Estimation of B0 inhomogeneity for improved magnetic resonance image acquisition and / or reconstruction

Abstract

This document discloses a medical system (100, 300, 500) comprising a memory (110) storing machine-executable instructions (120) and a B0 field estimation module (126); and a computing system (106). Execution of the machine-executable instructions causes the computing system to receive (200) an initial magnetic resonance image (122), the initial magnetic resonance image (122) including amplitude components and describing a first region of interest (326) of an object (118). Execution of the machine-executable instructions also causes the computing system to perform at least one of the following iterations: receiving (202) subsequent k-space data (124) describing a subsequent region of interest (328) of the object; calculating (204) an estimated B0 field mapping (128) for the subsequent region of interest based on the initial magnetic resonance image by inputting the initial magnetic resonance image into the B0 field estimation module; and reconstructing (206) a corrected magnetic resonance image (130) based on the subsequent k-space data and the estimated B0 field mapping.

Description

Technical Field

[0001] This invention relates to magnetic resonance imaging, and more specifically, to the estimation of B0 inhomogeneities for improved image acquisition and / or reconstruction. Background Technology

[0002] Magnetic resonance imaging (MRI) scanners use large static magnetic fields to align the nuclear spins of atoms as part of the process used to generate images of a patient's body. This large static magnetic field is called the B0 magnetic field, or the master magnetic field, or simply the B0 field. Producing high-quality MRI images depends on correcting for B0 inhomogeneities in the B0 field, such as using shimming, or compensating for B0 inhomogeneities during reconstruction.

[0003] Shi et al. disclosed a template-based method in “Template-Based Field Map Prediction for Rapid Whole Brain B0 Shimming” (Magnetic Resonance in Medicine 80:171-180 (2018)) to determine B0 inhomogeneities in the B0 field due to the presence of a human head. US Patent Application US 2006 / 0220645 discloses a B0 map specifically generated from actual acquisition (SSFP) with phase increment redundancy. Summary of the Invention

[0004] The method is provided in the independent claims as a medical system, computer program, and method. Embodiments are given in the dependent claims. For example, the medical system is implemented as a magnetic resonance imaging system or an MRI scanner. The patient to be examined may be located in the examination area of ​​the MRI scanner.

[0005] As mentioned above, determining the inhomogeneities in the B0 field is important for properly shimming the master magnet and reconstructing accurate magnetic resonance images. Typically, the B0 field is measured by performing magnetic resonance (MR) measurements at several different pulse times and recording the phase changes. This provides accurate results but is time-consuming. Another challenge is that the imaged object may move after the B0 image is acquired.

[0006] Embodiments may provide an improved apparatus for determining the B0 field mapping by calculating an estimated B0 field mapping using an initial magnetic resonance image to reconstruct a corrected magnetic resonance image from subsequently acquired k-space data. The initial magnetic resonance image can, for example, be a so-called reconnaissance image or survey image. The reconnaissance image or survey image may have a first region of interest capable of imaging a large portion of the object's body or at least a portion of the object's body within the imaging region. This can have several advantages. First, survey images are typically acquired to align with subsequent imaging procedures. It does not slow down the procedure. Another advantage is that, because the reconnaissance image or survey image has a larger field of view than subsequent images, it takes into account a large portion of the object when determining the degree of distortion of the object's body to the B0 field. Detailed magnetic resonance images imaging volumes within the object's body may not provide an accurate estimate of the B0 field because portions of the object outside the subsequent region of interest distort the B0 field and may not be taken into account. The insight of this invention is that the main contribution to the variability of B0 inhomogeneity stems from the different anatomical shapes and / or patient positioning within the examination area of ​​the medical system. Both of these will be captured by the survey image. After the first scan (typically the initial survey scan), the acquired diagnostic images can be different. As long as there is significant overlap in the regions of interest of the measurement scan and the diagnostic image, a B0-map for subsequent images can be predicted based on the first amplitude image (survey / coil survey / sensor reference, etc.). Specifically, this is to address motion and breathing issues of the subject. This invention is applicable to any amplitude image. It can even be used with a single SSFP scan. However, we want to use surveys and pre-scans, which have much shorter acquisition times and are required in any session. Furthermore, the B0-map generated according to this invention can be widely used for shimming (static and dynamic) and reconstruction of EPI / spirals, even in the presence of motion. This invention uses “shape information” in the amplitude image to predict the B0 distribution and does not require pixel-by-pixel analysis. This invention is able to predict the B0 distribution based on completely unrelated scans, thus independent of the acquisition being performed. This invention uses structural information captured in the amplitude image and is not applicable to pixel-by-pixel fitting known itself from US Patent Application US 2006 / 0220645. That is, this invention utilizes cross-voxel information for prediction. Even in the presence of motion or respiration within the sequence, this invention is able to predict the B0 distribution. Data acquired after the first scan (subsequent images) can involve new positions or changes in object shape, such as a rotated head or an inflated chest. Changes in object position or shape lead to changes in the B0 distribution. The updated B0 shimming settings derived from the calculated B0 distribution involve updates required due to these physiological changes.

[0007] In one aspect, the present invention provides a medical system comprising a memory storing machine-executable instructions and a computing system configured to control the medical system. In different examples, the medical system can take different forms. In one example, the medical system is a remote or cloud server for performing reconstruction of medical or magnetic resonance images. In another example, the computing system is a workstation used by a radiologist or other medical professional to review or process medical image data such as magnetic resonance images. In yet another example, the computing system is part of a medical system that includes a magnetic resonance imaging system. For example, the computing system can be part of a control system or a control system for controlling the magnetic resonance imaging system.

[0008] The execution of machine-executable instructions causes the computing system to receive an initial magnetic resonance image of a first region of interest describing the object. The initial magnetic resonance image includes an amplitude component. In some cases, the initial magnetic resonance image, as described below, is subsequently used to compute an estimated B0 field map.

[0009] The execution of the machine-executable instructions also causes the computing system to perform the estimation of the B0 field map. That is, to receive subsequent k-space data for different regions of interest describing the object. The subsequent regions of interest at least partially overlap with the first region of interest. The execution of the machine-executable instructions also causes the computing system to perform at least one iteration of calculating the estimated B0 field map for the subsequent regions of interest using the B0 field estimation module based on the initial magnetic resonance image. The execution of the machine-executable instructions also causes the computing system to perform at least one iteration of reconstructing the corrected magnetic resonance image based on the subsequent k-space data and the estimated B0 field map.

[0010] This embodiment can be advantageous because previous MRI images (in this case, the initial MRI image) can be used to generate an estimated B0 field map, which can then be used to reconstruct the corrected MRI image. This can reduce or eliminate the need for a preliminary B0 field map when performing an MRI examination.

[0011] One instance where this can be particularly advantageous is when the initial MRI image is a so-called survey or reconnaissance image. Survey or reconnaissance images are preliminary images, typically with lower resolution, used to confirm the location of specific anatomical regions of the object. The use of so-called survey or reconnaissance images can be beneficial because it allows imaging of large areas of the object. The distribution of the object's body within the B0 magnetic field causes distortion of the B0 magnetic field.

[0012] The determination of the estimated B0 field mapping can be performed in various different ways. For example, the method may also include motion correction. For instance, a sensor or camera can be used to measure the position of an object over time, and this data can be used to update the estimated B0 field mapping based on an initial magnetic resonance image. In other cases, subsequent k-space data can be used to construct a preliminary magnetic resonance image without motion or B0 correction. This can then be registered to the initial magnetic resonance image and used to determine motion parameters, and then a more accurate estimated B0 field mapping can be calculated.

[0013] The term "estimated B0 field mapping" refers to data describing the distortion of the B0 field caused by an object in the master magnet. Therefore, the term "estimated B0 field mapping" can be interpreted in several different ways. B0 residual maps can, for example, be used to adjust the magnetic shimming of a magnetic resonance imaging system to compensate for B0 inhomogeneities caused by the object. In another instance, the estimated B0 field mapping can refer to the residual B0 mapping (the magnetic inhomogeneities remaining after shimming). In this case, the estimated B0 field can be used to correct for the effects of B0 magnetic field inhomogeneities during reconstruction in high-sensitivity techniques such as echo-plane imaging and imaging utilizing non-Cartesian k-space trajectories like radial, helical, etc.

[0014] In another embodiment, the medical system further includes a magnetic resonance imaging system configured to acquire k-space data from an imaging region. The memory also contains a first pulse sequence command configured to acquire initial k-space data from a first region of interest. The memory also contains a set of second pulse sequence commands, each configured to acquire subsequent k-space data from a subsequent region of interest. Execution of the machine-executable instructions further enables the computing system to control the magnetic resonance imaging system to acquire the initial k-space data using the first pulse sequence commands.

[0015] The execution of the machine-executable instructions also enables the computational system to reconstruct the initial magnetic resonance image based on the initial k-space data. Furthermore, the execution of the machine-executable instructions enables the computational system to control the magnetic resonance imaging system to acquire subsequent k-space data for each iteration using one of a set of second pulse sequence commands. This embodiment can be advantageous because it provides a magnetic resonance imaging system capable of providing B0 inhomogeneity reconstruction for multiple images derived from a single initial magnetic resonance image.

[0016] In another embodiment, the initial magnetic resonance image is an amplitude image. It should be noted that the initial magnetic resonance image is an amplitude image, not a phase image.

[0017] In another embodiment, the subsequent region of interest is within the first region of interest. The volume of the subsequent region of interest is less than or equal to the volume of the first region of interest. Using images within a larger field of view or region of interest allows for a more accurate determination of the obtained estimated B0 field mapping. For example, if the initial MRI image or survey or reconnaissance image is an image of the entire head of an object, the distortion in the B0 or ​​main magnetic field caused by the head can be accurately calculated. If subsequent k-space data only image a small region of the head (e.g., part of the brain), the B0 field mapping from that subsequent region of interest may not provide an accurate value for the estimated B0 field mapping.

[0018] In another embodiment, the volume of the subsequent region of interest (ROI) is larger than the volume of the first ROI. In some cases, the initial MRI image can be used to determine the estimated B0 field mapping for regions larger than the first ROI. For example, template-based methods can match the initial MRI image to a larger region to compute the estimated B0 field mapping for the subsequent ROI. This also applies to AI-based methods. Some MRI protocols (such as EPI) can produce severely distorted images. The type of initial MRI image can be selected to minimize distortion. For example, a conventional gradient echo image can be acquired as the initial MRI image. Although the first ROI is smaller than the subsequent ROI, its lack of distortion allows for a more accurate estimated B0 field mapping.

[0019] In another embodiment, the magnetic resonance imaging system includes a main magnet for generating a B0 magnetic field in the imaging region. The system also includes an adjustable B0 magnetic field shimming configuration configured to shim the B0 magnetic field in the imaging region. Execution of machine-executable instructions further enables the computational system to perform the following operations before each subsequent acquisition of k-space data: Specifically, it computes an updated B0 shimming setting configured to reduce B0 inhomogeneities using the estimated B0 field mapping. This can be beneficial because correcting the shimming setting of the B0 magnetic field increases the signal-to-noise ratio in the k-space data. Once reconstruction is performed, this provides a better quality magnetic resonance image. This can have the effect of enabling a dynamic shimming process to improve the shimming of the B0 magnetic field after each acquisition.

[0020] In another embodiment, the execution of machine-executable instructions also enables the computing system to directly calculate an initial B0 field mapping for subsequent regions of interest based on the initial magnetic resonance image. This can, for example, be used to calculate the B0 map before the first acquisition of subsequent k-space data. This can, for example, be used to adjust the magnetic homogenization of the master magnet.

[0021] In another embodiment, the estimated B0 field mapping is calculated at least in part using the updated B0 shimming setting. Adjusting the B0 magnetic field shimming will change the uniformity of the B0 magnetic field. Therefore, it is advantageous to update the B0 field mapping using the updated 0 shimming setting.

[0022] In another embodiment, the execution of the machine-executable instructions further causes the computing system to perform the following steps during subsequent k-space data acquisition. One step performed during subsequent k-space data acquisition is receiving motion parameters describing the motion of the object during or after the subsequent k-space data acquisition. Another step performed during subsequent k-space data acquisition is calculating an estimated B0 field using at least part of the motion parameters. The estimated B0 field can then be used during image reconstruction.

[0023] In another embodiment, the medical system also includes a motion sensor system configured to at least partially measure motion parameters. Execution of machine-executable instructions further enables the computing system to control the motion sensor system to measure motion parameters. For example, a breathing belt or camera system can be used to directly measure the motion of an object. In other examples, reference markers may be present, which can be used to optically and precisely measure the position of the object, or, if MRI reference markers are present, can be used to measure the position of the object within the magnetic resonance image itself.

[0024] In another embodiment, execution of the machine-executable instructions also causes the computing system to reconstruct the intermediate image from subsequent k-space data. Execution of the machine-executable instructions also causes the computing system to calculate the registration between the intermediate image and the initial magnetic resonance image.

[0025] The execution of machine-executable instructions also enables the computing system to calculate motion parameters from the registration. This embodiment can be advantageous because no external system or motion sensor is required to determine the motion parameters. It also allows for very accurate determination of the position of subsequent regions of interest relative to the first region of interest. This can provide more accurate estimates of the B0 field calculation.

[0026] In another embodiment, at least in part, an analytical model is used to compute the estimated B0 field map, so as to compute the spatial transformation of the estimated B0 field map using motion parameters. For example, if the B0 field map is computed directly from the initial magnetic resonance image, it is possible to use the analytical model to transform the position of the B0 field after the object has been moved.

[0027] In another embodiment, the estimated B0 field is computed at least in part by inputting an initial magnetic resonance image and / or an estimated B0 field map and motion parameters into a trained neural network. In this example, the trained neural network can be used to predict or estimate the B0 field map and / or the residual B0 field map using the motion parameters.

[0028] In another embodiment, the estimated B0 field is computed at least in part by inputting the estimated B0 field mapping and motion parameters into a trained artificial intelligence algorithm. Similarly, in this embodiment, a trained artificial intelligence algorithm can be used to improve the value of the estimated B0 field mapping.

[0029] In another embodiment, the estimated B0 field is computed at least in part by feeding motion parameters into a trained support vector machine. The support vector machine can be trained or programmed to perform this transformation of motion parameters into the estimated B0 field graph.

[0030] In another embodiment, the memory also contains a system model configured to output time-correlation data describing the electromagnetic properties of the magnetic resonance imaging system in response to input of one of the second pulse sequence commands. Execution of the machine-executable instructions further enables the computational system to compute the time-correlation data by inputting subsequent pulse sequence commands into the system model. The calibrated magnetic resonance image is reconstructed based on subsequent k-space data, the estimated B0 field mapping, and the time-correlation data.

[0031] Various parameters that affect the quality of magnetic resonance imaging (MRI) reconstructions can be pre-calculated or modeled. Time- and temperature-dependent data are electromagnetic data derived from these quantities. Time / temperature-dependent data can describe any of the following: B0 field variations, eddy currents throughout the MRI system, amplitude and inhomogeneity of the transmitted RF field, gradient magnetic field amplitude and nonlinearity, receiver coil sensitivity variations, motion-dependent B0 inhomogeneities, associated magnetic field corrections, and combinations thereof.

[0032] In another embodiment, one of the second pulse sequence commands is configured to acquire subsequent k-space data according to the echo-plane imaging magnetic resonance imaging protocol.

[0033] In another embodiment, the second pulse sequence command is configured to acquire subsequent k-space data according to a multi-band magnetic resonance imaging protocol. Both echo-plane imaging and multi-band magnetic resonance imaging can benefit from shimming and better estimation of residual B0 inhomogeneities.

[0034] In another embodiment, the second pulse sequence command is configured to acquire subsequent k-space data in a spiral k-space sampling mode.

[0035] In another embodiment, the second pulse sequence command is configured to acquire subsequent k-space data in a non-Cartesian sampling mode.

[0036] However, it should be noted that the implementation can be beneficial for all k-space sampling modes or trajectories.

[0037] In another embodiment, the B0 field estimation module is implemented as a B0 modeling neural network. The B0 modeling neural network can be, for example, a neural network such as a convolutional neural network, which has been trained to generate an estimate of the B0 field in response to receiving one or more magnetic resonance images. The B0 modeling neural network can be trained, for example, by measuring the magnetic resonance images and then, before or after performing another magnetic resonance image, to measure the B0 image. Therefore, the training of the B0 modeling neural network would be a direct backpropagation or deep learning algorithm.

[0038] In another embodiment, the B0 field estimation module is implemented as a machine learning system.

[0039] In another embodiment, the B0 field estimation module is implemented as a B0-modeled random forest regression system.

[0040] In another embodiment, the B0 field estimation module is implemented as a B0 support vector machine learning system.

[0041] In another embodiment, the B0 field estimation module is implemented as a template-based B0 magnetic field predictor system. Template-based systems have been shown to accurately generate B0 magnetic field maps.

[0042] In another aspect, the present invention provides a computer program comprising machine-executable instructions executed by a computing system for controlling a medical system. Execution of the machine-executable instructions causes the computing system to receive an initial magnetic resonance image of a first region of interest describing the object. The initial magnetic resonance image includes an amplitude component.

[0043] The execution of machine-executable instructions also causes the computing system to perform at least one of the following iterations: First, it receives subsequent k-space data for subsequent regions of interest describing the object. These subsequent regions of interest at least partially overlap with the first region of interest.

[0044] The execution of the machine-executable instructions also causes the computational system to iteratively compute the estimated B0 field mapping for subsequent regions of interest based on the initial magnetic resonance image. The execution of the machine-executable instructions also causes the computational system to perform at least one iteration of reconstructing the calibrated magnetic resonance image based on subsequent k-space data and the estimated B0 field mapping.

[0045] In another aspect, the present invention provides a medical imaging method. The method includes receiving an initial magnetic resonance image of a first region of interest describing a subject. The initial magnetic resonance image includes an amplitude component. Execution of machine-executable instructions further causes the computing system to perform at least one iteration: the first step of the iteration is to receive subsequent k-space data of subsequent regions of interest describing the subject.

[0046] The subsequent region of interest (ROI) at least partially overlaps with the first ROI. The next step in the iteration is to compute an estimated B0 field mapping for the subsequent ROI using the B0 field estimation module based on the initial MRI image. The next step in the iteration is to reconstruct the corrected MRI image based on subsequent k-space data and the estimated B0 field mapping.

[0047] The magnetic resonance imaging system of the present invention is configured to reconstruct a set of magnetic resonance images based on echo signals, wherein the reconstruction software is installed in the computing system of the magnetic resonance examination system, or wherein the computing system has an interface to a remote reconstruction facility. The reconstruction software can be installed, for example, on a remote server in a healthcare facility with data network access, wherein the reconstruction software is available in the "cloud". In these remote configurations, the computing system is equipped with the capability to schedule the reconstruction of a set of magnetic resonance images at a remotely located reconstruction facility. Furthermore, the reconstruction of the magnetic resonance images can be accomplished through machine learning, for example, through a trained neural network, which can be incorporated into the computing system or accessed from a remote location.

[0048] It should be understood that one or more of the above embodiments of the present invention can be combined, as long as the combined embodiments are not mutually exclusive.

[0049] As will be understood by those skilled in the art, aspects of the present invention can be embodied as apparatus, method, or computer program product. Therefore, aspects of the present invention can take the form of entirely hardware embodiments, entirely software embodiments (including firmware, resident software, microcode, etc.), or embodiments combining software and hardware aspects, which are generally referred to herein as “circuit,” “module,” or “system.” Furthermore, aspects of the present invention can take the form of a computer program product embodied in one or more computer-readable media having computer-executable code embodied thereon.

[0050] Any combination of one or more computer-readable media can be used. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. As used herein, "computer-readable storage medium" includes any tangible storage medium that can store instructions executable by a processor or computing system of a computing device. A computer-readable storage medium can be referred to as a computer-readable non-transitory storage medium. A computer-readable storage medium can also be referred to as a tangible computer-readable medium. In some embodiments, a computer-readable storage medium may also store data accessible by a computing system of a computing device. Examples of computer-readable storage media include, but are not limited to: floppy disks, magnetic hard disk drives, solid-state drives, flash memory, USB thumb drives, random access memory (RAM), read-only memory (ROM), optical disks, magneto-optical disks, and register files of computing systems. Examples of optical disks include compact discs (CDs) and digital universal discs (DVDs), such as CD-ROMs, CD-RWs, CD-Rs, DVD-ROMs, DVD-RWs, or DVD-R discs. The term computer-readable storage medium also refers to various types of recording media accessible by a computer device via a network or communication link. For example, data can be retrieved via a modem, via the Internet, or via a local area network. Computer-executable code embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, fiber optic cable, RF, or any suitable combination thereof.

[0051] Computer-readable signal media may include propagated data signals carrying computer-executable code embodied therein, for example, in baseband or as part of a carrier wave. Such propagated signals may take any of a variety of forms, including but not limited to electromagnetic, optical, or any suitable combination thereof. A computer-readable signal medium may be any computer-readable medium that is not a computer-readable storage medium but is capable of communicating, propagating, or transmitting a program used by or in conjunction with an instruction execution system, apparatus, or device.

[0052] "Computer memory" or "memory" is an example of a computer-readable storage medium. Computer memory is any memory that a computing system can directly access. "Computer storage" or "storage" is another example of a computer-readable storage medium. Computer storage is any non-volatile computer-readable storage medium. In some embodiments, computer storage can also be computer memory, and vice versa.

[0053] As used herein, "computing system" encompasses electronic components capable of executing programs or machine-executable instructions or computer-executable code. References to computing systems that include the term "computing system" should be interpreted as potentially encompassing multiple computing systems or processing cores. A computing system can be, for example, a multi-core processor. A computing system can also refer to a collection of computing systems, either within a single computer system or distributed across multiple computer systems. The term computing system should also be interpreted as potentially referring to a collection or network of computing devices that include processors or computing systems. Machine-executable code or instructions can be executed by multiple computing systems or processors, which may reside within the same computing device or may even be distributed across multiple computing devices.

[0054] Machine-executable instructions or computer-executable code may include instructions or programs that cause a processor or other computing system to perform aspects of the invention. Computer-executable code for performing operations of the aspects of the invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, Smalltalk, C++, etc., or conventional programming languages ​​such as the "C" programming language, or similar programming languages, and compiled into machine-executable instructions. In some instances, the computer-executable code may be in the form of a high-level language or a pre-compiled form, and may be used in conjunction with an interpreter that dynamically generates machine-executable instructions. In other instances, the machine-executable instructions or computer-executable code may be in the form of programming for programmable gate arrays.

[0055] Computer executable code can execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer via any type of network (including a local area network (LAN) or a wide area network (WAN)), or it can be connected to an external computer (e.g., via the Internet provided by an Internet service provider).

[0056] Various aspects of the invention are described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each block or portion of a block in a flowchart, illustration, and / or block diagram can be implemented, where applicable, by computer program instructions in the form of computer-executable code. It should also be understood that combinations of blocks from different flowcharts, illustrations, and / or block diagrams can be combined, provided they are not mutually exclusive. These computer program instructions can be provided to a computing system of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus to produce a machine, such that instructions executable via the computing system of the computer or other programmable data processing apparatus create means for implementing the functions / actions specified in the blocks of the flowcharts and / or block diagrams.

[0057] These machine-executable instructions or computer program instructions may also be stored in a computer-readable medium that enables a computer, other programmable processing device or other equipment to operate in a particular manner, such that the instructions stored in the computer-readable medium produce an article of writing comprising instructions that implement the functions / actions specified in blocks of flowcharts and / or block diagrams.

[0058] Machine-executable instructions or computer program instructions may also be loaded onto a computer, other programmable data processing apparatus or other equipment to cause a series of operational steps to be performed on the computer, other programmable apparatus or other equipment to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide for implementing the functions / actions specified in the blocks of the flowchart and / or block diagram.

[0059] The term "user interface" as used in this article refers to an interface that allows a user or operator to interact with a computer or computer system. A "user interface" can also be called a "human-machine interface device." A user interface can provide information or data to and / or receive information or data from an operator. A user interface enables the computer to receive input from the operator and to provide output from the computer to the user. In other words, a user interface allows an operator to control or manipulate a computer, and the interface allows the computer to indicate the effects of the operator's control or manipulation. Displaying data or information on a monitor or graphical user interface is an example of providing information to an operator. Receiving data via a keyboard, mouse, trackball, touchpad, pointing stick, graphics tablet, joystick, gamepad, webcam, headset, foot pedal, wired gloves, remote control, and accelerometer are all examples of user interface components that allow information or data to be received from the operator.

[0060] As used herein, "hardware interface" encompasses the interfaces that enable a computer system to interact with and / or control external computing devices and / or devices. A hardware interface allows a computing system to send control signals or instructions to external computing devices and / or devices. It also enables a computing system to exchange data with external computing devices and / or devices. Examples of hardware interfaces include, but are not limited to: Universal Serial Bus (USB), IEEE 1394 port, parallel port, IEEE 1284 port, serial port, RS-232 port, IEEE-488 port, Bluetooth connectivity, wireless LAN connectivity, TCP / IP connectivity, Ethernet connectivity, control voltage interfaces, MIDI interfaces, analog input interfaces, and digital input interfaces.

[0061] As used herein, "display" or "display device" includes output devices or user interfaces suitable for displaying images or data. Displays can output visual, audio, and / or tactile data. Examples of displays include, but are not limited to: computer monitors, television screens, touchscreens, tactile electronic displays, Braille screens, cathode ray tubes (CRTs), memory tubes, bistable displays, electronic paper, vector displays, flat panel displays, vacuum fluorescent displays (VFs), light-emitting diode (LED) displays, electroluminescent displays (ELDs), plasma display panels (PDPs), liquid crystal displays (LCDs), organic light-emitting diode (OLED) displays, projectors, and head-mounted displays.

[0062] This paper defines K-space data as the recorded measurements of radio frequency signals emitted by atomic spins using the antenna of a magnetic resonance imaging (MRI) device during a magnetic resonance imaging (MRI) scan. MRI data is an example of tomographic medical image data.

[0063] This article defines magnetic resonance imaging (MRI) images, MR images, or MRI data as two-dimensional or three-dimensional visualizations of reconstructed anatomical data contained within MRI data. Such visualizations can be performed using a computer. Attached Figure Description

[0064] In the following preferred embodiments, the invention will be described by way of example only and with reference to the accompanying drawings, wherein:

[0065] Figure 1 The diagram illustrates an example of a medical system;

[0066] Figure 2 The operation is shown Figure 1 A flowchart of the methods used in medical systems;

[0067] Figure 3 The diagram illustrates another example of a medical system;

[0068] Figure 4The operation is shown Figure 3 A flowchart of the methods used in medical systems;

[0069] Figure 5 The diagram illustrates another example of a medical system;

[0070] Figure 6 The operation is shown Figure 5 A flowchart of the methods used in medical systems;

[0071] Figure 7 The diagram illustrates a medical imaging method;

[0072] Figure 8 The diagram illustrates an example of an image reconstruction module;

[0073] Figure 9 The diagram illustrates a method for training a template-based AI system to compute B0 non-uniformity maps; and

[0074] Figure 10 The diagram illustrates the use Figure 9 The AI ​​system generates the B0 non-uniformity map. Detailed Implementation

[0075] Elements with the same number in these figures are either equivalent elements or perform the same function. If the functions are equivalent, the elements discussed earlier need not be discussed in the following figures.

[0076] Figure 1 The diagram illustrates an example of medical system 100. Figure 1 The medical system 100 is described as including a computer 102. The computer has an optional hardware interface 104, which, for example, enables the computer 102 to communicate with or control other components within the medical system 100. The computer 102 is shown as including a computing system 106. The computing system 106 can be implemented as a processor or multiple processors and can also be distributed across multiple locations. The computing system 106 can also be a field-programmable gate array or other system capable of performing calculations. The computer 102 is also shown as including an optional user interface 108. The user interface 108 enables an operator to control the operation and functions of the computer 102 and the medical system 100. The computer 102 is also shown as including memory 110. Memory 110 is intended to represent different types of memory accessible to the computing system 106. The computing system 106 is shown as communicating with the hardware interface 104, the user interface 108, and the memory 110.

[0077] The medical system 100 can take different forms in different paradigms. In one paradigm, the medical system 100 can be a remote server or a cloud computing component. In other paradigms, the medical system 100 can be a workstation computer used by a physician or other medical professional. In still other paradigms, the medical system 100 can be integrated into a control system that controls a magnetic resonance imaging system.

[0078] Memory 110 is shown to contain machine-executable instructions 120. The machine-executable instructions 120 enable computing system 106 to provide and perform various computational tasks. For example, this may include basic data processing, image processing, and medical image reconstruction tasks. Memory 110 is shown to contain a received initial magnetic resonance image 122. For example, it may have been received by a data carrier, or it may have been received via a network or internet connection. Memory 110 is also shown to contain subsequent k-space data 124. The initial magnetic resonance image is acquired for a first region of interest (ROI) of the object. The subsequent k-space data describes a subsequent ROI of the object. In this specific example, the subsequent ROI is within the first ROI and has a volume less than or equal to the first ROI. In other examples, the subsequent ROI and the first ROI may overlap only partially.

[0079] Memory 110 is also shown to include a B0 field estimation module 126, configured to take at least the amplitude component of the magnetic resonance image as input. In response, it outputs an estimated B0 field map 128 for subsequent k-space data 124. Memory 110 is also shown to reconstruct a corrected magnetic resonance image 130 from the subsequent k-data 124 using the estimated B0 field map 128. The estimated B0 field map 128 can be used, for example, to correct for B0 inhomogeneities during the acquisition of subsequent k-space data 124.

[0080] Figure 2 The operation is illustrated in the diagram. Figure 1 A flowchart of a method for a medical system 100 is provided. First, in step 200, an initial magnetic resonance image 122 is received. The initial magnetic resonance image 122 describes a first region of interest of the object. The initial magnetic resonance image is an amplitude image. Next, the method proceeds to step 202. In step 202, subsequent k-space data 124 is received. Next, in step 204, an estimated B0 field mapping 128 is calculated by inputting the initial magnetic resonance image 122 into a B0 field estimation module 126. Then, in step 206, a corrected magnetic resonance image 130 is calculated using the subsequent k-space data 124 and the estimated B0 field mapping 128.

[0081] Figure 3 This illustrates another example of a medical system 300. Medical system 300 and... Figure 1 It is similar to the medical system 100 in the text, except that it additionally includes a magnetic resonance imaging system 302.

[0082] The magnetic resonance imaging system 302 includes a magnet 304. Magnet 304 can also be referred to as the main magnet. Magnet 304 is a superconducting cylindrical magnet with a hole 306 passing through it. Different types of magnets may also be used; for example, both split cylindrical magnets and so-called open magnets may be used. A split cylindrical magnet is similar to a standard cylindrical magnet, except that the cryostat is split into two parts to allow access to the equiplanar plane of the magnet; such magnets can be used, for example, in conjunction with charged particle beam therapy. An open magnet consists of two magnet parts, one on top of the other, with a sufficiently large space in between to receive the object: the arrangement of these two regions is similar to that of a Helmholtz coil. Open magnets are popular because they offer fewer object restrictions. Within the cryostat of the cylindrical magnet, there is an assembly of superconducting coils.

[0083] An imaging region 308 exists within the aperture 306 of the cylindrical magnet 304, in which the magnetic field is sufficiently strong and homogeneous for performing magnetic resonance imaging. The object 318 is shown supported by an object support 320, such that at least a portion of the object 318 is within the imaging region 308. A B0 magnetic field shimming coil 322, connected to a B0 magnetic field shimming power supply 324, is also visible within the aperture 306 of the magnet 304. The hardware interface 104 can be used to control and dynamically change the shimming of the main magnetic field of the magnet 304.

[0084] Within imaging region 308, a first region of interest 326 can be seen. It can be considered to almost encompass the entire head region of object 318. Because it covers and images the entire head region, it can provide a very good estimate of the B0 magnetic field inhomogeneity caused by placing object 318 within imaging region 308. Furthermore, a subsequent region of interest 328 can be seen within imaging region 308. This region is considered to be very tightly surrounding only a portion of the head of object 318. This can be used to provide clinical or more detailed MRI images. However, because only a small portion of the head is imaged, it will not accurately produce information about B0 inhomogeneity. Areas of object 318 outside of the subsequent region of interest 328 will also have the effect of distorting the B0 or ​​main magnetic field.

[0085] A set of magnetic field gradient coils 310 is also present within the aperture 306 of the magnet, which is used to acquire preliminary magnetic resonance data for spatial encoding of the magnetic spins within the imaging region 308 of the magnet 304. The magnetic field gradient coils 310 are connected to a magnetic field gradient coil power supply 312. The magnetic field gradient coil 310 is representative. Typically, the magnetic field gradient coil 310 comprises three separate sets of coils for spatial encoding in three orthogonal spatial directions. The magnetic field gradient power supply provides current to the magnetic field gradient coils. The current supplied to the magnetic field gradient coils 310 is controlled as a function of time and can be ramped or pulsed.

[0086] Adjacent to the imaging region 308 is an RF coil 314, which is used to manipulate the orientation of the magnetic spins within the imaging region 308 and to receive radio transmissions also from the spins within the imaging region 308. The RF antenna may comprise multiple coil elements. The RF antenna may also be referred to as a channel or antenna. The RF coil 314 is connected to an RF transceiver 316. The RF coil 314 and the RF transceiver 316 may be replaced by separate transmit and receive coils and separate transmitters and receivers. It should be understood that the RF coil 314 and the RF transceiver 316 are representative. The RF coil 314 is also intended to represent a dedicated transmit antenna and a dedicated receive antenna. Similarly, the transceiver 316 may also represent a separate transmitter and receiver. The RF coil 314 may also have multiple receive / transmit elements, and the RF transceiver 316 may have multiple receive / transmit channels. For example, if a parallel imaging technique such as SENSE is performed, the RF coil 314 will have multiple coil elements.

[0087] Transceiver 316, magnetic field gradient coil power supply 312, and gradient controller 312 are shown as hardware interface 106 connected to computer system 102.

[0088] Memory 110 is also shown as containing a first pulse sequence command 330. Memory 110 is also shown as containing initial k-space data 332, which is acquired from a first region of interest 326 by controlling the magnetic resonance imaging system 302 using the first pulse sequence command 330. Memory 110 is also shown as containing a set of second pulse sequence commands 334. These represent various different magnetic resonance imaging protocols capable of acquiring k-space data. Memory 110 is also shown as containing one or a selection of the second pulse sequence commands 336. This is just one of the sets of 334. Memory 110 is also shown as containing subsequent k-space data 338, which is acquired by controlling the magnetic resonance imaging system 302 for a subsequent region of interest 328 using one of the second pulse sequence commands 336. This process can be repeated for different selections of this set of second pulse sequence commands 334. This also implies that for each acquisition, there is a different subsequent region of interest 328. This could, for example, enable a very flexible system for various acquisition estimations of the B0 field mapping 128.

[0089] Figure 4 The operation is illustrated in the diagram. Figure 3 The flowchart describes a method for a medical system 300. First, in step 400, a first pulse sequence command 330 controls a magnetic resonance imaging system 302 to acquire initial k-space data 332. Next, in step 402, an initial magnetic resonance image 122 is reconstructed based on the initial k-space data 332. Then, the method proceeds to... Figure 2 Step 200. After step 200, step 404 is executed. In step 404, the magnetic resonance imaging system 302 is controlled to acquire subsequent k-space data 338 using a 336 command in the second pulse sequence. Then, the method continues. Figure 2 Steps 202, 204, and 206 are described. Step 406 is a decision box, and the question is whether more images need to be acquired. If the answer is yes, the method returns to step 404. If the answer is no, the method returns to step 408, where the method terminates.

[0090] Figure 5 This illustrates another example of medical system 500. Figure 5 Medical System 500 illustrated in Chinese Figure 3Similar to the medical system 300, except that it additionally includes a camera system 502. The camera system 502 can be used, for example, to directly measure motion parameters describing the motion of the object 318 during the acquisition of subsequent k-space data 338. The memory 110 is also shown as containing updated shimming settings 504. These updated B0 shimming settings can be configured to reduce B0 inhomogeneities by controlling the B0 magnetic field shimming power supply 324. The camera system 502 is a motion sensor system. The memory 110 is also shown as containing an intermediate image 508 that has been reconstructed from the subsequent k-space data 338, possibly without B0 correction and without motion correction. The memory 110 is also shown as containing an image registration 510 between the intermediate image 508 and the initial magnetic resonance image 112. This can be used to determine motion parameters 506, also shown as being stored in the memory 110. The camera system 502 is also capable of measuring motion parameters 506. In some cases, both the camera 502 and the image registration 510 can be used to derive or calculate motion parameters 506.

[0091] Memory 110 is also shown optionally containing system model 512. System model 512 contains a model of the magnetic resonance imaging system 302, whose electromagnetic properties are a function of time in response to pulse sequence commands. This can be used to model conditions such as B0 inhomogeneities caused by eddy currents, temperature-dependent gradient magnetic field nonlinearities, other time-dependent gradient magnetic field nonlinearities, variations in magnetic resonance coil sensitivity, motion-dependent B1 inhomogeneities, static magnetic gradient field nonlinearities, and associated magnetic field corrections. Memory 110 is also shown containing time-correlation data 514 output by system model 512 in response to one of the input second pulse sequence commands 336. This time-correlation data 514 can be used to improve the reconstruction of the corrected magnetic resonance image 130.

[0092] Figure 6 The diagram illustrates the operation. Figure 5 The flowchart shows the method of the medical system 500. First, it executes... Figure 4 Steps 400, 402, and 200 are shown. Next, step 600 is optionally performed. In step 600, an initial B0 field mapping is estimated based on the initial image. For example, if it is desired to shim the B0 magnetic field, the initial B0 field mapping can be used for the first traversal of the algorithm. Next, the method proceeds to step 602. Step 602 is also optional. In step 602, an updated B0 shimming setting is calculated, and the updated B0 shimming setting is configured to reduce B0 inhomogeneities using either the estimated B0 field mapping or the initial B0 field mapping.

[0093] Next, optional step 604 is performed. In step 604, the B0 magnetic field is homogenized by controlling the adjustable B0 magnetic field homogenization using the updated B0 homogenization settings. After step 606, the method proceeds to... Figure 4 Step 404. Then, as... Figure 4 As shown, step 202 is also performed. Then, the method proceeds to step 608. Step 608 is also optional. In step 608, motion parameters 506 are received. The motion parameters describe the motion of object 318 during or after the acquisition of subsequent k-space data 338. After step 608, the previously performed... Figure 4 Step 204 as described herein. Following step 204, optional step 610 is performed. In step 610, time-related data 514 is calculated by inputting one of the second pulse sequence commands, 336, into the system model 512. Following step 610, the following is performed: Figure 4 Step 206 is shown. During the reconstruction of the corrected magnetic resonance image 130, if time-related data 514 is available, it can be used.

[0094] Figure 7 The illustration depicts a medical imaging method. In this example, a so-called intelligent survey is acquired corresponding to an initial magnetic resonance image 122. This involves using the survey scan to calculate the B0 magnetic field. Before each subsequent acquisition of k-space data 124, an artificial intelligence module corresponding to the B0 field estimation module 126 is used to correct and set hardware values, such as the homogenization of the B0 field. Images from the earlier scan 124 can be used to estimate motion and / or hardware settings. Finally, data from the various acquisitions are used to reconstruct the corrected magnetic resonance image 130.

[0095] Figure 8 The diagram illustrates an example of the image reconstruction module 800. Raw k-space data or subsequent k-space data 124 is input. First, reconstruction is performed, reconstructing an intermediate image 508. This intermediate image 508 is reconstructed without motion correction and possibly without correction for B0 inhomogeneities. The next step in the module is to estimate motion 608. Motion parameters are received, either from an external camera system 502 or from image registration 510. These are then input into the next part of the module, predicting 204 a new B0 for the new location of the object. Optionally, a system may also exist that models the electromagnetic properties of the magnetic resonance imaging system and models the hardware defects 610. The modeled electromagnetic properties of the magnetic resonance imaging device are time-correlated data 514, which can be calculated by inputting pulse sequence commands into the system model 512. The module then performs high-fidelity reconstruction 206, which uses the subsequent k-space data 124, the motion estimate from 608, the new B0 field from 204, and the hardware defects 610.

[0096] Figure 9 The diagram illustrates the training of the AI ​​block. In this case, the AI ​​block uses a machine learning algorithm trained using a measured image and a measured B0 field inhomogeneity 902 to co-register to a template 904. Training involves transforming parameters of the image 900 and the measured B0 field inhomogeneity 902. This combination of translations of the image and the B0 field inhomogeneity 902 can be used to train a machine language regression algorithm with these two parameters. Training on the measured B0 field inhomogeneity 902 is accomplished using co-registration parameters and spherical harmonic coefficients.

[0097] Figure 10 The diagram illustrates the use Figure 9 The system shown calculates the B0 field mapping 128. The initial magnetic resonance image 122 is co-registered to the template 904. This produces the transformed initial magnetic resonance image 122. These transformations are then input to... Figure 9 In the artificial intelligence block, the spherical harmonic coefficients can be predicted, and an estimated B0 field mapping 128 is obtained. The rotation, translation, and stretch tilt calculated for the initial magnetic resonance image can be updated using motion parameters 506, and then used to calculate the corrected estimated B0 field mapping.

[0098] As previously mentioned, magnetic resonance imaging (MRI) scanners use B0 pre-scans to improve surface-to-noise ratio (SNR) and image quality during reconstruction (e.g., reducing image distortion in EPI, MB-SENSE, and spirals). However, pre-scans can add significant time (up to 10%) to the overall scan duration. Furthermore, despite the use of B0 pre-scans, MRI images can sometimes be affected by artifacts, ranging from subtle to severe image degradation. Typically, motion and breathing, combined with scanner hardware limitations, produce artifacts. The presence of artifacts necessitates repeated scans, leading to reduced scanner throughput and decreased patient comfort. Moreover, it affects measurement reliability by degrading image quality. This is a widespread problem affecting virtually all MRI systems and all application areas.

[0099] As in Figure 6 or Figure 8 The exemplary image acquisition / reconstruction framework described herein addresses this problem comprehensively by explaining the fundamental physics behind issues in a broad class of fast MRI scans (EPI, MB-SENSE, spiral, etc.). By eliminating the B0 preparatory scan required for MRI examinations, the paradigm can also directly reduce scan time. AI can be a key enabler in both the acquisition and reconstruction phases. The paradigm offers the combined benefits of improved image quality and reduced scan time.

[0100] Even in well-calibrated MRI scanners, artifacts can affect image quality to varying degrees, depending on the scan type and the anatomical structure of interest. The root causes of artifacts can be subdivided into two main types, listed below along with their causes:

[0101] 1. System defects (examples are given below)

[0102] a. Eddy currents (capable of fully modeling all acquisition conditions in space and time)

[0103] b. Gradient nonlinearity (fully models all conditions)

[0104] 2. Related to human physiology (examples are given below)

[0105] a. Magnetic field inhomogeneity (depending on the region of interest)

[0106] b. Movement, respiration, etc. (difficult to describe, but can be measured using imaging tools)

[0107] System defects can be adequately quantified by performing calibration scans during setup, and their effects are typically mitigated during acquisition and / or image reconstruction. However, physiologically induced artifacts are difficult to predict. Preparatory scans (B0 pre-scans) mitigate some sources, such as magnetic field inhomogeneities. However, this results in increased scan time and can lead to significant overhead. Most attempts to mitigate artifacts attempt to use image processing, or there are no solutions at all, requiring repeated scans. Specifically, motion-induced artifacts are difficult to quantify and correct, especially in dynamic scans that acquire a series of image volumes over time.

[0108] One often overlooked but important issue is that all sources of artifacts are interactive and cannot be corrected using post-processing. For example, in the case of brain scans, head movements alter magnetic field inhomogeneities, leading to varying effects of eddy currents and gradient nonlinearities. One way to maintain image fidelity under such conditions is to incorporate all artifact sources into a single image reconstruction framework. This has resulted in an excessive number of post-processing options that attempt to correct the different artifacts piecemeal, without any single approach providing consistent image fidelity to account for all the causes occurring simultaneously during reconstruction.

[0109] The paradigm addresses two technical issues. A) In some paradigms, it eliminates the need for pre-scanning of B0 using AI-based predictive acquisition. B) Even in motion scenarios, it provides a true-fidelity reconstruction by using the same AI-based predictive strategy to predict residual B0 inhomogeneities after shimming, which will then be used as input to the reconstruction algorithm.

[0110] The B0 pre-scans routinely performed on the scanner prior to multiple scans do not directly provide any diagnostic value. However, they are currently crucial for improving SNR (spot rinsing). They are also used to provide input to reconstruction algorithms to improve image quality. However, this comes with the assumption that the patient does not move between the pre-scan and main scan. This assumption is often violated in practice, leading to a deterioration in image quality.

[0111] The paradigm eliminates the need for a pre-scan of B0, resulting in a reduction in direct scan time. This can save 10% or more of the total scan time. Furthermore, it provides a way to predict B0 inhomogeneities, thereby directly improving image quality across a large class of scans, even in the presence of motion. Therefore, the paradigm can provide a motion-stabilized, scan-time-reduced acquisition and reconstruction framework.

[0112] Recent studies have highlighted that human B0 heterogeneity is remarkably similar across subjects, regardless of the anatomy considered. The main variations in B0 heterogeneity stem from differences in intra-scan anatomical shape and patient positioning. Both of these will be captured via investigative scans (initial MRI images 122) performed during each MR examination to aid in planning.

[0113] Examples can be beneficial for very widely used fast imaging scans, such as echo-planar imaging (EPI) and its variants (e.g., MB-SENSE), as well as scans with great clinical potential that have not yet been widely adopted due to their susceptibility to artifacts (e.g., spiral imaging). Both types of scans typically capture image volume within seconds.

[0114] The examples also provide implementations of a comprehensive image acquisition / reconstruction framework, such as Figure 6 and Figure 8 As shown in the figure. This framework can be used to account for time-dependent data 514, such as eddy currents, gradient nonlinearities, and magnetic field inhomogeneities. In the example, the reconstruction framework can include motion-related effects (direct and indirect), but can also reduce scan time by eliminating the need for a preparatory scan performed to measure magnetic field inhomogeneities (by determining the estimated B0 field mapping 128).

[0115] One obstacle in developing a comprehensive image reconstruction framework that incorporates motion is based on assumptions made by current image reconstruction techniques. For example, current EPI and spiral image reconstruction perform a pre-scan to measure magnetic field inhomogeneities, but inherently assumes no motion between the pre-scan and the main scan (EPI / spiral) acquisition. However, this assumption can be violated, particularly in dynamic scans involving multiple volumes, where object motion often leads to unintended artifacts. Contrary to the disclosed paradigms, current frameworks often require repeated pre-scans to measure magnetic field inhomogeneities, which is entirely impractical in dynamic scans where motion occurs continuously (especially in the torso region).

[0116] Recent research has shown that in brain scans, motion-induced magnetic field inhomogeneities can be predicted if the initial distribution of magnetic field inhomogeneities is known. While this concept has been used to correct EPI images in post-processing, this paradigm can be directly incorporated into the reconstruction framework presented herein. The paradigm can use AI-assisted survey scans to predict magnetic field inhomogeneities, thereby eliminating the need for preparatory scans to measure these inhomogeneities.

[0117] The example collection / reconstruction framework may involve one or more of the following steps:

[0118] a) Perform an investigation scan for planning (e.g., EPI / spirometry) (receive 200 initial MRI images 122)

[0119] b) Predict magnetic field inhomogeneities from survey scans using AI (compute estimated B0 field mapping 204 or compute initial B0 field mapping 600)

[0120] c) Improve magnetic field homogeneity by setting appropriate shimming currents (calculate 602 updated shimming settings 504 and shimming of the B0 magnetic field 604), and also predict residual magnetic field inhomogeneities (using B). 0 The shim setting 606 is varied to adjust the estimated B0 field mapping. This can be accomplished using simple calculations.

[0121] d) Perform EPI / spiral dynamic scanning (multiple volumes over time, such as DTI, fMRI, injection tracking, etc.) (acquire 404 subsequent k-space data 124)

[0122] e) Using the residual magnetic field obtained from step c and leveraging knowledge of system defects, the first image volume in the sequence is reconstructed with high fidelity. It should be noted that there may be no time gap between the prediction in step c and the first volume in the acquisition time series. (Reconstruction of 206-corrected magnetic resonance image 130)

[0123] f) For each subsequent image volume

[0124] 1. Perform low-quality reconstruction (reconstructing intermediate image 508) using an existing framework instead of the proposed framework, and estimate motion parameters (translation, rotation, tilt, and stretching) using image co-registration (registration 510) of the first volume (first region of interest 326). Alternatively, reliable motion estimates may be obtained from other sources, such as camera system 502.

[0125] 2. Use motion parameters to predict residual magnetic field inhomogeneities using AI (using B0 field estimation module 126 to calculate 128 estimated B0 field mapping 128)

[0126] 3. High-fidelity reconstruction (corrected MRI image 130) is obtained by using motion parameter 506 and the predicted residual magnetic field inhomogeneity and updated system defects (to interpret motion) within the integrated reconstruction framework.

[0127] Some paradigms can provide B0 pre-scan predictions from the survey scan (estimated B0 field mapping 128), and thus enable its elimination from the actual scan, regardless of the scanner (field intensity and variants), scan type, and underlying anatomical structures (although our initial tests were limited to brain scans). This pre-scan elimination directly leads to a reduction in scan time.

[0128] Other paradigms can provide for incorporating residual B0 in image reconstruction applications into all field intensities and variants, and those major classes of MRI scans currently in use and newly developed in the field, such as EPI distortion correction, SENSE, MB-SENSE, CSENSE, etc., which are already in use in the field, can benefit from the proposed comprehensive reconstruction framework, provided that the acquisition volume can be acquired in a few seconds (<10 seconds), which is what they typically do.

[0129] Although the invention has been described in detail in the accompanying drawings and the foregoing description, such description should be considered illustrative or exemplary rather than restrictive; the invention is not limited to the disclosed embodiments.

[0130] Those skilled in the art, in practicing the claimed invention, will be able to understand and implement other variations of the disclosed embodiments through study of the drawings, disclosure, and appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite articles "a" or "an" do not exclude multiple. A single processor or other unit can perform the functions of several items listed in the claims. The fact that certain measures are listed merely in mutually different dependent claims does not indicate that combinations of these measures cannot be used advantageously. Computer programs can be stored / distributed on suitable media, such as optical storage media or solid-state media provided with or as part of other hardware, but can also be distributed in other forms, such as via the Internet or other wired or wireless telecommunications systems. Any reference numerals in the claims should not be construed as limiting the scope of protection.

[0131] List of reference numerals

[0132] 100 Medical System

[0133] 102 Computer

[0134] 104 hardware interface

[0135] 106 Computing System

[0136] 108 User Interface

[0137] 110 memory

[0138] 120 Machine Executable Instructions

[0139] 122 Initial Magnetic Resonance Images

[0140] 124 Subsequent k-space data

[0141] 126B0 Field Estimation Module

[0142] 128-estimated B0 field mapping

[0143] 130-corrected magnetic resonance images

[0144] 200 receives an initial magnetic resonance image of a first region of interest describing the object, wherein the initial magnetic resonance signal is an amplitude image.

[0145] 202 receives subsequent k-space data of a subsequent region of interest describing the object, wherein the subsequent region of interest is within the first region of interest.

[0146] 204 By inputting the initial magnetic resonance image into the B0 field estimation module, the estimated B0 field mapping of the subsequent region of interest is calculated from the initial magnetic resonance image.

[0147] 206 Corrected magnetic resonance images reconstructed from subsequent k-space data and estimated B0 field mapping

[0148] 300 Medical System

[0149] 302 Magnetic Resonance Imaging System

[0150] 304 main magnet

[0151] 306 magnet hole

[0152] 308 imaging area

[0153] 310 magnetic gradient coil

[0154] 312 magnetic field gradient coil power supply

[0155] 314 RF coil

[0156] 316 transceiver

[0157] 318 objects

[0158] 320 object support

[0159] 322B0 magnetic field homogenizing coil

[0160] 324B0 Magnetic Field Uniform Power Supply

[0161] 326 First Region of Interest

[0162] Areas of interest following 328

[0163] 330 First Pulse Sequence Command

[0164] 332 Initial k-space data

[0165] 334 A set of second pulse sequence commands

[0166] 336 Second Pulse Sequence Command

[0167] 338 Subsequent k-space data

[0168] 400 uses the first pulse sequence command to control the magnetic resonance imaging system to acquire initial k-space data.

[0169] 402 Reconstructing the initial magnetic resonance image from initial k-space data

[0170] 404 uses one of the second pulse sequence commands in this set to control the magnetic resonance imaging system to acquire subsequent k-space data for each iteration.

[0171] 500 Medical System

[0172] 502 Camera System

[0173] 504 Updated Shimming Settings

[0174] 506 motion parameters

[0175] 508 intermediate image

[0176] 510 registration

[0177] 512 system module

[0178] Data related to 514 time

[0179] 600 Calculate the initial from the initial magnetic resonance image B0 Field mapping

[0180] 602 uses the estimated B0 field mapping to calculate the updated B0 shimming setting configured to reduce B0 inhomogeneities.

[0181] 604 homogenizes the B0 magnetic field by controlling the adjustable B0 magnetic field homogenization with the updated B0 homogenization settings.

[0182] 606 uses variations in the B0 shimming setting to adjust the estimated B0 field mapping.

[0183] 608 receives motion parameters describing the motion of the object during or after the acquisition of subsequent k-space data.

[0184] 610 Calculation Time Related Data

[0185] 800 Reconstruction Module

[0186] 900 measured image

[0187] B0 field inhomogeneity measured by 902

[0188] 904 template

Claims

1. A medical system (100, 300, 500), comprising: - A memory (110) storing machine-executable instructions (120) and a B0 field estimation module (126), the B0 field estimation module (126) being configured to output an estimated B0 field map (128) in response to receiving at least one amplitude component of an initial magnetic resonance image as input. as well as - A computing system (106) configured to control the medical system, wherein the execution of the machine-executable instructions causes the computing system to receive (200) the initial magnetic resonance image (122), wherein the initial magnetic resonance image describes a first region of interest (326) of an object (118). The execution of the machine-executable instructions also causes the computing system to perform at least one of the following operations: - Receive (202) subsequent k-space data (124) describing subsequent different regions of interest (328) of the object, wherein the subsequent regions of interest at least partially overlap with the first region of interest; - Determine motion parameters based on measurements received from the motion sensor system and / or registration of the initial magnetic resonance image with an intermediate image generated based on the subsequent k-space data; - By inputting the initial magnetic resonance image and the motion parameters into the B0 field estimation module, the estimated B0 field mapping (128) for the subsequent region of interest is calculated (204); and - Reconstruct (206) the corrected magnetic resonance image (130) based on the subsequent k-space data, the motion parameters and the estimated B0 field mapping.

2. The medical system according to claim 1, wherein, The medical system further includes a magnetic resonance imaging system (302) configured to acquire k-space data from an imaging region (308), wherein the memory further contains a first pulse sequence command (330) configured to acquire initial k-space data (332) from a first region of interest, wherein the memory further contains a set of second pulse sequence commands (334), each second pulse sequence command configured to acquire subsequent k-space data from a subsequent region of interest, wherein the execution of the machine-executable instructions further enables the computing system to: - The magnetic resonance imaging system (400) is controlled to acquire the initial k-space data using the first pulse sequence command; and - Reconstruct the initial magnetic resonance image (402) based on the initial k-space data; and - Using one (336) of the set of second pulse sequence commands, control (404) the magnetic resonance imaging system to acquire the subsequent k-space data for each iteration, wherein each iteration includes reconstructing a corrected magnetic resonance image based on the subsequent k-space data and the estimated B0 field mapping.

3. The medical system according to claim 2, wherein, The magnetic resonance imaging system includes a main magnet (304) for generating a B0 magnetic field in the imaging region, wherein the magnetic resonance imaging system further includes adjustable B0 magnetic field shims (322, 324), the adjustable B0 magnetic field shims being configured to shim the B0 magnetic field in the imaging region. The execution of the machine-executable instructions also causes the computing system to perform the following operations before each acquisition of subsequent k-space data: - The calculation (602) is configured to use the estimated B0 field mapping to reduce the updated B0 shimming setting (504) of B0 inhomogeneity; and - The B0 magnetic field is homogenized by controlling the adjustable B0 magnetic field homogenization using the updated B0 homogenization setting (604).

4. The medical system according to claim 3, wherein, The estimated B0 field mapping is calculated (606) using at least part of the updated B0 shimming settings.

5. The medical system according to claim 3 or 4, wherein, The medical system also includes a motion sensor system (502) configured to at least partially measure the motion parameters, wherein execution of the machine-executable instructions further causes the computing system to control the motion sensor system to measure the motion parameters.

6. The medical system according to claim 3 or 4, wherein, The execution of the machine-executable instructions also enables the computing system to: - Reconstruct the intermediate image based on the subsequent k-space data (508); - Calculate the registration between the intermediate image and the initial magnetic resonance image (510); and - Calculate the motion parameters based on the registration.

7. The medical system according to claim 3 or 4, wherein, The estimated B0 field mapping is calculated in part by at least one of the following: - Use the analytical model to calculate the spatial transformation of the estimated B0 field mapping using the motion parameters; - The initial magnetic resonance image and the motion parameters are input into the trained neural network.

8. The medical system according to any one of claims 2 to 4, wherein, The memory also includes a system model (512) configured to output time-related data (514) describing the electromagnetic properties of the magnetic resonance imaging system in response to input of one of the set of second pulse sequence commands, wherein execution of the machine-executable instructions further enables the computing system to compute the time-related data by inputting the subsequent pulse sequence commands into the system model, wherein the corrected magnetic resonance image is reconstructed based on the subsequent k-space data, the estimated B0 field mapping, and the time-related data.

9. The medical system according to claim 8, wherein, The time-related data describes any of the following: B0 non-uniformity caused by eddy currents, temperature-related gradient magnetic field non-linearity, time-related gradient magnetic field non-linearity, changes in magnetic resonance coil sensitivity, motion-related B1 non-uniformity, static magnetic gradient field non-linearity, accompanying magnetic field corrections, and combinations thereof.

10. The medical system according to any one of claims 2 to 4, wherein, The second pulse sequence command is configured to acquire the subsequent k-space data according to any of the following: - According to the echo-planar imaging magnetic resonance imaging protocol; - According to the multi-band magnetic resonance imaging protocol; - Utilizing a spiral k-space sampling mode; - Utilizing a non-Cartesian sampling mode; and - A combination of the above items.

11. The medical system according to any one of claims 1 to 4, wherein, The B0 field estimation module is implemented as any one of the following: - B0 modeling machine learning system; -B0 models neural networks; - B0 modeling of random forest regression system; -B0 Support Vector Machine Learning System; as well as - Template-based B0 magnetic field predictor system.

12. The medical system according to any one of claims 1 to 4, wherein, Any one of the following: - in, The subsequent region of interest is within the first region of interest, wherein the subsequent region of interest has a volume less than or equal to the first region of interest; and - Wherein, the subsequent region of interest has a volume larger than the first region of interest.

13. A computer program product storing a computer program, said computer program comprising machine-executable instructions for execution by a computing system (106) controlling a medical system (100, 300, 500), wherein, The execution of the machine-executable instructions causes the computing system to: receive (200) an initial magnetic resonance image (122) of a first region of interest (326) of a describing object (118), wherein the initial magnetic resonance image includes an amplitude component; The execution of the machine-executable instructions also causes the computing system to perform at least one of the following operations: - Receive (202) subsequent k-space data (124) describing subsequent different regions of interest (328) of the object, wherein the subsequent regions of interest at least partially overlap with the first region of interest; - Determine motion parameters based on measurements received from the motion sensor system and / or registration of the initial magnetic resonance image with an intermediate image generated based on the subsequent k-space data; - By inputting the initial magnetic resonance image and the motion parameters into the B0 field estimation module, calculate (204) an estimated B0 field mapping (128) for the subsequent region of interest, wherein the B0 field estimation module is configured to output the estimated B0 field mapping in response to receiving at least the amplitude component of the initial magnetic resonance image; and - be configured to reconstruct (206) a corrected magnetic resonance image (130) based on the subsequent k-space data, the motion parameters and the estimated B0 field mapping.

14. A medical imaging method, wherein, The method includes receiving (200) an initial magnetic resonance image (122) of a first region of interest (326) of a descriptive object (118), wherein the initial magnetic resonance image includes an amplitude component; The method includes performing the following operations at least once: - Receive (202) subsequent k-space data (124) describing subsequent different regions of interest of the object, wherein the subsequent regions at least partially overlap with the first region of interest; - Determine motion parameters based on measurements received from the motion sensor system and / or registration of the initial magnetic resonance image with an intermediate image generated based on the subsequent k-space data; - By inputting the initial magnetic resonance image and the motion parameters into the B0 field estimation module, an estimated B0 field mapping (128) for the subsequent region of interest is calculated (204), wherein the B0 field estimation module is configured to output the estimated B0 field mapping in response to receiving at least the amplitude component of the initial magnetic resonance image as input; and - Reconstruct (206) the corrected magnetic resonance image (130) based on the subsequent k-space data, the motion parameters and the estimated B0 field mapping.

Images