY2AX000060 awarded by the Advanced Research Projects Agency for Health. The government has certain rights in the invention.FIELD

[0003] The present disclosure relates generally to systems and methods for imaging a subject and, more particularly, to systems and methods for multi-source static computed tomography (CT).BACKGROUND

[0004] Computed Tomography (CT) is one of the most common diagnostic imaging modalities used in modern medicine, enabling rapid, non-invasive image acquisition at high resolution, plays a pivotal role in diagnostic medical imaging. However, conventional CT systems with a rotating gantry typically have a complex architecture and require heavy and sophisticated control hardware and electronics, which increases cost and limits availability in developing countries, or in resource-limiting environments, such as a battlefield. In addition, the temporal resolution attained using conventional CT systems is also limited by the time required for the gantry to mechanically complete a significant portion of an angular rotation. As a result, the reduced temporal resolution amplifies motion artifacts due to breathing or involuntary motion of apatient, affecting the spatial resolution achievable and degrading the quality of the reconstructed image.

[0005] To overcome such limitations, multi-source static (or motion-free or rotation-free) CT systems have been developed, for example, as described in U. S. Patent No. 10,492,744, issued December 3, 2019, which is incorporated herein by reference in its entirety. Multi-source static CT is an emerging technology that can use an array of x-ray sources to achieve projections for different view angles rather than a single source on a rotating gantry. In practice, this is made possible through the use of carbon nanotube x-ray sources which are inexpensive and compact. The advantages of multi-source static CT include light weight, robust mechanical design, lower duty cycle per source for heat dissipation, fast rotation speeds through electronic switching, and more. However, there can be challenges for image quality for static CT, in particular in a volume of interest, based on the acquired low dose (low source current) and sparse view data

[0006] There is a need for systems and methods for improving and controlling image quality and diagnostic accuracy for static CT imaging using, for example, adaptive fluence modulation for acquisition of CT data using a static CT imaging system.SUMMARY

[0007] In accordance with an embodiment, a multi-source static computed tomography (CT) imaging system including a gantry having a bore configured to receive a subject and defining a central axis extending along the bore and a radial plane extending transversely to the central axis extending along the bore, a plurality of source modules coupled to the gantry at fixed radial locations about the bore for directing X-ray beams toward the subject, a plurality of beam modulators coupled to the gantry at fixed radial locations about the bore and positioned in front of the plurality of source modulators, each beam modulator configured to shape the X-ray beam from one or more corresponding source modules, a plurality of detector modules coupled to the gantry at fixed radial locations about the bore such that one of the plurality of detector modules is arranged in diametric opposition to and on opposite sides of the radial plane of one of the plurality of source modules, and a control system coupled to the plurality of source modules and configured to control the plurality of source modules to perform an acquisition sequence to acquire CT image data from the plurality of detector modules.

[0008] In accordance with another embodiment, a method for generating a computed tomography (CT) image of a subject using fluence modulation in a multi-source static CT system having a plurality of beam modulators associated with a plurality of source modules includes acquiring, using the static CT system a first set of CT data from the subject using a low dose CT acquisition with uniform fluence for each of the plurality of source modules and with a predetermined amount of a total radiation dose, generating, using a processor device, a first CT image based on the first set of CT data, identifying, using the processor device, a target in the first CT image, generating, using the processor device, a fluence plan based at least on the identified target, acquiring, using the static CT system, a second set of CT data based on the fluence plan and with a remaining amount of the total radiation dose, and reconstructing, using the processor device, a second CT image based on the second set of CT data.BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The present disclosure will hereafter be described with reference to the accompanying drawings, wherein like reference numerals denote like elements.

[0010] FIGs. 1A and 1B illustrate an example motion-free (or static or rotation-free) computed tomography (CT) system having multiple X-ray sources in accordance with an embodiment;

[0011] FIG. 2A illustrates the frontal and side (cross-section) views of an example motion-free CT gantry design in accordance with an embodiment;

[0012] FIG. 2B is a schematic depicting a cross-sectional view of an example source-detector geometrical arrangement in accordance with an embodiment;

[0013] FIG. 3A is an illustration of an example modular design showing the insertion of source and detector modules into the gantry in accordance with an embodiment;

[0014] FIG. 3B is a detailed illustration of the example source and detector modules shown in FIG. 4B in accordance with an embodiment;

[0015] FIG. 4A illustrates an example motion-free CT gantry design depicting a subject illumination using a single source module in accordance with an embodiment;

[0016] FIG. 4B illustrates another example motion-free CT gantry design depicting a subject illumination using multiple source modules in accordance with an embodiment;

[0017] FIG. 5A is a schematic depicting a cross-sectional view of an example source-detector and beam modulator arrangement in accordance with an embodiment;

[0018] FIGs. 5B and 5C are schematic block diagrams of example source module and beam modulator arrangements in accordance with an embodiment;

[0019] FIG. 6 illustrates an example beam modulator arrangement and resulting fluence field in accordance with an embodiment;

[0020] FIG. 7A illustrates an example uniform fluence field generated using a static CT system without beam modulators;

[0021] FIGs. 7B-7E illustrate example fluence fields generated using a static CT system with beam modulators in accordance with an embodiment;

[0022] FIG. 8 illustrates a method for generating a CT image of a subject using adaptive fluence modulation in a multi-source static CT system in accordance with an embodiment;

[0023] FIG. 9 illustrates a method for generating a CT image of a subject using adaptive fluence modulation in a multi-source static CT system to minimize uncertainty in generative reconstruction in accordance with an embodiment; and

[0024] FIG. 10 is a block diagram of an example computer system in accordance with an embodiment.DETAILED DESCRIPTION

[0025] FIGs. 1A and 1B illustrates an example motion-free (or static or rotation-free) computed tomography (CT) system having multiple X-ray sources in accordance with an embodiment. As used herein, a “rotation-free,” “motion-free,” or “static” CT system refers to a system, such as described herein, where rotation of the gantry is not necessary to perform tomographic imaging. The example motion-free or static CT system 100 may be used to perform the methods described herein. The CT system 100 includes a gantry 102 that extends about a bore 103. The gantry 102 has multiple X-ray source modules 104 and detector modules 106 that are circumferentially arranged about the gantry 102, and coupled therein in a stationary fashion. That is, positions of the source modules 104 and detector modules 106 are at locations fixed in relation to a subject 112 arranged at a specific position along the gantry 102. The X-ray source modules 104 project an x-ray beam 108, which may be a fan-beam or cone-beam of X-rays, towards respective detector modules 106 on the opposite side of the gantry 102. In certain aspects, the source modules 104 are configured to operate at one or more X-ray energy levels.

[0026] The detector modules 106 may include multiple X-ray detector elements 110. For the sake of clarity, however, FIG. 1B only depicts a few X-ray source modules 104 and a fewcorresponding detector modules 106. Together, the X-ray detector elements 110 sense the projected X-rays 108 that pass through or are scattered by a subject 112, such as a medical patient or an object undergoing examination, that is positioned in the bore 103 of the CT system 100. Each X-ray detector element 110 produces an electrical signal that may represent the intensity of an impinging X-ray beam and, hence, the attenuation or scattered components of the beam as they passes through or scatter from the subject 112. In some configurations, each X-ray detector 110 may be capable of counting the number of X-ray photons that impinge upon the detector 110. That is, the detector 110 may include photon counting and / or energy discriminating detectors. However, the detector 110 may also include energy integrating detectors.

[0027] The CT system 100 also includes an operator workstation 116, which typically includes a display 118; one or more input devices 120, such as a keyboard and mouse; and a computer processor 122. The computer processor 122 may include a commercially available programmable machine running a commercially available operating system. In some embodiments, the operator workstation 116 may be a computer system 1000 as described below with respect to FIG. 10. The operator workstation 116 provides the operator interface that enables scanning control parameters to be entered into the CT system 100. The operator workstation 116 may be in communication with a data store server 124 and an image reconstruction system 126, or the functions of data storage and image reconstruction may be performed on the operator workstation 116. By way of example, the operator workstation 116, data store sever 124, and image reconstruction system 126 may be connected via a communication system 128, which may include any suitable network connection, whether wired, wireless, or a combination of both. As an example, the communication system 128 may include both proprietary or dedicated networks, as well as open networks, such as the internet.

[0028] The operator workstation 116 is also in communication with a control system 130 that controls operation of the CT system 100. The control system 130 generally includes an x-ray controller 132 and, optionally, may include a table controller 134, an optional gantry controller 136, and a data acquisition system (DAS) 138. In some embodiments, various aspects of the control system 130 may be implemented on one or more processor devices, for example, as described below with respect to FIG. 10. The X-ray controller 132 provides power and timing signals to respective X-ray source modules 104, in accordance with a desired scanning pattern or acquisition sequence. In some modes of operation, the X-ray controller 132 may be configured toconcurrently operate multiple source modules 104 to achieve a uniform scanning pattern impinging on a subject. In other modes of operation, each of the multiple source modules 104 may be operated in a time-varying fashion based on a selected illumination pattern. If included, the table controller 134 controls a table 140 to position the subject 112 in the gantry 102 of the CT system 100. In some situations, the table controller 134 may be mechanically controlled, such as by manually actuated levers or other controls. Furthermore, powered or manual control may be foregone. For example, in some situations, the table 140 may be a stretcher or make-shift table and the patient and table 140 adjusted as needed, such as may occur in battlefield or other deployments. In some alternative configurations, an optional gantry controller 136 may control the position of the gantry 102 with respect to the subject 112.

[0029] The DAS 138 samples data from the detector modules 106 and converts the data to digital signals for subsequent processing. For instance, digitized X-ray data is communicated from the DAS 138 to the data store server 124. The image reconstruction system 126 then retrieves the X-ray data (or CT image data) from the data store server 124 and reconstructs an image therefrom. The image reconstruction system 126 may include a commercially available computer processor, or may be a highly parallel computer architecture, such as a system that includes multiple-core processors and massively parallel, high-density computing devices. In some embodiments, the image reconstruction system 126 may be implemented on one or more processor devices, for example, as described below with respect to FIG. 10. Optionally, as mentioned, image reconstruction can also be performed on the processor 122 in the operator workstation 116. If not reconstructed at the operator workstation 116, reconstructed images can be communicated back to the data store server 124 for storage or to the operator workstation 116 to be displayed to the operator or clinician.

[0030] The CT system 100 may also include one or more networked workstations 142. By way of example, a networked workstation 142 may include a display 144; one or more input devices 146, such as a keyboard and mouse; and a processor 148. The networked workstation 142 may be located within the same facility as the operator workstation 116, or in a different facility, such as a different healthcare institution or clinic.

[0031] The networked workstation 142, whether within the same facility or in a different facility as the operator workstation 116, may gain remote access to the data store server 124 and / or the image reconstruction system 126 via the communication system 128. Accordingly, multiplenetworked workstations 142 may have access to the data store server 124 and / or image reconstruction system 126. In this manner, x-ray data, reconstructed images, or other data may be exchanged between the data store server 124, the image reconstruction system 126, and the networked workstations 142, such that the data or images may be remotely processed by a networked workstation 142. This data may be exchanged in any suitable format, such as in accordance with the transmission control protocol (TCP), the internet protocol (IP), or other known or suitable protocols.

[0032] Referring to FIGs. 2A and 2B, a schematic illustration of the gantry 102 of the present disclosure is provided. In contrast to previous designs, arrangement of multiple source modules 104 and detector modules 106 is achieved along separate rings of the gantry 102, by way of a first 200 and second 202 circumference configured in the gantry 102, respectively. As shown in FIG. 2B, the first 200 and second 202 circumference are axially separated by a lateral distance 204, and spaced from a radial plane 206 of the gantry 102 that runs perpendicular or transverse to a central axis 208 of the bore 103. As illustrated, the source modules 104 and detector modules 106 are angled toward the radial plane 206 such that each source module 104 is diametrically opposed to at least one detector module 106, to facilitate capture, among other X-ray beam components, attenuated beam components traversing an object, or subject. That is, each pair of source modules 104 and detector modules 106 includes one source module 104 and diametrically opposed detector module 106 arranged on opposite sides of the radial plane 206.

[0033] Additionally, a CT system in accordance with the present disclosure, may utilize multiple modules to enable a normal system operation even when some components are damaged. This feature can be useful in resource-limited environments such as a developing country or a battlefield. For this purpose, a gantry 102 of a system, as described, includes multiple source modules 104 and detector modules 106 that may be redundant and / or can be easily replaced without disintegrating the whole gantry 102. For instance, each source module 106 may be designed to enable easy replacement of damaged X-ray generating elements.

[0034] Particularly referring to FIG. 3A, a non-limiting example of the gantry is illustrated. The non-limiting example includes 9 source modules 104 and 9 detector modules 106. However, as may be appreciated, the number of modules may be varied depending on the size of system, targeted spatial resolution and accuracy. In the illustrated example, all modules are omitted, except ones at the bottom of the gantry 102, for clarity. As shown, the source modules 104 anddetector modules 106 can be inserted and replaced with new ones upon malfunction or damage in a manner similar to replacement of laser toner cartridges.

[0035] Referring now to FIG. 3B, a perspective view of a source module 104 and detector module 106, in accordance with an embodiment, is shown. Specifically, each source module 104 may include multiple X-ray source units, for example, 10, although various values may be possible, arranged in a linear array, although other configurations may also be possible. The source modules 104 may include two, or more, layers of shielding elements 300 configured for minimizing an X-ray contamination in proximate detector modules 106 or escaping from the gantry 102. In addition, each source module 104 may further include a source module controller (not shown) in communication with the X-ray controller 132 described above. In certain aspects, the source modules 104 may further configured to operate at more than one X-ray energy levels. In some embodiments, the source module 104 can be carbon nanotube (CNT) x-ray source array.

[0036] As illustrated, the detector module 106 may include a curved detector arranged under an anti-scatter grid 302, although other shape configurations may be possible, and may include multiple layers forming the anti-scatter grid 302. The anti-scatter grid 302 may be fashioned using a variety of suitable materials to block the X-rays scattered along a longitudinal, or Z-direction, while preserving at least some scattering along the other directions. As described, each source module 104 and detector module 106 may be configured along a slanted angle with respect to each other to guarantee that the trajectory of a transmitted X-ray beam from each source hits detector elements diametrically opposed to the subject. In addition, each detector module 106 may further include a detector module controller (not shown) in communication with a DAS 138, as described with reference to FIG. 1B. As mentioned, in some embodiments, the detector module 106 can be an energy integrating detector.

[0037] Referring to FIGs. 4A and 4B, aspects of a static CT system in accordance with an embodiment are shown. As mentioned, static CT system refers to a system, such as described herein, where rotation of the gantry is not necessary to perform tomographic imaging.Specifically, FIG. 4A shows the trajectory of an X-ray beam 402 emitted from a single source module 104. As the X-ray beam 402 is incident on the subject, which may include a patient or inanimate object, such as a package, some of the X-rays of the X-ray beam 402 pass through the subject as an attenuated X-ray beam 404, while other X-rays form scatter 406. By contrast, in conventional CT systems, only the attenuated X-ray beam 404 is measured, while the scatteredX-rays 406 are considered as sources of noise. However, the scattered x-rays contain information about the subject, and can be used to enhance the soft-tissue contrast and increase the spatial resolution. Thus, in accordance with an embodiment, the scattered X-rays 406 may be collected by the multiple detector modules 106 distributed about the circumference of the gantry 102.

[0038] In some embodiment, the above-described process can be performed sequentially or in parallel. By way of example, FIG. 4B illustrates one specific scanning pattern, whereby the multiple source modules 104 distributed along the circumference of the gantry 102 can be operated simultaneously in order to illuminate the subject uniformly. In other modes of operation, the source modules 104 can be operated in a time-varying fashion to achieve a selected illumination pattern. In this manner, multiplexed images can be decoded in a reconstruction step using the known information of system configuration including the illumination pattern. In some embodiments, compressed sensing techniques can then be used to reduce the number of images required for tomographic reconstruction by using a priori information of the subject.

[0039] As mentioned, the static (or motion-free) CT system 100 can interrogate a subject with X-rays from multiple angles, either simultaneously or in a time-varying fashion, and measure the X-ray attenuation and scattering. A selected illumination pattern can be uniform, i.e. all the sources are switched on, or varying with time, i.e., different subsets of the sources are switched on. In some embodiments, an exact number of required sources and their arrangement can be determined based on the imaging geometry, the categories of imaged subject, a required accuracy, and so forth. However, the acquired low dose and sparse view data can present challenges for image quality including for a targeted volume of interest (VOI) in the subject.

[0040] The present disclosure describes an apparatus, systems and methods for adaptive imaging with a multi-source static (or rotation free or motion-free) CT imaging system using adaptive fluence modulation. In some embodiments, a plurality of beam modulators can be positioned in front of a plurality of X-ray source modules in the gantry of the static CT imaging system. Each beam modulator can be configured to modify the X-ray beam from one or more of the plurality of source modules, for example, to shape, block, or adjust the spectrum of the X-ray beam. In some embodiments, a beam modulator can be, for example, a filter or a collimator. In some embodiments, a beam modulator can be positioned in front of each source module and. in other embodiments, a beam modulator can be positioned in front of a portion of a first source module and a portion of a secondsource module. The type of beam modulator provided for each source module can be the same or can be varied, for example, a first subset of source modules can be provided with one type of beam modulator and a second subset of source modules can be provided with a different type of beam modulator. In some embodiments, the beam modulators can be spotlight collimators that block positive or negative fan angles of even and odd indexed source modules, respectively. The beam modulators and an acquisition sequence or pattern (e.g. for the activation of the plurality of source modules) selected for a specific subject or volume of interest in the subject can be configured to modulate the fluence fields, for example, to deliver higher fluence to specific regions of the subject. In some embodiments, the disclosed apparatus, systems and methods can enable fluence field modulation with no moving parts (e.g., with static beam modulators that have no moving parts). In some embodiments, one or more beam modulators can be programmable or controllable beam modulators such as, for example, multi-leaf collimators. As discussed further below, advantageously, the beam modulators can enable volume of interest imaging by increasing the relative exposure for overlapping views.

[0041] The present disclosure also describes a method for adaptive fluence modulation using the static CT system with beam modulators. Advantageously, the method for adaptive fluence modulation can be configured to determine a fluence plan specific to, for example, a particular subject, a particular volume of interest in the subject, a particular application or goal for the CT imaging. The apparatus, systems and methods disclosed herein, can improve image quality and diagnostic accuracy of multi-source static CT imaging. In addition, the disclosed apparatus, systems and methods can be used for applications such as, for example, volume of interest imaging, CT material decomposition, and reducing or minimizing hallucinations in generative reconstructions. The disclosed apparatus, system and methods can be used to improve diagnostic quality of CT images in both military and civilian applications.

[0042] FIG. 5A is a schematic depicting a cross-sectional view of an example source-detector and beam modulator arrangement in accordance with an embodiment. In FIG. 5A, a beam modulator 150 can be positioned in front of each source module 104 proximate to the source module 104, and between the source module 104 and the corresponding and opposing detector module 106. Each beam modulator 150 can be configured to modify or change the X-ray beam (e.g., X-ray beam 108 shown in FIG. 1B or X-ray beam 402 shown in FIGs. 4A and 4B) emitted by the source module 104. For example, the beam modulator 150 can be configured to shape theprojected X-ray beam (e.g., shape the X-ray beam to the desired fluence field in the image reconstruction domain), block (or absorb) a portion of the X-ray beam, or adjust the spectrum (e g., the color) of all or a portion of the X-ray beam. Accordingly, in some embodiments, the beam modulator 150 can be configured to modify various aspects of the X-ray beam emitted from the source module 104, for example, 1) the beam modulator 150 can change the fluence coming out of the X-ray beam and modulate it spatially, or 2) the beam modulator 150 can change the spectrum coming out of the X-ray beam and modulate it spatially. In some embodiments, the beam modulator 150 for each source module 104 can have the same spatial and / or spectral pattern. In some embodiments, a beam modulator 150 with a different spatial or spectral pattern can be used for each source module 104. In some embodiments, different subsets of source modules 104 can each be provided with a beam modulator with the same spatial and / or spectral pattern. For example, a first set of source modules 104 can each have a beam modulator 150 with a first spatial and / or spectral pattern and a second set of source modules 104 can each have a beam modulator 150 with a second spatial and / or spectral pattern. In some embodiments, the type of beam modulator 150 used for each source module 104 can be varied for example, for multi-spectral acquisitions.

[0043] In some embodiments, the beam modulator 150 can be a filter, for example, an aluminum filter or a k-edge filter with a desired thickness. The filter can be configured to provide a desired spatial pattern, for example, a portion of the filter can be completely blocking and another portion of the filter can be more permissive and allow a portion of an x-ray beam to pass through. In some embodiments, the filter can be configured to provide a desired spectral pattern, for example, the filter can provide a desired spectra (e.g., a color) for the entire X-ray beam emitter from a source module 104 or different portions of the filter can provide or create different spectra (e.g., colors) for different portions of the of the X-ray beam emitted by the source module. In some embodiments, each filter (e.g., beam modulator 150) can have the same spatial and / or spectral pattern or, in other embodiments, different source modules 104 can have filters with different spatial and / or spectral patterns. In one example, filters that produce different colors could be placed in front of select ones of source modules 104 so that different subsets of the source modules are associated with different colors or, in another example, each source module 104 could be associated with a different color.

[0044] In some embodiment, the beam modulator 150 may be a collimator which can be formed from, for example, metal and can be configured to completely block a portion of an X-ray beam from a source module 104 and allow another portion of the X-ray beam from the source module 104 to pass through. As mentioned, each collimator (e.g., beam modulator 150) can have the same spatial pattern or, in other embodiments, different source modules 104 can have collimators with different spatial patterns. The collimators are static and do not need to move during an acquisition.

[0045] In some embodiments, beam modulators 150 (e.g., a filter, a collimator) can be static and do not need to move during an acquisition. In some embodiments, the beam modulator 150 can be controllable (e.g., programmable) and the spatial and / or spectral pattern can be changed during an acquisition. In some embodiments, the programmable beam modulator 150 can be a multi-leaf collimator (MLC). An MLC can comprise multiple leaves that move independently to form a beam shape and configured the fluence field. A MLC controller (e.g., as part of control system 130 of static CT system 100 shown in FIG. IB) can be used to control the movement of each of the leaves. The leaves of the MLC may be formed out of a radiation absorbing material (e.g., lead, tungsten, etc.). In some embodiments, a controllable beam modulator 150 can be provided for each source module 104 or for a subset of the source modules 104. Each controllable beam modulator 150 (e.g., an MLC) can be controlled to provide the same spatial pattern or, in other embodiments, different controllable beam modulators 150 can be controlled to provide different spatial patterns.

[0046] In some embodiments, the beam modulator 150 for each source module 104 can be the same type of beam modulator, for example, type of device, material used for the device, etc.. In some embodiments, a different type of beam modulator 150 can be used for each source module 104. In some embodiments, different subsets of source modules 104 can each be provided with a the same type of beam modulator 150. For example, a first set of source modules 104 can each have a static collimator and a second set of source modules 104 can each have a controllable multi -leaf collimator. In another example, a first set of source modules 104 can each have an aluminum filter and a second set of source modules 104 can each have a copper filter.

[0047] In some embodiments, a beam modulator 150 may be positioned in front of each source module 104 as illustrated in FIG. 5B. In some embodiments, beam modulator 150 may be integrated with a source module 104. As shown in FIG. 5C, in some embodiments,, one beammodulator 150 can be positioned in front of two source modules 104 so that a portion of the beam modulator 150 is in front of a portion of each of two source module 104 as shown in FIG.5C. Accordingly, the beam modulator 150 can modify a portion of the x-ray beam from each of the two source modules 104, which is discussed further below with respect to FIG. 6. For simplicity, FIGs. 5 A and 5B illustrate only two source modules 104 and two beam modulators 150 and FIG. 5C illustrates only two source modules 104 and one beam modulator 150.

[0048] In addition to providing a plurality of beam modulators 150 that can be used to shape the desired fluence field, the control system 130 (shown in FIG, IB) of the static CT system 100 can be used to dynamically control fluence fields and / or spectra for an acquisition. The control system 130 can be configured to selectively control the relative exposure applied by each source module 104 during an acquisition (or scan), for example, based on an imaging or fluence plan that can indicate which source modules 104 to activate, the timing or order for activation of each of the selected source modules 150, a spatial and / or spectral pattern of programmable beam modulators, and other acquisition parameters (e g., power, current, etc.). Advantageously, the source modules 104 and corresponding beam modulators 150 can be used to modulate, for example, the spatial distribution of the X-ray flux used for each projection and / or the spectrum of X-ray beam used for each projection. In some embodiments, the plurality of source modules 104 and the corresponding plurality of beam modulators 150 in a static CT system can be configured to build any fluence field in a tomographic slice. In addition, as discussed further below, the disclosed systems and methods can enable the acquisition of an optimal set of projections for a given radiation dose budget.

[0049] In some embodiments, the source modules 104 can be controlled to focus exposure on a particular region of the subject, for example, a region of interest (ROI) or volume of interest (VOI) of the subject. For example, the region of interest can be a particular side of the subject such as, for example, the right side, the left side, the anterior side, or the posterior side. For example, if a volume of interest that includes a target (e.g., a medical anomaly) is known to be on a particular side of the subject, the exposure can be focused on the relevant side to provide a higher amount of the total radiation dose for the acquisition to this side of the subject. In another example, the exposure from a plurality of the source modules 104 can be focused on the region or volume of interest from one or more sides of the subject. In such embodiments, the beam modulator 150 may be referred to as a spotlight beam modulator. For beam modulators 150 thatare configured to provide a spectral pattern, the control system 130 can be configured to activate different source modules 104 to use different colors to illuminate different portions or regions of the subject. Accordingly, the control system 130 can not only be used to dynamically control the fluence, but also to control the spectral separability for applications such as, for example, spectral CT material decomposition. In some embodiments, the control system 130 can be configured to utilize any type of activation pattern for the source modules 104, for example, a circular activation pattern or a non-circular activation pattern (e.g., a star pattern). In some embodiments, the fluence plan for an acquisition may utilize uniform fluence for each source module 104 or non-uniform fluence. Accordingly, an adaptive acquisition can be provided by, for example, 1) adjusting the power of each source module 104 (e.g., to adjust the exposure); 2) selecting which source modules 104 to activate (e.g., energize) and what projections (e.g., views) to obtain; and 3) providing the beam modulators 150 in front of each source module 104 to modulate the beam. The adaptive fluence modulation techniques described herein aim to allocate a higher proportion of the X-ray dose of the acquisition to the VOI, maximizing image quality where it is most needed for, for example, medical diagnosis, without increasing the total x-ray exposure.

[0050] As mentioned, a beam modulator 150 can be positioned, in some embodiments, in front of each source module 104 in the gantry 102, e.g., either fully in front of the source module 104 or partially in front of the source module 104. FIG. 6 illustrates an example beam modulator arrangement and resulting fluence field in accordance with an embodiment. In the example illustrated in FIG. 6, a beam modulator 150 can be positioned in front of and between two neighboring source modules 158, 160 and configured to block a portion of the X-ray beam from each of the two neighboring source modules 104. For example, in some embodiments, a single beam modulator 150 can be positioned in front of each even 158 and odd 160 indexed source module 104 and configured to block half of the x-ray beam of the even indexed source module 158 and half of the X-ray beam of the odd indexed source module 160 as illustrated in the exploded view 156. It should be understood that the half illumination approach described with respect to FIGs. 6 and 7A-7E is just one example of a beam modulator arrangement and spatial pattern. As described herein, in various embodiments, other arrangements, spatial patterns, spectral patterns, acquisition patterns, etc. can be provided using the disclosed adaptive fluence modulation systems and methods for static CT.

[0051] In the example illustrated in FIG. 6, as mentioned the beam modulator 150 can be a collimator configured to block the positive or negative fan angles of the even 158 and odd 160 indexed source modules, respectively. In FIG. 6, the even index source modules 158 can be collimated to pass positive fan angles only (i.e., blocks negative fan angles) and the odd source modules 160 can be collimated to pass negative fan angles only (i.e., blocks positive fan angles). For example, when the source module 158 is activated, the X-ray beam 162 (solid line) emitted by the source module 158 is directed to a first side (or half) of the subject and when the source module 160 is activated, the X-ray beam 164 (dotted line) emitted by the source module 160 is directed to a different second side (or half) of the subject. Accordingly, even and odd indexed source modules 158, 160 can illuminate different sides or halves of the subject. In an example, if a target volume 152 (e.g., a medical anomaly) is known to be on the right side of the subject, the X-ray exposure on the left side of the subject can be reduced and the X-ray exposure on the right side can be correspondingly increased by controlling which source modules 104 are activated and modulating the current of each source module 104. Accordingly, the source modules 104, beam modulators 150 and the control system 130 (shown in FIG. IB) can be used to selectively illuminate certain regions of the subject with higher effective exposure. An example resulting fluence field 134 for the described example half-illumination embodiment is also illustrated in FIG. 6. The example fluence field 134 in FIG. 6 can be referred to as a spotlight illumination (and fluence) pattern. By modulating the current of each source module 104, the spotlight fluence field 134 can be designed to have higher effective exposure for a volume of interest, for example, target volume 152. The spotlight fluence field 134 can improve image quality in the targeted half of the subject and reduces the image quality in the opposite half of the subject. Various acquisition patterns can be used for activating the even and odd source modules.

[0052] As mentioned, the example half illumination approach can be used to focus exposure on different sides of a subject. FIG. 7A illustrates an example uniform fluence field 702 generated using a static CT system without beam modulators. Each source module 104 can be activated with uniform fluence to generate the uniform fluence field 702. In addition, the uniform fluence field 702 can represent a certain total radiation that was applied in a uniform way. FIGs. 7B-7E illustrate example fluence fields generated using a static CT system with beam modulators in accordance with an embodiment. FIG. 7B illustrates a fluence field 704 focused on a right side (e g., a spotlight right fluence field) of a subject that can be generated with the half-illuminationapproach described above with respect to FIG. 6. The same total radiation exposure used for the uniform fluence field 702 can be used for the spotlight right fluence field 704, but with more exposure applied to the right side (or half) of the subject (e.g., three quarters of the radiation exposure) than the left side (or half) of the subject (e.g., one quarter of the radiation exposure). FIG. 7C illustrates a fluence field 706 focused on a left side of a subject (e.g., a spotlight left fluence field) that can be generated with the half-illumination approach described above with respect to FIG. 6. The same total radiation exposure used for the uniform fluence field 702 can be used for the spotlight left fluence field 706, but with more exposure applied to the left side (or half) of the subject than the right side (or half) of the subject. FIG. 7D illustrates a fluence field 708 focused on an anterior side of a subject (e.g., a spotlight anterior fluence field) that can be generated with the half-illumination approach described above with respect to FIG. 6. The same total radiation exposure used for the uniform fluence field 702 can be used for the spotlight anterior fluence field 708, but with more exposure applied to the anterior side (or half) of the subject than the posterior side (or half) of the subject. FIG. 7E illustrates a fluence field 710 focused on a posterior side of a subject (e.g., a spotlight posterior fluence field) that can be generated with the half-illumination approach described above with respect to FIG. 6. The same total radiation exposure used for the uniform fluence field 702 can be used for the spotlight posterior fluence field 710, but with more exposure applied to the posterior side (or half) of the subject than the anterior side (or half) of the subject.

[0053] FIG. 8 illustrates a method for generating a CT image using adaptive fluence modulation in a multi-source static CT system in accordance with an embodiment. Although the blocks of the process of FIG. 8 are illustrated in a particular order, in some embodiments, one or more blocks may be executed in a different order than illustrated in FIG. 8, or may be bypassed.

[0054] At block 802, a first set of CT data (or CT image data or projection data) can be acquired from the subject using a low dose CT scan (e.g., a CT scout scan) with a predetermined amount of the total radiation dose and unform fluence on all source modules (e.g., source modules 104) and to all parts of the subject, e.g., collecting all of the views). In one example, the predetermined amount of the total radiation dose for the low dose CT scan can be 10% of the total radiation dose. The low dose CT scan can be acquired using a multi-source static CT system with a plurality of beam modulators (e.g., beam modulators 150) for the X-ray source modules such as, for example, static CT system 100 described above with respect to FIGs. 1A-7E. Insome embodiments, the first set of CT data can be stored in data storage, for example, data store server 124 of the static CT system 100 shown in FIG. IB.

[0055] At block 804, a first CT image (e.g., a tomographic image) can be generated using the first set of CT data. The first CT image can be generated using, for example, known CT image reconstruction techniques such as model based iterative reconstruction (MBIR). In some embodiments, the first CT image can be generated (or reconstructed) using the image reconstruction system 126 or operator workstation 116 shown in FIG. 1 or a different computer system such as shown in FIG. 10. In some embodiments, the first CR image can be stored in data storage, for example, data store server 124 of the static CT system 100 shown in FIG. 1. At block 806, the first CT image can be analyzed to identify one or more targets, for example, a region or volume of interest, a material of interest in the subject, or a parameter or feature of the subject or the first CT image. In some embodiments, the target can be identified automatically using, for example, a classification or segmentation technique. In some embodiments, the classification or segmentation technique can be implemented on, for example, image reconstruction system 126 or operator workstation 112 shown in FIG. 1 or another computer system such as computer system 1000 shown in FIG. 10. The identified or detected target can be stored in data storage, for example, data store server 124 of the static CT system 100 shown in FIG. 1 or in data storage 1016 of a computer system 1000 shown in FIG. 10.

[0056] At block 808, a fluence plan for the subject can be generated based on at least the identified target(s) in the first CT image. In some embodiments, other factors such as, for example, the particular application or goal of the CT imaging (e g., volume of interest imaging, spectral CT material decomposition, minimizing hallucinations (or uncertainty) in generative reconstructions), the type of subject, characteristics of the static CT system, etc. can also be considered when generating the fluence plan. In one example, for volume of interest imaging, the fluence plan can be configured to apply higher fluence to the identified target. In another example, for spectral CT material decomposition, the fluence plan can be configured to control which regions of the subject are exposed to which colors of X-ray beam to enhance the image quality of specific materials, depending on the clinical task. The fluence plan can be generated for the remaining amount of the total radiation dose. For example, if the low dose CT scan is conducted with 10% of the total dose (or exposure), the remaining amount of the total radiation dose (or exposure) would be 90% of the total radiation dose. As mentioned, the fluence plan canalso include, for example, which source modules 104 to activate, the timing or order for activation of each of the selected source modules 104, a spatial and / or spectral pattern of programmable beam modulators, and other acquisition parameters (e.g., power, current, etc.). In some embodiments, the fluence plan can be stored in data storage, for example, data store server 124 of the static CT system 100 shown in FIG. 1 or in data storage 1016 of a computer system 1000 shown in FIG. 10.

[0057] At block 810, a second set of CT data (or CT image data or projection data) can be acquired from the subject using the remaining amount of the total radiation dose and based on the generated fluence plan. As mentioned, in one example the remaining amount of the total radiation dose can be 90% of the total radiation dose. As mentioned, in some embodiments, the fluence plan can be configured to use the remaining amount of the dose budget to target one or more of the identified targets from the first set of CT data. Accordingly, the acquisition can put more energy at a desired region or target. The second set of CT data can be acquired using a multi-source static CT system with a plurality of beam modulators (e.g., beam modulators 150) for the X-ray source modules such as, for example, static CT system 100 described above with respect to FIGs. 1A-7E. The second set of CT data can be stored in data storage, for example, data store server 124 of the static CT system 100 shown in FIG. 1.

[0058] At block 812, a CT image (e.g., a tomographic image) of the subject can be reconstructed based on the second set of CT data. The second CT image can be generated using, for example, a known CT image reconstruction technique such as model based iterative reconstruction (MBIR). In some embodiments, the second CT image can be generated (or reconstructed) using the image reconstruction system 126 or operator workstation 116 shown in FIG. 1 or a different computer system such as shown in FIG. 10. The second CT image can be stored in data storage, for example, data store server 124 of the static CT system 100 shown in FIG. 1 or in data storage 1016 of a computer system 1000 shown in FIG. 10.

[0059] As mentioned, the disclosed systems and methods can be used for minimizing uncertainty (or hallucinations) in generative reconstructions. FIG. 9 illustrates a method for generating a CT image using adaptive fluence modulation in a multi-source static CT system to minimize uncertainty in generative reconstruction in accordance with an embodiment. Although the blocks of the process of FIG. 9 are illustrated in a particular order, in some embodiments, one or more blocks may be executed in a different order than illustrated in FIG. 9, or may be bypassed.

[0060] At block 902, a first set of CT data (or CT image data or projection data) can be acquired from the subject using a low dose CT scan (e.g., a CT scout scan) with a predetermined amount of the total radiation dose and unform fluence on all source modules (e.g., source modules 104) and to all parts of the subject, e.g., collecting all of the views). In one example, the predetermined amount of the total radiation dose for the low dose CT scan can be 10% of the total radiation dose. The low dose CT scan can be acquired using a multi-source static CT system with a plurality of beam modulators (e.g., beam modulators 150) for the X-ray source modules such as, for example, static CT system 100 described above with respect to FIGs. 1A-7E. As discussed below, the first set of CT data can advantageously be used to determine where there is the most uncertainty (e.g., hallucinations) in a reconstruction performed using generative reconstruction. In some embodiments, the first set of CT data can be stored in data storage, for example, data store server 124 of the static CT system 100 shown in FIG. IB.

[0061] At block 904, a generative reconstruction algorithm, referred to herein as Langevin Posterior Sampling (LPS), can be performed on the first set of CT image data from the low dose CT scan of the subject. LPS can be used to generate many samples from the posterior to quantify the uncertainty of the generative model, commonly referred to as hallucinations. A generative reconstruction technique, such as LPS, can hallucinate when there is little information in the CT data measurements. That is, many possible anatomical structures could plausibly underlie any noisy measurements, and so the variance of the posterior random walk is high. LPS can utilize a physics-based measurement likelihood term (e.g., a physics-based likelihood model) and a probabilistic prior term from a pre-trained score-based diffusion model (e.g., a score-based diffusion prior). In some embodiments, LPS can use the score-based diffusion model and the physics-based likelihood model to sample a posterior random walk. LPS can be configured for Bayesian characterization of the posterior associated with the trained prior. One or more aspects of LPS can be implemented using one or more trained machine learning models that can be trained using known methods. In some embodiments, the LPS reconstruction can be performed using the image reconstruction system 126 or operator workstation 116 shown in FIG. 1 or a different computer system such as shown in FIG. 10.

[0062] LPS is configured to perform a random walk on the distribution p(x|y) where x are CT images and y are sinogram measurements. In some embodiments, LPS can be implemented using the following differential equation:where a is the step size, ^Xtlogptxt— xs~) is the prior score function which can be approximated with a trained neural network (e.g., trained using known methods), Vxlogp(y\x — xs) is the likelihood score function, and dwsis an amount of white noise. To derive the various components of equation (1), diffusion posterior sampling (DPS) can be implemented with the acquired first set of CT data using known methods. A variance-exploding process can be used as shown below:dxt= gdwt(2)and to sample from the posterior with DPS the reverse process can be used.whereXtlogpt(t) is a prior score function which can be approximated with a trained neural network (e.g., trained using known methods) andxlogp(y\x~) is the likelihood score function. For independent Poisson measurements, the score function can be given by:

[0063] As mentioned, LPS can be applied to the first set of CT data to generate a plurality of samples from the posterior to quantify the uncertainty of the generative model (i.e., hallucinations) in reconstructing an image from the first set of CT data. LPS can utilize an annealed Langevin sampling algorithm described.1

[0064] The form of equation (1) is similar to reverse diffusion sampling but with a factor of - in front of the score term. As a result, the distribution will not change over time but the samples will be a random walk on the posterior distribution. In the LPS stochastic differential equation (equation (6)) the coefficient y has been removed. This is because it is set to y=l for principled Bayesian inference. That is, the proper weights are being set on the prior and likelihood terms such that the random walk is truly over the posterior distribution. In many other methods using DPS, this coefficient may be set to less than one so the step size can be increased without numerical instability. The disclosed LPS generative reconstruction (equation (6) can achieve y = 1 by reducing the step size a and running many iterations. This stochastic process can result in a random walk on the posterior distribution defined by a diffusion-based prior score function and a physics model-based likelihood score function. This random walk can allow for Bayesianestimation by generating many samples from the posterior distribution of high-quality images given low-quality measurements. By taking the standard deviation across realizations of the random walk, the magnitude and position of hallucinations in the LPS generative reconstructions can be characterized.

[0065] At block 906, a hallucination map can be generated based on the characterized magnitude and position of the hallucinations. In some embodiments, the hallucination map can indicate the variance of the LPS samples. In some embodiments, the magnitude and position of hallucinations in the reconstruction of the first set of CT data and the generated hallucination map can be stored in data storage, can be stored in data storage, for example, data store server 124 of the static CT system 100 shown in FIG. 1 or in data storage 1016 of a computer system 1000 shown in FIG. 10.

[0066] At block 908, a fluence plan for the subject can be generated based on the hallucination map. For example, in some embodiments, the hallucination map can be forward projected to the projection domain to generate the fluence plan for the remaining amount of the total radiation dose. For example, if the low dose CT scan is conducted with 10% of the total dose (or exposure), the remaining amount of the total radiation dose (or exposure) would be 90% of the total radiation dose. In some embodiments, the fluence can be made proportional to the magnitude of the forward projected hallucination map for the fluence plan. In some embodiments, other factors such as, for example, the type of subject, characteristics of the static CT system, etc. can also be considered when generating the fluence plan. As mentioned, the fluence plan can also include which source modules 104 to activate, the timing or order for activation of each of the selected source modules 104, a spatial and / or spectral pattern of programmable beam modulators, and other acquisition parameters (e.g., power, current, etc.). In some embodiments, the fluence plan can be stored in data storage, for example, data store server 124 of the static CT system 100 shown in FIG. 1 or in data storage 1016 of a computer system 1000 shown in FIG. 10.

[0067] At block 910, a second set of CT data (or CT image data or projection data) can be acquired from the subject using the remaining amount of the total radiation dose and based on the fluence plan. As mentioned, in one example the remaining amount of the total radiation dose can be 90% of the total radiation dose. The fluence plan can be configured to use the remaining amount of the dose budget to target one or more of the identified regions of uncertainty from the first set of CT data (e.g., hallucinations) and to reduce that uncertainty. Accordingly, theacquisition can put more energy where it is expected to have more uncertainty in the reconstructed image. The second set of CT data can be acquired using a multi-source static CT system with a plurality of beam modulators (e.g., beam modulators 150) for the X-ray source modules such as, for example, static CT system 100 described above with respect to FIGs. 1 A-7E. The second set of CT data can be stored in data storage, for example, data store server 124 of the static CT system 100 shown in FIG. 1. At block 912, a CT image (e.g., a tomographic image) of the subject can be reconstructed based on the second set of CT data. In some embodiments, a generative reconstruction technique can be used such as, for example, the LPS technique described above. The generative reconstruction can be performed using the image reconstruction system 126 or operator workstation 116 shown in FIG. 1 or a different computer system such as shown in FIG. 10. The reconstructed CT image can be stored in data storage, for example, data store server 124 of the static CT system 100 shown in FIG. 1 or in data storage 1016 of a computer system 1000 shown in FIG. 10.

[0068] FIG. 10 is a block diagram of an example computer system in accordance with an embodiment. Computer system 1000 may be used to implement the systems and methods described herein. In some embodiments, the computer system 1000 may be a workstation, a notebook computer, a tablet device, a mobile device, a multimedia device, a network server, a mainframe, one or more controllers, one or more microcontrollers, or any other general-purpose or application-specific computing device. The computer system 1000 may operate autonomously or semi-autonomously, or may read executable software instructions from the memory or storage device 1016 or a computer-readable medium (e.g., a hard drive, a CD-ROM, flash memory, etc.), or may receive instructions via the input device 1020 from a user, or any other source logically connected to a computer or device, such as another networked computer or server. Thus, in some embodiments, the computer system 1000 can also include any suitable device for reading computer-readable storage media.

[0069] Data, such as data acquired with an imaging system (e.g., a computer tomography (CT) system) may be provided to the computer system 1000 from a data storage device 1016, and these data are received in a processing unit 1002. In some embodiment, the processing unit 1002 includes one or more processors. For example, the processing unit 1002 may include one or more of a digital signal processor (DSP) 1004, a microprocessor unit (MPU) 1006, and a graphics processing unit (GPU) 1008. The processing unit 1002 also includes a data acquisition unit 1010that is configured to electronically receive data to be processed. The DSP 1004, MPU 1006, GPU 1008, and data acquisition unit 1010 are all coupled to a communication bus 1012. The communication bus 1012 may be, for example, a group of wires, or a hardware used for switching data between the peripherals or between any components in the processing unit 1002.

[0070] The processing unit 1002 may also include a communication port 1014 in electronic communication with other devices, which may include a storage device 1016, a display 1018, and one or more input devices 1020. Examples of an input device 1020 include, but are not limited to, a keyboard, a mouse, and a touch screen through which a user can provide an input. The storage device 1016 may be configured to store data, which may include data such as, for example, CT data, CT images, adaptive acquisition sequences, fluence plans, etc. whether these data are provided to, or processed by, the processing unit 1002. The display 1018 may be used to display images and other information, such as magnetic resonance images, patient health data, and so on.

[0071] The processing unit 1002 can also be in electronic communication with a network 1022 to transmit and receive data and other information. The communication port 1014 can also be coupled to the processing unit 1002 through a switched central resource, for example the communication bus 1012. The processing unit can also include temporary storage 1024 and a display controller 1026. The temporary storage 1024 is configured to store temporary information. For example, the temporary storage 1024 can be a random access memory.

[0072] The following example sets forth, in detail, ways in which the present disclosure was developed and evaluated and ways in which the present disclosure may be used or implemented, and will enable one of ordinary skill in the art to more readily understand the principles thereof. The following example is presented by way of illustration and are not meant to be limiting in any way.

[0073] In this example study, simulation-based experiments show that spotlight collimators can be used to apply higher exposure to areas with more hallucinations, resulting in higher image quality. In this example study, computer simulations were used to investigate Volume of Interest (VOI) imaging, where the goal is to enhance image quality within a region with suspected medical anomalies

[0074] In the example study, a simulation-based experiment of head-CT imaging for stroke detection was conducted that demonstrated that spotlight beam modulators (e.g., a spotlightcollimator, for example, as described above with respect to FIGs. 6 and 7B-7E) can effectively reduce the standard deviation and worst-case scenario hallucinations in reconstructed images. Compared to uniform fluence (e.g., as illustrated in the example of FIG. 7A), the adaptive fluence modulation technique described herein can provide a significant reduction in posterior standard deviation. Accordingly, spotlight beam modulators and generative reconstruction can improve image quality and diagnostic accuracy of multi-source static CT.

[0075] In this example study, fluence field modulation with spotlight collimators in static CT was simulated. A nonlinear forward model can be used to simulate noise and measurements as shown below:y(x) = Io° exp —Ax (6)is a flattened vector representation of the ground truth image of attenuation coefficients assuming a 60keV monoenergetic source spectrum, A G ^MxN^> is a forward projection matrix, Iois the number of photons per line of response, and y(x) G ^Mxl^ is the expected photon counts. The ° symbol indicates elementwise multiplication and the exponential operation I =s elementwise. The measurements are considered to follow an independent Poisson distribution y\x ~lP(y(%)).

[0076] To model the spotlight collimators, A was modeled by deleting the matrix rows corresponding to lines of response that are blocked by the collimator. In terms of practical implementation, this can be accomplished by forward / back projecting even and odd views together and limiting the fan angle appropriately, followed by an operation to stitch the data together in a common projection domain.

[0077] For fluence modulation, the exposure of each source, as modeled by Io, can be increased and decreased where needed for the given imaging task. In this example study, volume of interest (VOI) imaging can be accomplished by forward-projecting the volume to get a sinogram domain mask, and then reducing current on the sources that are not overlapping the VOL After normalizing by total exposure, the effect will be higher fluence in the volume of interest relative to a uniform fluence field.

[0078] In this example study, the disclosed generative reconstruction algorithm LPS was used to generate many samples from the posterior to quantify the uncertainty of the generative model, i.e., hallucinations. As mentioned, LPS can utilize a physics-based measurement likelihood term(e., a physics-based likelihood model) and a probabilistic prior term from a pre-trained scorebased diffusion model (e g., a score-based diffusion prior). The score-based diffusion model and the physics-based likelihood model of LPS was used to sample a posterior random walk. By taking the standard deviation across realizations of the random walk, the magnitude and position of hallucinations in the generative reconstructions can be characterized.

[0079] In this example study, a simulation-based experiment was conducted to evaluate whether fluence modulation with spotlight collimators could be used to reduce hallucinations in generative reconstructions. A static CT system with a full ring of 72 sources with 400 mm source-to-axis distance and 400 mm source-to-detector distance using the LEAP cone beam forward projector was simulated. In this example study, a scenario was considered where 10% exposure is used for a low-dose CT (e.g., a CT scout scan) using uniform fluence on all sources (with spotlight collimators). Then, the LPS algorithm was run to sample a random walk on the posterior distribution of head CT images given the low7-dose scout measurements The regions of greatest uncertainty will be subject to the greatest variability. The example study focused on hallucinations in the brain tissue which can be critical for diagnosing many pathologies.Accordingly, a threshold-based bone mask was applied to the standard deviation map to create a hallucination map. Then, the hallucination map was forward projected to the projection domain to generate a fluence plan for the remaining 90% exposure. For this example study, the fluence was set proportional to the magnitude of the forward projected hallucination map. Then, the LPS algorithm was run again conditioned on the measurements from both the scout and final scans. For a comparison case, in this example study uniform fluence was also simulated with no spotlight collimators and normalized exposure to 106photons per pixel per view on average for both cases.

[0080] For training data, in this example study 10,800 axial head CT images with shape 256 x 256 were extracted from Task 1 of the SynthRAD2023 dataset. 10,000 images w7ere used for training and 800 for evaluation. There is some noise and blur in these images, but it is much higher image quality than what is possible with low-dose sparse view CT simulated in this example study, so it. can be treated as ground truth. For the diffusion backbone network, HuggingFace conditional U-Net was used. The prior score estimator was trained with 10,000 epochs with 1,000 iterations per epoch and 16 full 256 256 images. To initialize LPS, DPS was run with 256 evenly spaced time steps using a second order Heun sampler. Then, 1024 iterations of LPS was run, setting the step size dynamically such that the magnitude of the score term isalways 5 HU on average across all pixels. To evaluate the resulting image distributions, root mean squared error (RMSE), root mean squared bias (RMS bias), and standard deviation of the posterior samples were computed for both spotlight fluence and uniform fluence cases. The 90thpercentile standard deviation across all pixels of the standard deviation map were also evaluated to evaluate the worst-case- scenario hallucinations. For all of these metrics, in this example study a brain mask (0 to 80 HU) was used to evaluate only the errors in the brain region to avoid the bone regions dominating the error metrics. This process was repeated for 800 evaluations and report the mean and standard deviation of each error metric over the patient population.

[0081] The results of the example study demonstrate a significant reduction in hallucinations when using the disclosed spotlight fluence modulation compared to uniform fluence. The standard deviation across the posterior samples was lower for the spotlight fluence case, indicating reduced variability in the reconstructed images. Reductions in RMSE and RMS Bias were not found to be statistically significant in this study but the p-value for RMSE is close to significance at a p-value of 0.0547. The most dramatic improvement was over the 90th percentile standard deviation, a measure of worst-case scenario hallucinations, is also significantly reduced, highlighting the effectiveness of the spotlight fluence method,

[0082] In this example study, the disclosed fluence modulation in multi-source static CT systems using spotlight collimators simulated and evaluated. The disclosed LPS algorithm for Bayesian characterization of the posterior associated with a trained prior was also developed and evaluated. In this example study, the spotlight fluence method demonstrated a significant reduction in posterior standard deviation compared to uniform fluence. These findings indicate that adaptive fluence modulation can improve multi-source static CT image quality.

[0083] Computer-executable instructions for adaptive fluence modulation for a multi-source static CT system according to the above-described methods may be stored on a form of computer readable media. Computer readable media includes volatile and nonvolatile, removable, and nonremovable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer readable media includes, but is not limited to, random access memory (RAM), read-only memory (ROM), electrically erasable programmable ROM (EEPROM), flash memory or other memory technology, compact disk ROM (CD-ROM), digital volatile disks (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magneticstorage devices, or any other medium which can be used to store the desired instructions and which may be accessed by a system (e g., a computer), including by internet or other computer network form of access.

[0084] The present technology has been described in terms of one or more preferred embodiments, and it should be appreciated that many equivalents, alternatives, variations, and modifications, aside from those expressly stated, are possible and within the scope of the invention.- 21 -

Claims

CLAIMS1. A multi-source static computed tomography (CT) imaging system, the system comprising:a gantry having a bore configured to receive a subject and defining a central axis extending along the bore and a radial plane extending transversely to the central axis extending along the bore;a plurality of source modules coupled to the gantry at fixed radial locations about the bore for directing X-ray beams toward the subject;a plurality of beam modulators coupled to the gantry at fixed radial locations about the bore and positioned in front of the plurality of source modules, each beam modulator configured to shape the X-ray beam from one or more corresponding source modules;a plurality of detector modules coupled to the gantry at fixed radial locations about the bore such that one of the plurality of detector modules is arranged in diametric opposition to and on opposite sides of the radial plane of one of the plurality of source modules; anda control system coupled to the plurality of source modules and configured to control the plurality of source modules to perform an acquisition sequence to acquire CT image data from the plurality of detector modules.

2. The system according to claim 1, wherein the control system is further configured to reconstruct one or more tomographic images from the acquired CT image data.

3. The system according to claim 1, wherein at least one of the plurality of source modules and a corresponding beam modulator are configured to modulate a spatial distribution of an X-ray flux of the X-ray beam emitted by the at least one source module.

4. The system according to claim 1, wherein at least one of the plurality of source modules and a corresponding beam modulator are configured to modulate a spectrum of the X-ray beam emitted by the at least one source module.

5. The system according to claim 1, wherein each of the plurality of beam modulators is one of a filter, a collimator, or a multi-leaf collimator (MLC).

6. The system according to claim 1, wherein each of the plurality of beam modulators is a same type of beam modulator.

7. The system according to claim 1, wherein one or more of the plurality of beam modulators are different types of beam modulator.

8. The system according to claim 1, wherein the acquisition sequence is configured to provide a first amount of radiation dose to a first region of the subject and a second amount of radiation dose to a second region of the subject.

9. The system according to claim 8, wherein the first region of the subject includes a volume of interest and the first amount of radiation dose is greater than the second amount of radiation dose.

10. The system according to claim 8, wherein the second region of the subject is opposite to the first region of the subject.

11. The system according to claim 8, wherein each beam modulator in the plurality of beam modulator is positioned in front of two of the plurality of modules and is configured as a spotlight beam modulator.

12. The system according to claim 1, wherein each of the plurality of beam modulators is associated with one or more of the plurality of source modules.

13. The system according to claim 1, wherein the acquisition sequence is configured to include a non-circular activation pattern for the plurality of source modules.

14. The system according to claim 1, wherein the activation sequence is configured to adaptively control an amount of energy applied to each source module in the plurality of source modules.

15. A method for generating a computed tomography (CT) image of a subject using fluence modulation in a multi-source static CT system having a plurality of beam modulators associated with a plurality of source modules, the method comprising:acquiring, using the static CT system, a first set of CT data from the subject using a low dose CT acquisition with uniform fluence for each of the plurality of source modules and with a predetermined amount of a total radiation dose;generating, using a processor device, a first CT image based on the first set of CT data; identifying, using the processor device, a target in the first CT image;generating, using the processor device, a fluence plan based at least on the identified target;acquiring, using the static CT system, a second set of CT data based on the fluence plan and with a remaining amount of the total radiation dose; andreconstructing, using the processor device, a second CT image based on the second set of CT data.

16. The method according to claim 15, wherein the fluence plan is configured to provide a first amount of the remaining amount of the total radiation dose to a first region of the subject and a second amount of the remaining amount of the total radiation dose to a second region of the subject.

17. The method according to claim 16, wherein the first region of the subject includes the target and the first amount of the remaining amount of the total radiation dose is greater than second amount of the remaining amount of the total radiation dose.

18. The method according to claim 15, wherein the fluence plan comprise a circular activation pattern for the plurality of source modules.

19. The method according to claim 15, wherein the fluence plan is configured to apply a noncircular activation pattern for the plurality of source modules.

20. The method according to claim 15, wherein the fluence plan is configured to provide X-ray beams of a first spectrum to a first region of the subject and X-ray beams of a second spectrum to a second region of the subject.

Images