Shimming method and shimming apparatus
A patient-specific B1 magnetic field shimming method using projection data to determine shimming parameters in MRI reduces scan time to 1 second, improving image quality and diagnostic accuracy by addressing the inefficiencies of current methods.
Patent Information
- Application Number
- JP2024038535
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-04-28
- Filing Date
- 2024-03-13
- Publication Date
- 2024-11-08
AI Technical Summary
Current B1 magnetic field shimming methods in MRI are lengthy, taking tens of seconds to minutes, which is undesirable for clinical applications requiring faster scan times.
A patient-specific B1 magnetic field shimming method that acquires B1 projection data along determined directions using MRI apparatus, determining shimming parameters based on this data to achieve uniform B1 field distribution, reducing scan time to around 1 second.
The method significantly shortens scan time to 1 second, improving image quality by reducing dielectric artifacts and ensuring accurate RF excitation, thus enhancing diagnostic accuracy and patient safety.
Smart Images

Figure 2024159513000001_ABST
Abstract
Description
[Technical field]
[0001] The embodiments disclosed herein and in the drawings relate to a shimming method and apparatus.
[0002] The present disclosure relates to patient-specific radio frequency (RF) magnetic field (i.e., "B1 magnetic field") shimming in Magnetic Resonance Imaging (MRI) devices. B1 magnetic field shimming is performed in vivo based on projection data acquired along one or more projections determined according to the patient's anatomy throughout the patient being imaged. [Background technology]
[0003] The discussion of the background art provided herein is intended to provide a general overview of the background of the present disclosure, and the inventors' work is not admitted, expressly or impliedly, as prior art to the present disclosure, to the extent described in this Background section, as well as aspects of the specification that are not admitted as prior art at the time of filing.
[0004] As part of the procedure to obtain MRI images of the inside of a patient's body, the MRI machine uses a static magnetic field (B0) to tune the nuclear spins of atoms. During a scan, RF pulses generated by an RF transmitter perturb the local magnetic field, and RF signals emitted by the nuclear spins are detected by an RF receiver.
[0005] In clinical practice, the strength of the B0 magnetic field is becoming higher (from 1.5T to 3T or higher) to achieve diagnostic images with high spatial and contrast resolution. However, at higher B0 magnetic fields, the RF behavior in the patient becomes more complicated. For example, the dielectric properties of the human body can cause local perturbations in the B1 magnetic field, resulting in inhomogeneous excitation. This can lead to errors in the contrast of the obtained diagnostic images, which can lead to misdiagnosis.
[0006] To address this issue, it is common to generate a map of the RF transmit field and perform shimming correction according to the map to obtain a more uniform B1 field distribution. In a spatially resolved B1 map, each pixel represents an index of the transmit field B1 at its location. In addition to B1 field shimming, the B1 map may also be used for RF transmit calibration (for accurate RF pulse flip angles), parallel transmission (pTx) RF pulse control (e.g., pTx is typically required for 7T MRI), and correction of quantitative relaxometry maps (typically associated with longitudinal relaxation time (T1) mapping). The B1 map is acquired during a prescan procedure, and the B1 map may then be used to calibrate, design, or correct data in the acquisition of subsequent sequences in the protocol.
[0007] To shorten the scan time, it is effective to shorten the prescan as much as possible. Although there are several B1 mapping techniques, a common drawback is that the measurement time is unrealistically long. At clinical magnetic field strengths (e.g., 1.5T, 3T, etc.), a B1 shimming method that can be performed within a scan time of about 1 second and correct B1 magnetic field fluctuations has not yet been established.
[0008] It would therefore be desirable to overcome these and other shortcomings in current approaches. [Prior art documents] [Patent documents]
[0009] [Patent Document 1] US Patent Application Publication No. 2013 / 0134972 [Patent Document 2] U.S. Patent No. 9,575,153 [Patent Document 3] US Patent Application Publication No. 2005 / 0054910 [Patent Document 4] US Patent Application Publication No. 2021 / 10161382 [Patent Document 5] U.S. Patent No. 10,126,395 [Patent Document 6] European Patent Application Publication No. 2734855 Summary of the Invention [Problem to be solved by the invention]
[0010] One of the problems to be solved by the embodiments disclosed in this specification and the drawings is to shorten the scan time of the prescan. However, the problems to be solved by the embodiments disclosed in this specification and the drawings are not limited to the above problem. Problems corresponding to the effects of each configuration shown in the embodiments described later can also be positioned as other problems. [Means for solving the problem]
[0011] A shimming method according to an embodiment is a shimming method for performing patient-specific B1 magnetic field shimming in a magnetic resonance imaging apparatus, and includes the steps of acquiring patient information of a patient to be imaged by the magnetic resonance imaging apparatus, determining a direction of projection based on the acquired patient information, acquiring B1 projection data along the determined direction of projection using the magnetic resonance imaging apparatus, and determining a set of B1 shimming parameters based on the acquired B1 projection data. [Brief description of the drawings]
[0012] Various embodiments of the present disclosure will now be described in detail, by way of example only, with reference to the following drawings, in which like reference numerals indicate like elements and in which:
[0013] [Figure 1] FIG. 1 illustrates an exemplary dielectric artifact in an MRI image. [Diagram 2]FIG. 2 illustrates exemplary B1 field maps acquired in four different imaging anatomy: head, abdomen, chest, and thigh. [Diagram 3] FIG. 3 illustrates a non-limiting example of a block diagram of a B1 field shimming apparatus according to one embodiment of the present disclosure. [Figure 4] FIG. 4 illustrates a non-limiting example flow chart of a B1 field shimming method according to one embodiment of the present disclosure. [Diagram 5] FIG. 5 illustrates a non-limiting example of a block diagram of a projection determination circuit that determines one or more projections for which projection data is acquired by an MRI device, according to one embodiment of the present disclosure. [Figure 6] FIG. 6 illustrates a non-limiting example of a flow chart of a process for determining one or more projections for which projection data is acquired by an MRI device, according to an embodiment of the present disclosure. [Figure 7] FIG. 7 is a diagram illustrating non-limiting examples of different projections and projection data acquired along each of the projections, according to one embodiment of the present disclosure. [Figure 8] FIG. 8 is a diagram illustrating non-limiting examples of different projection combinations identified for different imaged anatomies according to one embodiment of the present disclosure. [Figure 9] FIG. 9 illustrates a non-limiting example of a block diagram of a shimming parameter analysis circuit for analyzing a set of shimming parameters based on acquired projection data in accordance with an embodiment of the present disclosure. [Figure 10] FIG. 10 illustrates a non-limiting example flow chart of a process for determining a set of shimming parameters based on acquired projection data in accordance with an embodiment of the present disclosure. [Figure 11] FIG. 11 illustrates a non-limiting example of a block diagram of a shimming parameter analysis circuit for analyzing a set of shimming parameters based on acquired projection data in accordance with an embodiment of the present disclosure. [Figure 12]FIG. 12 illustrates a non-limiting example of a flowchart of a process for analyzing a set of shimming parameters based on acquired projection data in accordance with an embodiment of the present disclosure. [Figure 13] FIG. 13 is a diagram illustrating a non-limiting example of obtaining profiles of projection data along two or more directions for different sets of shimming parameters according to an embodiment of the present disclosure. [Figure 14] FIG. 14 is a diagram illustrating a non-limiting example of obtaining profiles of projection data along one or more directions for different sets of shimming parameters according to an embodiment of the present disclosure. [Figure 15] FIG. 15 illustrates a non-limiting example of a block diagram of a shimming parameter analysis circuit for analyzing a set of shimming parameters based on acquired projection data in accordance with an embodiment of the present disclosure. [Figure 16] FIG. 16 illustrates a non-limiting example of a flowchart of a process for analyzing a set of shimming parameters based on acquired projection data in accordance with an embodiment of the present disclosure. [Figure 17] FIG. 17 is a schematic block diagram of a magnetic resonance imaging apparatus according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0014] The present disclosure relates to a shimming method for performing patient-specific B1 magnetic field shimming in a magnetic resonance imaging apparatus, the shimming method including the step of acquiring patient information of a patient imaged by the magnetic resonance imaging apparatus. The shimming method further includes the step of determining a direction of projection based on the acquired patient information. The shimming method further includes the step of acquiring B1 projection data along the determined direction of projection using the magnetic resonance imaging apparatus. The shimming method further includes the step of determining a set of B1 shimming parameters based on the acquired B1 projection data.
[0015] The present disclosure also relates to a shimming device for performing patient-specific B1 magnetic field shimming in a magnetic resonance imaging device. The shimming device includes a processing circuit. The processing circuit acquires patient information of a patient imaged by the magnetic resonance imaging device. The processing circuit further determines a direction of projection based on the acquired patient information. The processing circuit further acquires B1 projection data along the determined direction of projection using the magnetic resonance imaging device. The processing circuit further determines a set of B1 shimming parameters based on the acquired B1 projection data.
[0016] It is noted that the above description does not identify all embodiments and / or additional novel aspects of the present disclosure or claimed invention. Instead, the above description merely provides an introduction to various embodiments and equivalent novel aspects. For further details and / or possible aspects of the present disclosure and embodiments, please refer to the detailed description and corresponding drawings further described below.
[0017] The following disclosure presents many different embodiments or examples for implementing various features of the presented subject matter. To facilitate understanding of the disclosure, specific examples of components and arrangements are described below. These are, of course, merely examples and are not intended to be limiting.
[0018] For example, although the order of description of various steps as described herein is presented for purposes of clarity, in general these steps may be performed in any suitable order. Also, although different features, techniques, configurations, etc. may be described herein in different parts of the disclosure, each concept may be performed independently of one another or in combination with one another. Thus, the disclosure may be embodied and viewed in many different ways.
[0019] Furthermore, as used herein, the term "a" or "an" is generally intended to mean "one or more" unless otherwise specified.
[0020] When electromagnetic waves encounter an object in an MRI machine, dielectric effects occur: wavelengths shorten, electric currents are generated, and waves are reflected / refracted at tissue interfaces. Under high B0 fields (3T and above), the homogeneity of the time-varying RF magnetic field (B1 field) is highly dependent on the electrical properties of the object (e.g., the patient's body). Standing wave currents flow in opposite directions from both sides of the patient, generating patterns of destructive interference (dark areas) and constructive interference (light areas) separated by about a quarter wavelength. These abnormal bright and dark areas due to the inhomogeneity of the B1 field are commonly called "dielectric artifacts".
[0021] 1(a) and (b) show exemplary dielectric artifacts in MRI images of the human body. As can be seen from FIG. 1(a), the center and front of the brain image are abnormally bright. In FIG. 1(b), the center of the abdominal image is abnormally dark. These dielectric artifacts increase the difficulty of analyzing MRI images and obtaining clinical information, and may cause misdiagnosis if they become severe.
[0022] To reduce dielectric artifacts, one method is to obtain information describing the B1 field distribution for each patient so that B1 shimming can be performed for each patient, which can improve the homogeneity of the B1 field and, as a result, the image quality and, in some cases, the Specific Absorption Rate (SAR) or B1 rms safety.
[0023] So far, many B1 shimming methods have been developed. Typically, these methods involve acquiring spatially resolved maps using standard Cartesian phase encoding. Because the B1 field varies slowly spatially, the encoding matrix can be small (corresponding to low resolution). Typically, the matrix size is, for example, 32x32, 48x48, or 64x64. A single slice acquired at the isocenter is often used to represent the B1 field distribution of the volume. It is also possible to acquire 3D B1 maps, at the expense of longer scan times.
[0024] However, even with small matrix sizes (e.g., 32x32, 48x48, or 64x64), it takes tens of seconds to minutes to acquire enough data to construct a B1 map. Although some fast methods (e.g., Bloch-Siegert Shift) are available, 64 shots (32 phase encoding x 2 frequency offsets) are still required to create a low-resolution B1 map. For example, with TR=200ms, it would take 12.8 seconds to acquire B1 data per channel. If separate B1 maps are acquired for the two channels, the total prescan time would be about 26 seconds.
[0025] In a typical clinical B1 shimming, a full Cartesian encoded map is not required. Typically, there are only two control parameters: amplitude attenuation and phase offset {A,φ} (i.e., the setting of the TX2 channel relative to the TX1 channel, or the setting of the two drive ports on a quadrature TX coil). Furthermore, it is known that the B1 field varies slowly spatially with a characteristic + / - pattern due to dielectric effects. In addition, an estimate of the shape of the B1 field is known. Figures 2(a)-(d) show exemplary B1 field maps acquired in four different imaging anatomy, namely, head, abdomen, chest, and thigh. Each B1 map illustrated in Figures 2(a)-(d) has a characteristic distribution that corresponds to the + / - pattern caused by dielectric effects. Maps from individual patients conform to similar characteristic patterns, although there are differences in aspect ratio, spatial scaling, and B1 inhomogeneity scaling.
[0026] Thus, when used for B1 shimming purposes, the B1 map does not contain any useful information beyond that of a 2x2 image. By acquiring projection data instead of Cartesian encoding, sufficient information can be extracted from one or more 1D projections. In particular, the symmetry of the projection data can reveal adequate information to calculate the optimal set of {A,φ}. To solve for {A,φ}, no map is needed, all that is needed is a few metrics of the B1 spatial distribution. In this way, by acquiring only a few projection data (e.g., 1, 2, 3, 4) instead of 32 phase encoding, the scan time can be significantly reduced from tens of seconds to around 1 second.
[0027] FIG. 3 illustrates a non-limiting example of a block diagram of a B1 magnetic field shimming device according to an embodiment of the present disclosure. The B1 magnetic field shimming device 300 includes a projection determination circuit 310, a projection data acquisition circuit 320, and a shimming parameter analysis circuit 330. The projection determination circuit 310 acquires information about a patient to be imaged (patient information), determines one or more projections for which projection data are acquired, determines gradient magnetic field signals required to realize the projections, and applies the gradient magnetic field signals to a gradient magnetic field coil driver of the MRI device. The projection data acquisition circuit 320 acquires projection data from an RF receiver of the MRI device and transmits the projection data to the shimming parameter analysis circuit 330. The shimming parameter analysis circuit 330 analyzes a set of shimming parameters based on the projection data (and optionally the patient information), and outputs the set of shimming parameters to an RF transmitter of the MRI device.
[0028] FIG. 4 illustrates a non-limiting example of a flow chart of a B1 field shimming method according to an embodiment of the present disclosure. The B1 field shimming method process 400 begins by acquiring patient information at step 410. The patient information includes the patient's anatomy to be imaged. At step 420, the direction and number of projections are determined based on the imaged anatomy. The projection directions may be initialized with respect to the patient's anatomical axial directions, such as left-right, head-foot, anterior-posterior, etc. The patient information may include information regarding the patient's orientation within the MRI device, such as prone versus supine, and head-foot position. The patient's orientation in addition to the anatomical projection directions is sufficient to define the projections in the device's hardware coordinates, such as x, y, and z gradient directions. Such geometric transformations are well known in MRI. At step 430, the MRI device acquires projection data along each of the determined projections. At step 440, a set of shimming parameters is analyzed based on the acquired projection data. Optionally, the patient information is also used in the analysis of the shimming parameters. In step 450, the analyzed shimming parameters are applied to RF transmitters to control individual transmit channels of the MRI system.
[0029] The components of FIG. 3 and the processes performed by those components will now be described in more detail with reference to FIGS.
[0030] FIG. 5 shows a non-limiting example of a block diagram of a projection determination circuit for determining one or more projections for which projection data is acquired by an MRI device, according to an embodiment of the present disclosure. The projection determination circuit 310 includes an imaging anatomy extraction circuit 510, an anatomy projection look-up table (LUT) 520, and a gradient signal generation circuit 530. The imaging anatomy extraction circuit 510 receives information about a patient (patient information) and extracts the anatomy of the patient to be imaged. A large number of anatomy projection pairs are pre-stored in the anatomy projection look-up table 520. By comparing the imaging anatomy against the entries stored in the anatomy projection look-up table 520, one or more matching projections are searched for and sent to the gradient signal generation circuit 530. Upon receiving the matching projections, the gradient signal generation circuit 530 generates the gradient signals required to realize the projections and outputs them to the gradient coil driver of the MRI device.
[0031] FIG. 6 illustrates a non-limiting example of a flow chart of a process for determining one or more projections for which projection data is acquired by an MRI device, according to an embodiment of the present disclosure. On the left side of FIG. 6 is a process 620 for determining a projection according to received patient information. In step 622, information about the imaged anatomy is extracted from the patient information. In step 624, a lookup is performed in the anatomy projection lookup table 520 to acquire one or more projections that match the imaged anatomy. In step 626, gradient magnetic field signals required to realize each projection are generated. In step 628, the generated gradient magnetic field signals are applied to the gradient coil drivers of the MRI device.
[0032] On the right side of FIG. 6 is a process 660 for offline preparation of an anatomy projection lookup table 520 used in process 620. In step 662, an electromagnetic (EM) simulation is performed to obtain B1 magnetic field distributions corresponding to different imaging anatomy. In step 664, a Principal Component Analysis (PCA) is performed on each B1 magnetic field distribution. In step 666, one or more characteristic projections are identified for each B1 magnetic field distribution. In step 668, matched anatomy projection pairs are created and stored in the anatomy projection lookup table 520. Although in FIG. 6, an electromagnetic simulation is used to prepare the anatomy projection lookup table 520, a person skilled in the art will understand that other methods are possible. For example, data for creating anatomy projection pairs may be collected from other physical simulations, research experiments on phantoms or volunteers, and / or clinical treatments on patients.
[0033] 7(a)-(f) are diagrams illustrating non-limiting examples of different projections and B1 projection data acquired along each of the projections, according to one embodiment of the present disclosure.
[0034] Typically, in B1 shimming, the goal is to solve for two fixed parameters, i.e., {A,φ}. This process is a low-resolution approach to the solution with fewer degrees of freedom. The B1 projection data along the selected direction exhibits characteristic signals in the projection data and contains sufficient spatially resolved information about the pattern of the B1 magnetic field. If information about the scale of the B1 magnetic field and its spatial distribution is known, it is possible to use the B1 projection data to calculate the B1 shimming parameters {A,φ}.
[0035] Various metrics or objectives are possible when selecting the shimming parameters, i.e., the selection of {A,φ}. One natural objective is to obtain the shimming parameters that create the smallest RF field amplitude inhomogeneity across a slice, as an integral across the slice. Another criterion can be to minimize the peak-to-peak variation across the slice. Additionally, other criteria are possible. Asymmetry of the B1 field variation may be of particular interest. The human body, such as the brain, has a remarkable left-right symmetry. This symmetry may aid the physician in detecting various clinical conditions. Due to pathology, some parts of the brain, for example, may appear brighter or darker in an MRI image than the corresponding anatomical region on the opposite side. Reducing other confounding technical sources of image asymmetry can make the task of diagnosis easier. Thus, even if the overall image appearance is slightly inhomogeneous due to RF transmit field inhomogeneity, the image may be pleasing to the clinician, provided that the remaining inhomogeneity largely preserves left-right symmetry.
[0036] When considering RF field asymmetry, it is possible to determine how close a single projection in the left-right direction is to symmetry, or to determine the asymmetry between two projections by reversing the B1 field values along one of the two directions, even if the two projections are located at different angular orientations, provided that they correspond to a straight line of left and right mirror pairs.
[0037] The choice of projection direction can be important for the analysis process. Typically, in human subjects, there is a principal axis in the B1 field distribution. Using methods such as principal component analysis (PCA), the optimal projection direction can be identified.
[0038] PCA is an algorithm that sorts and compares elements of variation in data. The following is an example of using PCA to select the direction of projection: A collection of B1+ map images is collected to cover a variety of human subjects and a variety of shimming parameter settings. Analysis is performed to highlight the most significant variations in the population and / or aspects of variation that can be best corrected by applying alternative shim value settings. Reformat each image to be similar in center, size, and intensity. Apply a common mask to primarily preserve overlapping pixels across multiple images. Convert each image to a long 1D feature vector. Add the 1D feature vectors (row-wise or column-wise) to a single rectangular array. Apply a Singular Value Decomposition (SVD) algorithm to identify singular vectors ordered by decreasing singular value. The SVD step is similar to eigenvalue-eigenvector decomposition, but can be performed on rectangular arrays. Extract the largest few 1D singular vectors. For each, reformat again from 1D feature vector to 2D image. Each image thus obtained becomes a principal component. If the maps are projected along their radial readout directions, examine the first few PCA images in the radial directions that showed the most global variations. These directions are the preferred and most accurate directions for obtaining the projections of the shimming.
[0039] 8(a)-(c) show non-limiting examples of different projections identified for different imaged anatomy. Each anatomy exhibits a unique B1 field pattern. The number of projections may be specific for each imaged anatomy. Different models may be selected and incorporated for each imaged anatomy to cover "corner cases" (e.g., metal implants, amputations / excisions, etc.).
[0040] The direction of the projections may also be specific for each imaging anatomy. The direction may be selected, for example, via PCA, to match the specific B1 field pattern of the anatomy. As illustrated in (a)-(c) of FIG. 8, for the head (8(a)), two orthogonal projections oriented at about 45° may be selected. For the torso (8(b)), two orthogonal projections or four projections separated by about 45° may be selected. For both the chest and thighs, a single horizontal projection may be used due to the simplicity of the B1 field pattern. It should be noted that the specific angles shown in (a)-(c) of FIG. 8 are not limiting. Other angular orientations are possible without departing from the scope of the present disclosure. For example, instead of the 45° shown in (a) of FIG. 8, angles of 40°, 60°, etc. may be used.
[0041] In a non-limiting embodiment, the minimum number of projections required for B1 field shimming is determined to save scan time. For example, for the head, two projections may be determined. If the Bloch-Siegert shift method is used, only four shots (two projections and two frequency offsets per projection) are required to obtain sufficient information. Using this projection-based shimming method, B1 shimming data can be obtained within one second. In another embodiment, many projections may be obtained to provide stability to the solution. In this respect, the directions and numbers of projections illustrated in (a) to (c) of FIG. 8 are not limiting of the present disclosure.
[0042] Several techniques for solving the final output of {A,φ} from the B1 projection data will now be described with reference to FIGS.
[0043] In one non-limiting embodiment, the conversion from B1 projection data to the final output of shimming parameters {A,φ} is performed by searching in a dictionary or look-up table. It is recognized that most patients share the same pattern of B1 field. For example, axial slices of the head are known to have a limited range of elliptical aspect ratios, e.g., ranging from a circle to an ellipse. For a given size and aspect ratio, the patterns are very similar.
[0044] Thus, for example, a group of B1 magnetic field distributions for different dimensional scales and aspect ratios may be predefined for each imaging anatomy. More elements may be added to take into account more dimensional features, such as body fat composition, male / female, age (e.g., child, adolescent, adult, elderly), etc. Such a group of B1 magnetic field distributions may be built offline by electromagnetic simulation using Sim4Life or a similar platform. Projection data for each of these B1 magnetic field distributions may be generated, and a solution for {A,φ} may be solved a priori for each B1 magnetic field distribution. For this purpose, a lookup table may be prepared, in which each entry contains a B1 magnetic field distribution, the physical features associated with that B1 magnetic field distribution, the projection data, and the solution for {A,φ}. A dictionary search may then be performed (by brute force search or other methods) in the lookup table to obtain a B1 magnetic field distribution with matching B1 projection data and matching physical features. For the matched B1 magnetic field distribution, the optimized {A,φ} is known.
[0045] FIG. 9 illustrates a non-limiting example of a block diagram of a shimming parameter analysis circuit that analyzes a set of shimming parameters based on acquired closely matched projection data according to an embodiment of the present disclosure. The shimming parameter analysis circuit 330 includes a physical feature extraction circuit 910, a projection distribution lookup table (LUT) 920, and a shimming parameter determination circuit 930. The physical feature extraction circuit 910 extracts one or more physical features from the patient information. The physical features may be, for example, a dimensional scale and / or an aspect ratio. Based on the projection data received from the projection data acquisition circuit 320 and the extracted physical features, a matching B1 magnetic field distribution is searched for in the projection distribution lookup table 920. The shimming parameter determination circuit 930 determines a set of shimming parameters based on the matched B1 magnetic field distribution and applies them to the RF transmitter.
[0046] FIG. 10 shows a non-limiting example of a flow chart of a process for analyzing a set of shimming parameters based on acquired projection data according to an embodiment of the present disclosure. On the left side of FIG. 10 is a process 1020 for analyzing shimming parameters using a projection distribution lookup table 920. In step 1022, projection data is received from the projection data acquisition circuitry 320. In step 1024, one or more physical characteristics (e.g., dimensional scale and / or aspect ratio, etc.) are extracted from the patient information. In step 1026, a search is performed in the pre-stored projection distribution lookup table 920 to obtain a B1 magnetic field distribution that matches the projection data and matches the physical characteristics. In step 1028, the set of shimming parameters associated with the acquired B1 magnetic field distribution is applied to the RF transmitter of the MRI device.
[0047] On the right side of FIG. 10 is a process 1060 for preparing the projection distribution lookup table 920 offline. In step 1062, electromagnetic simulations are performed to construct B1 magnetic field distributions for different physical features (including but not limited to dimensional scales and / or aspect ratios). In step 1064, projection data is generated for each B1 magnetic field distribution. In step 1066, a shimming parameter solution is solved for each B1 magnetic field distribution. In step 1068, the B1 magnetic field distributions are stored in the projection distribution lookup table 920, where each entry includes the B1 magnetic field distribution and the physical features, projection data, and shimming parameter solutions associated with the B1 magnetic field distribution. Although electromagnetic simulations are used to prepare the projection distribution lookup table 920 in FIG. 10, those skilled in the art will understand that other methods are possible. For example, data for creating the distribution entries may be collected from other physical simulations, research experiments on phantoms or volunteers, and / or clinical treatments on patients.
[0048] In an alternative embodiment, machine learning is used to map one set of information to another domain. In this case, the physical features and the measured spatial B1 projection data are input to a neural network. The output of the neural network is a set of optimal shimming parameters {A,φ}. Instead of searching for a solution for {A,φ} in a predefined dictionary or lookup table, the machine learning approach learns the mapping using a neural network and supervised learning. The neural network learns the transfer function in advance from training data including the B1 field distribution, the physical features, the projection data, and the set of shimming parameters {A,φ}. This approach has the added advantage of speed, since it is computationally faster to perform machine learning inference than a dictionary lookup.
[0049] FIG. 11 illustrates a non-limiting example of a block diagram of a shimming parameter analysis circuit that analyzes a set of shimming parameters based on acquired projection data according to an embodiment of the present disclosure. The shimming parameter analysis circuit 330 includes an individual projection data receiving circuit 1110, a cost function score calculation circuit 1120, and a shimming parameter determination circuit 1130. The shimming parameter determination circuit 1130 selects a set of different shimming parameters to be applied to the RF transmitter. The individual projection data receiving circuit 1110 acquires projection data from the projection data acquisition circuit 320 and outputs the individual projection data acquired along one or more directions with the different shimming parameter sets to the cost function score calculation circuit 1120. The cost function score calculation circuit 1120 calculates a cost function score for each of the different shimming parameter sets. The shimming parameter determination circuit 1130 identifies a set of shimming parameters with the lowest cost function score based on the cost function scores and outputs the set of shimming parameters to the RF transmitter of the MRI device.
[0050] FIG. 12 illustrates a non-limiting example of a flowchart of a process 1200 for analyzing a set of shimming parameters based on acquired projection data according to an embodiment of the present disclosure. At step 1210, projection data is acquired along one or more directions for different sets of shimming parameters. For example, projection data may be acquired along two directions (e.g., about 45° and about 135°, about 0° and about 90°, etc.) for different sets of shimming parameters (e.g., one set of parameters may be {A=0 dB, φ=50°} and another set of parameters may be {A=3 dB, φ=120°}). At step 1220, a cost function score is calculated for each set of different shimming parameters. As described below, the cost function score may be based on the L1 norm or the L2 norm, but other types of cost function scores may be used, for example. At step 1230, a set of shimming parameters with the lowest cost function score is identified based on the cost function scores calculated at step 1220. In step 1240, the identified set of shimming parameters is determined as the final set of shimming parameters to be applied to the RF transmitter.
[0051] 13(a)-(d) show non-limiting examples of obtaining profiles of projection data along two or more directions for different sets of shimming parameters according to an embodiment of the present disclosure. FIG. 13(a) shows two exemplary directions 1 and 2 with a first set of shimming parameters applied to the RF transmitter. FIG. 13(c) shows the profiles of projection data generated along the two directions 1 and 2. Similarly, FIG. 13(b) shows the two directions 1 and 2 with a second set of shimming parameters applied, and FIG. 13(d) shows the corresponding profiles of projection data obtained. Those skilled in the art will understand that the number of sets of shimming parameters and directions is not limited thereto. Other numbers of sets of shimming parameters and directions are possible without departing from the scope of the present disclosure.
[0052] A cost function score is calculated for each set of shimming parameters to measure the difference between the profiles of the projection data acquired along directions 1 and 2, according to equation (1) below.
[0053]
number
[0054] In the above formula, L=1 or 2, x means the index of pixels along the profile of the projection data, and Nx indicates the total number of those pixels. Alternatively, Nx indicates the number of pixels within a certain aggregation range. For example, the aggregation range may be a range where there is good data, a range where the signal magnitude is higher than some threshold or some noise level, or a range that is windowed or trimmed to exclude potentially poor quality data at both ends of the curve of the profile (e.g., to keep the central 80% of the data and discard 10% of the values at both ends), etc. When L=1, a score based on the L1 norm is obtained from the calculation of formula (1). The L1 norm is the sum of the absolute values of the differences between the profiles of the projection data along directions 1 and 2.
[0055] Here, the score based on the L1 norm is essentially an index of the symmetry of the B1 magnetic field. The difference between the profiles of the two projection data shown in FIG. 13(d) is smaller than that between the profiles of the two projection data shown in FIG. 13(c), so the score based on the L1 norm calculated for the second set of shimming parameters is smaller than the score based on the L1 norm calculated for the first set of shimming parameters. The smaller cost function score indicates that the second set of shimming parameters is better, since it provides a more constant B1 magnetic field distribution. Thus, it is possible to identify the best set of shimming parameters by comparing the cost function scores calculated for a number of different sets of shimming parameters.
[0056] To identify the best set of shimming parameters, it is also possible to evaluate the cost function score in other forms than the L1 norm based score. Figures 14(a)-(d) show non-limiting examples of acquiring profiles of projection data along one or more directions for different sets of shimming parameters according to an embodiment of the present disclosure. Figure 14(a) shows two exemplary directions 1A and 2A with a first set of shimming parameters (e.g., A=0 dB; φ=50°) applied to the RF transmitter. Figures 14(c) and 14(d) show corresponding profiles of projection data 1A and 2A acquired along the two directions, respectively. Similarly, Figure 14(b) shows two directions 1B and 2B with a second set of shimming parameters (A=3 dB; φ=120°) applied, and Figures 14(c) and 14(d) show corresponding acquired profiles of projection data 1B and 2B.
[0057] For each set of shimming parameters, a cost function score based on the L2 norm may be calculated according to equation (2) below.
[0058]
number
[0059] In the above formula, x and Nx have the same meaning as in formula (1), and B1 target shows a preferred B1 projection data profile obtained by any feasible means. The cost function score in this embodiment is based on the L2 norm, which represents the sum of squares of the difference between the projection data profile along one direction and the fixed B1 magnetic field data of interest. The projection data profiles obtained along one or both of the two directions shown in Fig. 14(a) and (b) may be used in the calculation of the L2 norm-based score for two different sets of shimming parameters. Again, by comparing the cost function scores calculated with the different sets of shimming parameters, it is possible to identify the best set of shimming parameters.
[0060] 11 and 12, it is possible to select two projections (e.g., as shown in Fig. 13(a)-(d)) or only one projection (e.g., as shown in Fig. 14(a)-(d)) to optimize the B1 shimming. Alternatively, more projections may be added to improve the robustness of the solution.
[0061] FIG. 15 illustrates a non-limiting example of a block diagram of a shimming parameter analysis circuit that analyzes a set of shimming parameters based on acquired projection data, according to an embodiment of the present disclosure. The shimming parameter analysis circuit 330 includes an individual projection data receiving circuit 1510, a projection profile symmetry analysis circuit 1520, and a shimming parameter determination circuit 1530. The shimming parameter determination circuit 1530 generates shimming parameters that are applied to the RF transmitter. In particular, the shimming parameter determination circuit 1530 can switch on one transmission channel of a plurality of transmission channels and switch off other transmission channels by applying appropriate shimming parameters. The individual projection data receiving circuit 1510 receives projection data acquired with different transmission channels switched on. The projection profile symmetry analysis circuit 1520 evaluates the effect of different transmission channels on the symmetry of the profile of the projection data. The shimming parameter determination circuit 1530 determines a set of final shimming parameters that maximizes the symmetry of the profile of the projection data based on the results of the evaluation.
[0062] FIG. 16 illustrates a non-limiting example flow chart of a process 1600 for analyzing a set of shimming parameters based on acquired projection data according to an embodiment of the present disclosure. The process begins at step 1610, where shimming parameters applied to the RF transmitter are controlled to operate each of the multiple transmit channels independently. At step 1620, projection data is acquired from the projection data acquisition circuitry 320. Individual projection data are acquired with one different transmit channel of the multiple transmit channels turned on and the rest turned off. At step 1630, the effect of each of the multiple transmit channels on the symmetry of the profile of the projection data is analyzed. At step 1640, a set of shimming parameters {A, φ} is solved. The set of shimming parameters maximizes the symmetry of the profile of the projection data. At step 1650, the solved set of shimming parameters {A, φ} is applied to the RF transmitter as a final set of shimming parameters.
[0063] Similar to the approach shown in Figures 11 and 12, this approach does not require pre-computed look-up tables or trained neural networks. Instead, this approach uses the symmetry of the projection data profile (or the cost function score calculated for each set of shimming parameters in the approach shown in Figures 11 and 12) as a surrogate measure of the homogeneity of the B1 field distribution. By evaluating the effect of each transmit channel on the symmetry of the B1 data in the profile, the best set of {A,φ} can be solved to maximize the symmetry.
[0064] Preferably, the direction of the projection is selected along a characteristic direction of the B1 field distribution. More preferably, the direction of the projection is selected along a characteristic direction that exhibits the least symmetry, so as to add subtlety to the method. Using this method, it is possible to select only one projection. Alternatively, more projections may be added to improve the robustness of the solution.
[0065] It should be noted that in the techniques shown in Figures 15 and 16, the choice of the direction of projection may be important. If the "hot" and "cold" spots (+ / - regions) overlap in the B1 field distribution, it becomes difficult to differentiate the effects of B1 shimming. Taking the projections shown in Figures 7(a)-(c) as an example, Figure 7(a) is a good choice because the + / - regions in the projection are almost not overlapped (i.e., the amplitudes of the B1 fields are combined along rays perpendicular to the readout direction of the projection). The task of this combination to form the projection is similar to taking the average value, and the amplitude of the projection data does not increase only with the length of the rays through the object. Figure 7(b) is also a good choice because the + / - regions in the projection are almost not overlapped. However, Figure 7(c) is not a good choice because the + / - regions overlap more than Figures 7(a) and 7(b).
[0066] In all the above-mentioned embodiments, any B1 measurement method can be used, including but not limited to the Double Angle Method (DAM), Actual Flip Angle Imaging (AFI), Dual Refocusing Echo Acquisition Mode (DREAM), Phase Sensitive (PS), Saturation Recovery, and Bloch-Siegert Shift. The method presented in this embodiment will perform best with a B1 measurement method that is not affected by tissue composition and relaxation bias. If different tissue types are mixed in the B1 data, the projection data will contain errors because all tissues are substantially integrated along the projection. Therefore, the Bloch-Siegert Shift method is preferred because it is generally not significantly affected by a wide range of tissue compositions.
[0067] Preferably, a region of interest may be selected within the projection. For example, only the central 80% of the projection is considered in the solution to avoid subcutaneous fat. Conventional Cartesian B1 mapping methods can typically be supported with a region of interest centered at any location within the volume. With a fully encoded map, regions of interest of any shape, size and location are possible. In contrast, the projection-based approaches described in this disclosure generally have limited flexibility in the selection of the location of the region of interest due to the nature of the projection. The B1 intensity information along the vertical line is inherently combined in a basic one-dimensional readout. Without the projection incorporating additional features into the shimming collection, there would be no way to collect data along only some limited portion of the vertical ray. In other words, data in a dimension spatially orthogonal to the dimension of the projection readout is essentially averaged when measuring the projection. Therefore, the location of this orthogonal dimension is not analyzed, and as a result, the boundary of the region of interest in that dimension is not analyzed.
[0068] Furthermore, a portion of the volume may be selected. For example, an internal volume technique (e.g., 2D spatial RF selection or "pencil beam" technique) using RF slice selection along different gradient axes may be used. Pencil beam techniques (sometimes called "internal volume selection") use two or more slice selection RF pulses with fully or partially orthogonal RF selection gradient directions to excite MR signals only from the intersecting geometrical region selected by the RF pulses. This is an effective way to generate signals only along a rectangular row and then apply a readout direction along the length of the row.
[0069] In addition, more than one projection may be acquired per repetition time (TR) using multiple readouts per excitation ("bowtie" technique). In this "bowtie" readout technique, following RF excitation, a pattern of RF pulses and gradient waveforms in a pulse sequence is designed to measure two or more echo signals along different spatial trajectories. Specifically, after excitation of a slice, a conventional readout may collect data along a line through the center of k-space at several angles in k-space. After the data for that line is collected, additional gradient lobes in several other directions may move the net k-space encoding to a position farthest from the center of k-space at several other angular orientations. Another readout gradient and data sampling may then trace through the center of k-space along the second angular orientation to a geometrically opposite position in k-space. This configuration is sometimes referred to as a "bowtie readout technique" because it uses two or more readout directions at different angles that intersect at the center of k-space.
[0070] Using the techniques described in this disclosure, much shorter pre-scan times can be achieved compared to conventional Cartesian B1 maps. If Bloch-Siegert shifting is used and the repetition time (TR) is 200 ms, B1 shimming for the chest or thigh takes 0.4 seconds (1 projection x 2 frequencies), B1 shimming for the head takes 0.8 seconds (2 projections x 2 frequencies), and B1 shimming for the torso takes 1.6 seconds (4 projections x 2 frequencies). As a result, B1 projection data can be acquired in about 1 second compared to about 26 seconds for the lowest resolution Cartesian B1 map (32 encodings x 2 frequencies x 2 channels if a separate map is measured for each channel). Also, it is less expensive to select more projections to add more stability to the solution, if necessary. The added cost is about an increase of about 1 second.
[0071] Furthermore, the techniques described in this disclosure may eliminate the need for breath-holding, since each measurement is a simple projection through the center of the volume, following the same principles as free-breathing radial acquisitions. Also, the scan time of shimmed acquisition projections is generally much shorter, reducing the chance of data being unstable or corrupted by patient motion. The amount of patient motion in a window of about one second is insignificant compared to the total amount of motion that routinely occurs over tens of seconds.
[0072] Although some computation time is required to convert the B1 projection data into an {A,φ} solution, both the lookup table approach and the machine learning approach require a finite amount of computation time to compute the {A,φ} solution. This computation time can be hidden during other pre-scans (e.g., coil maps). Traditional B1 shimming solutions require similar computation time to solve each individual channel of the B1 map input.
[0073] In general, the techniques described in this disclosure provide significant scan time savings compared to conventional methods, and do not require acquisition of a full B1 map to calculate a simple two-parameter setting for B1 shimming. By measuring several projections along characteristic directions, sufficient information can be obtained from the projection data to determine the B1 shimming parameters {A,φ}.
[0074] Referring now to FIG. 17, a non-limiting example of a magnetic resonance imaging (MRI) device 100 is shown. The MRI device 100 shown in FIG. 17 includes a gantry 101 (shown in a schematic cross-sectional view) and various associated MRI system components 103 interconnected via interfaces. Typically, at least the gantry 101 is installed in a shielded room. The geometric configuration of the MRI device 100 shown in FIG. 17 includes a substantially coaxial cylindrical configuration of a static magnetic field B0 magnet 111, a Gx, Gy and Gz gradient coil set 113, and a large Whole-Body RF Coil (WBC) 115. An imaging volume 117 is shown along the horizontal axis of each element included in this cylindrical arrangement to substantially surround the head of a patient 119 supported by a patient table 120.
[0075] Within the imaging volume 117, one or more small array RF coils 121 may be connected to be closer to the patient's head (referred to herein, for example, as a "scan target" or "target"). As will be appreciated by those skilled in the art, coils and / or arrays that are relatively small compared to WBC (whole body coils), such as surface coils, are often customized for specific parts of the body (e.g., arms, shoulders, elbows, wrists, knees, feet, chest, spine, etc.). Such small RF coils are referred to herein as array coils (AC) or phased-array coils (PAC). These may include at least one coil configured to transmit RF signals to the imaging volume 117 and multiple receiver coils configured to receive RF signals from a target, such as the patient's head, within the imaging volume 117.
[0076] The MRI apparatus 100 includes an MRI system control 130 having input / output ports connected to a display 124, a keyboard 126, and a printer 128. As will be appreciated by those skilled in the art, the display 124 may be of the touch screen type to also provide control input, and a mouse or other I / O device may also be provided.
[0077] The MRI system control 130 interfaces with the MRI sequence control 140, which in turn controls the Gx, Gy, and Gz gradient coil drivers 132, as well as the RF transmitter 134 and the transmit / receive switch 136 (if the same RF coil is used for both transmit and receive). The RF transmitter 134 may be configured with two or more transmit channels to drive two or more RF transmit coils or ports on the coils, such as those used for RF shimming. The MRI sequence control 140 includes a suitable program code structure 138 for performing techniques of MRI imaging (also known as nuclear magnetic resonance imaging, or NMR imaging), including B1 field shimming. The MRI sequence control 140 may be configured for MR imaging with or without parallel imaging. In addition, the MRI sequence control 140 may perform one or more prescan and scan sequences for acquiring the main scan Magnetic Resonance (MR) images (called diagnostic images). The MR data obtained by the pre-scan may be used, for example, to determine shimming parameters of the whole-body RF coil 115 and / or the array RF coil 121 .
[0078] The MRI system components 103 include an RF receiver 141 that provides input to an MRI data processor 142 to produce processed image data that is transmitted to the display 124. The MRI data processor 142 is configured to access previously generated MR data, images, and / or projection data, such as projection data acquired with different perturbations of shimming parameters, projection data acquired with different transmit channels, and / or system configuration parameters 146, as well as program code structures 144 and 150.
[0079] In one embodiment, the MRI data processor 142 includes processing circuitry that may include devices such as Application-Specific Integrated Circuits (ASICs), programmable logic devices (e.g., Simple Programmable Logic Devices (SPLDs), Complex Programmable Logic Devices (CPLDs), and Field Programmable Gate Arrays (FPGAs)), and other circuit components configured to perform the functions described in this disclosure, such as those described with respect to Figures 3, 5, 9, 11, and 15.
[0080] The MRI data processor 142 executes one or more sequences of one or more instructions contained in the program code structures 144 and 150. Alternatively, the instructions may be read from another computer-readable medium, such as a hard disk or a removable media drive. One or more processors in a multi-processing arrangement may be used to execute the sequences of instructions contained in the program code structures 144 and 150. In alternative embodiments, hardwired circuitry may be used in place of or in combination with software instructions. Thus, embodiments of the present disclosure are not limited to any particular combination of hardware circuitry and software. For example, the program code structure 150 stores instructions for executing the process 400.
[0081] Furthermore, as used herein, the term "computer readable medium" refers to any non-transitory medium that provides instructions to the MRI data processor 142 for execution. Computer readable media can be in many forms, including but not limited to non-volatile or volatile media. Non-volatile media include, for example, optical, magnetic, and magneto-optical disks, or removable media drives. Additionally, volatile media includes dynamic memory.
[0082] 17 shows a generalized MRI system program storage (memory) 150 in which stored program code structures, such as instructions for executing process 400, are stored on a non-transitory computer readable storage medium callable by various data processing components of the MRI apparatus 100. As will be appreciated by those skilled in the art, the MRI system program storage 150 may be partitioned and configured, at least in part, to be directly connected to different processing computers within the MRI system components 103 that have the most immediate need for the stored program code structures during normal operation (i.e., rather than being shared and directly connected to the MRI system controller 130).
[0083] Furthermore, the MRI device 100 shown in Fig. 17 can be utilized to implement the exemplary embodiments described herein above. The MRI system components 103 may be divided into different logical collections of "boxes" and may typically include a number of Digital Signal Processors (DSPs), microprocessors, and special purpose processing circuits (e.g., processing circuits for high speed A / D (Analog to Digital) conversion, Fast Fourier Transform, array processing). Each of these processing circuits is typically a clocked "state machine" in which the physical data processing circuitry progresses from one physical state to another physical state every clock cycle (or a predetermined number of clock cycles).
[0084] Moreover, not only do the physical states of processing circuits (e.g., CPUs, registers, buffers, arithmetic units, etc.) change progressively from one clock cycle to another during the course of operation, but also the associated physical states of data storage media (e.g., bit stores in magnetic storage media) change from one state to another during operation of such systems. For example, at the end of an image reconstruction process and / or sometimes an image reconstruction map (e.g., coil sensitivity map, unfolding map, ghost map, distortion map, etc.) generation process, an array of computer-readable and accessible data value stores contained in a physical storage medium changes from some previous state to a new state, where the physical states of the physical portions of such arrays change between minimum and maximum values to represent real-world physical events and conditions. As will be appreciated by those skilled in the art, such arrays of stored data values represent and constitute physical structures, in the same way that a particular structure of computer control program code, when sequentially loaded into instruction registers and executed by one or more CPUs of the MRI device 100, changes its state to generate a particular sequence of states operable within the MRI device 100.
[0085] Numerous modifications and variations are possible in light of the above description, and therefore, it will be understood that within the scope of the appended claims, the presently disclosed technology may be practiced otherwise than as specifically described herein.
[0086] The embodiments of the present disclosure can also be additionally described as follows.
[0087] (Appendix 1) A shimming method for performing patient-specific B1 magnetic field shimming in a magnetic resonance imaging apparatus, comprising: acquiring patient information of a patient to be imaged by the magnetic resonance imaging apparatus; determining a direction of projection based on the acquired patient information; acquiring B1 projection data along the determined projection direction using the magnetic resonance imaging device; determining a set of B1 shimming parameters based on the acquired B1 projection data; A shimming method comprising:
[0088] (Appendix 2) and controlling the magnetic resonance imaging apparatus based on the determined set of B1 shimming parameters. The shimming method described in (Appendix 1).
[0089] (Appendix 3) The step of determining a direction of projection comprises: extracting specific imaging anatomy from the patient information; obtaining the direction of the projection from a first lookup table using the extracted imaged anatomy as a key; Further comprising: The step of acquiring B1 projection data includes: generating magnetic field gradient signals based on the orientation of the acquired projections; applying the generated gradient magnetic field signals to a gradient magnetic field coil driver of the magnetic resonance imaging device; Further comprising: The shimming method described in (Appendix 1).
[0090] (Appendix 4) determining respective B1 magnetic field distributions corresponding to different specific imaging anatomy; performing a principal component analysis on each of the determined B1 field distributions; identifying a characteristic projection of each of the determined B1 field distributions based on results of the principal component analysis; storing the identified characteristic projections and the corresponding imaged anatomy as matched pairs in the first lookup table; Further comprising: The shimming method described in (Appendix 3).
[0091] (Appendix 5) determining the B1 magnetic field distribution further comprises determining the B1 magnetic field distribution based on physical simulations, phantom experiments, in vivo experiments, and / or data collected during a clinical procedure performed on a patient; The shimming method described in (Appendix 4).
[0092] (Appendix 6) The step of determining the set of B1 shimming parameters comprises: extracting specific physical features from the patient information; obtaining a B1 magnetic field distribution that matches the extracted physical feature and the acquired B1 projection data from a second lookup table that associates a B1 magnetic field distribution, a specific physical feature associated with the B1 magnetic field distribution, and B1 projection data associated with the B1 magnetic field distribution; determining a set of specific B1 shimming parameters associated with the acquired B1 magnetic field distribution as the determined set of B1 shimming parameters; Further comprising: The shimming method described in (Appendix 3).
[0093] (Appendix 7) determining B1 magnetic field distributions for different corresponding physical features respectively; generating corresponding projection data for each of the determined B1 magnetic field distributions; determining, for each of the determined B1 magnetic field distributions, a corresponding set of B1 shimming parameters; storing each of the determined B1 magnetic field distributions in the second look-up table in association with the corresponding physical characteristic, the corresponding projection data, and the corresponding set of B1 shimming parameters; Further comprising: The shimming method described in (Appendix 6).
[0094] (Appendix 8) the extracted physical features include at least one of a dimensional scale of the imaged anatomy, an aspect ratio of the imaged anatomy, a body fat composition of the patient, a gender of the patient, and an age of the patient; The shimming method described in (Appendix 6).
[0095] (Appendix 9) The step of determining the set of B1 shimming parameters comprises: extracting specific physical features from the patient information; applying the extracted physical features and the acquired projection data to a trained neural network; determining the set of B1 shimming parameters from an output of the neural network; Further comprising: The shimming method described in (Appendix 3).
[0096] (Appendix 10) determining B1 magnetic field distributions for different corresponding physical features respectively; generating corresponding projection data for each of the determined B1 magnetic field distributions; determining for each of said determined B1 magnetic field distributions a corresponding set of B1 shimming parameters; training the neural network using the determined B1 magnetic field distribution, the corresponding physical characteristics, the corresponding projection data, and the corresponding set of B1 shimming parameters as training data; Further comprising: The shimming method described in (Appendix 9).
[0097] (Appendix 11) The step of determining the set of B1 shimming parameters comprises: receiving B1 projection data corresponding to each of a plurality of sets of B1 shimming parameters separately along the determined direction of projection; calculating a cost function score corresponding to each of the plurality of sets of B1 shimming parameters separately based on the received B1 projection data; identifying a particular set of B1 shimming parameters that corresponds to the lowest cost function score; determining the identified set of specific B1 shimming parameters as the determined set of B1 shimming parameters; Further comprising: The shimming method described in (Appendix 1).
[0098] (Appendix 12) determining a direction of projection further comprises determining a first direction of projection and a second direction of projection; The step of receiving B1 projection data includes: receiving first B1 projection data corresponding to each of the plurality of sets of B1 shimming parameters along a first direction of projection separately; receiving second B1 projection data corresponding to each of the plurality of sets of B1 shimming parameters along a second direction of projection separately; Further comprising: calculating the cost function score further comprises calculating, for each of the plurality of sets of B1 shimming parameters, the cost function score based on the received first B1 projection data related to the first direction of projection and the received second B1 projection data related to the second direction of projection. The shimming method described in (Appendix 11).
[0099] (Appendix 13) and calculating the cost function score further comprises calculating the cost function score for each of the plurality of sets of B1 shimming parameters based on the received B1 projection data and target B1 projection data. The shimming method described in (Appendix 11).
[0100] (Appendix 14) The step of determining the set of B1 shimming parameters comprises: controlling a set of B1 shimming parameters to switch on each individual transmit channel of an RF transmitter of the magnetic resonance imaging apparatus and to switch off other transmit channels; receiving specific B1 projection data acquired in each of the switched-on individual transmit channels separately; analyzing the received particular B1 projection data to assess the effect of each of the individual transmit channels on a symmetry of a profile of the B1 projection data; calculating a particular set of B1 shimming parameters that maximizes the symmetry of a profile of the B1 projection data; determining the calculated set of specific B1 shimming parameters as the determined set of B1 shimming parameters; Further comprising: The shimming method described in (Appendix 1).
[0101] (Appendix 15) determining the direction of projection further comprises determining the direction of projection such that a profile of B1 projection data acquired along the direction of projection has minimal symmetry; The shimming method described in (Appendix 14).
[0102] (Appendix 16) the set of B1 shimming parameters includes RF amplitude modulation and RF phase modulation applied to an RF transmission channel of an RF transmitter of the magnetic resonance imaging apparatus; The shimming method described in (Appendix 1).
[0103] (Appendix 17) The step of acquiring the B1 projection data further includes acquiring the B1 projection data using a Bloch-Siegert shift method, a double angle method, a substantially flip angle method, a dual refocusing echo acquisition mode method, a phase sensitive method, or a saturation recovery method. The shimming method described in (Appendix 1).
[0104] (Appendix 18) and wherein the acquiring step further comprises performing a 2D spatial selection to select a portion within the patient volume along the determined direction of the projection. The shimming method described in (Appendix 1).
[0105] (Appendix 19) The step of acquiring the B1 projection data further includes performing two or more readouts per excitation to acquire two or more projections per repetition time. The shimming method described in (Appendix 1).
[0106] (Appendix 20) A shimming device for performing patient-specific B1 magnetic field shimming in a magnetic resonance imaging device, comprising: acquiring patient information of a patient to be imaged by the magnetic resonance imaging apparatus; determining a direction of projection based on the acquired patient information; acquiring B1 projection data along the determined projection direction using the magnetic resonance imaging apparatus; determining a set of B1 shimming parameters based on the acquired B1 projection data; A shimming apparatus comprising a processing circuit.
[0107] Those skilled in the art will also understand that there may be many variations made to the operation of the techniques described above while still achieving the same objectives of the present disclosure. Such variations are intended to be included within the scope of the present disclosure. Thus, the above description of the embodiments of the present disclosure is not intended to be limiting. Rather, any limitations to the embodiments of the present disclosure are set forth in the following claims.
[0108] According to at least one of the embodiments described above, the scan time of the prescan can be reduced.
[0109] Although some embodiments have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, substitutions, modifications, and combinations of embodiments can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope of the invention and its equivalents as described in the claims, as well as in the scope and spirit of the invention. [Explanation of symbols]
[0110] 100 Magnetic resonance imaging device 142 MRI Data Processor
Claims
1. 1. A shimming method for performing patient-specific B1 field shimming in a magnetic resonance imaging apparatus, comprising: acquiring patient information of a patient to be imaged by the magnetic resonance imaging apparatus; determining a direction of projection based on the acquired patient information; acquiring B1 projection data along the determined projection direction using the magnetic resonance imaging device; determining a set of B1 shimming parameters based on the acquired B1 projection data; A shimming method comprising:
2. and controlling the magnetic resonance imaging device based on the determined set of B1 shimming parameters. The shimming method of claim 1 .
3. The step of determining a direction of projection comprises: extracting specific imaging anatomy from the patient information; using the extracted imaged anatomy as a key to obtain the direction of the projection from a first lookup table; Further comprising: The step of acquiring B1 projection data includes: generating magnetic field gradient signals based on the orientation of the acquired projections; applying the generated gradient magnetic field signals to a gradient magnetic field coil driver of the magnetic resonance imaging device; Further comprising: The shimming method of claim 1 .
4. determining respective B1 magnetic field distributions corresponding to different particular imaging anatomy; performing a principal component analysis on each of the determined B1 field distributions; identifying a characteristic projection of each of the determined B1 field distributions based on results of the principal component analysis; storing the identified characteristic projections and the corresponding imaged anatomy as matched pairs in the first lookup table; Further comprising: The shimming method according to claim 3.
5. determining the B1 magnetic field distribution further comprises determining the B1 magnetic field distribution based on physical simulations, phantom experiments, in vivo experiments, and / or data collected during a clinical procedure performed on a patient; The shimming method according to claim 4.
6. The step of determining a set of B1 shimming parameters comprises: extracting specific physical features from the patient information; obtaining a B1 magnetic field distribution that matches the extracted physical feature and the acquired B1 projection data from a second lookup table that associates a B1 magnetic field distribution, a specific physical feature associated with the B1 magnetic field distribution, and B1 projection data associated with the B1 magnetic field distribution; determining a set of specific B1 shimming parameters associated with the acquired B1 magnetic field distribution as the determined set of B1 shimming parameters; Further comprising: The shimming method according to claim 3.
7. determining respective B1 magnetic field distributions for different corresponding physical features; generating corresponding projection data for each of the determined B1 magnetic field distributions; determining, for each of the determined B1 magnetic field distributions, a corresponding set of B1 shimming parameters; storing each of the determined B1 magnetic field distributions in the second look-up table in association with the corresponding physical characteristic, the corresponding projection data, and the corresponding set of B1 shimming parameters; Further comprising: The shimming method according to claim 6.
8. the extracted physical features include at least one of a dimensional scale of the imaged anatomy, an aspect ratio of the imaged anatomy, a body fat composition of the patient, a gender of the patient, and an age of the patient; The shimming method according to claim 6.
9. The step of determining a set of B1 shimming parameters comprises: extracting specific physical features from the patient information; applying the extracted physical features and the acquired projection data to a trained neural network; determining the set of B1 shimming parameters from an output of the neural network; Further comprising: The shimming method according to claim 3.
10. determining respective B1 magnetic field distributions for different corresponding physical features; generating corresponding projection data for each of the determined B1 magnetic field distributions; determining, for each of the determined B1 magnetic field distributions, a corresponding set of B1 shimming parameters; training the neural network using the determined B1 magnetic field distribution, the corresponding physical characteristics, the corresponding projection data, and the corresponding set of B1 shimming parameters as training data; Further comprising: The shimming method of claim 9.
11. The step of determining a set of B1 shimming parameters comprises: receiving B1 projection data corresponding to each of a plurality of sets of B1 shimming parameters separately along the determined direction of projection; calculating a cost function score corresponding to each of the plurality of sets of B1 shimming parameters separately based on the received B1 projection data; Identifying a particular set of B1 shimming parameters that corresponds to the lowest cost function score; determining the identified set of specific B1-shimming parameters as the determined set of B1-shimming parameters; Further comprising: The shimming method of claim 1 .
12. determining a direction of projection further comprises determining a first direction of projection and a second direction of projection; The step of receiving B1 projection data includes: receiving first B1 projection data corresponding to each of the plurality of sets of B1 shimming parameters along a first direction of projection separately; receiving second B1 projection data corresponding to each of the plurality of sets of B1 shimming parameters along a second direction of projection separately; Further comprising: calculating the cost function score further comprises calculating, for each of the plurality of sets of B1 shimming parameters, the cost function score based on the received first B1 projection data related to the first direction of projection and the received second B1 projection data related to the second direction of projection. The method of claim 11.
13. and calculating the cost function score further comprises calculating the cost function score for each of the plurality of sets of B1 shimming parameters based on the received B1 projection data and target B1 projection data. The shimming method of claim 11.
14. The step of determining a set of B1 shimming parameters comprises: controlling a set of B1 shimming parameters to switch on each individual transmit channel of an RF transmitter of the magnetic resonance imaging apparatus and to switch off other transmit channels; receiving specific B1 projection data acquired on each of the switched-on individual transmit channels separately; analyzing the particular B1 projection data received to assess the effect of each of the individual transmit channels on a symmetry of a profile of the B1 projection data; calculating a particular set of B1 shimming parameters that maximizes the symmetry of a profile of the B1 projection data; determining the calculated set of specific B1-shimming parameters as the determined set of B1-shimming parameters; Further comprising: The shimming method of claim 1 .
15. determining the direction of projection further comprises determining the direction of projection such that a profile of B1 projection data acquired along the direction of projection has minimal symmetry; The method of claim 14.
16. the set of B1 shimming parameters includes RF amplitude modulation and RF phase modulation applied to an RF transmission channel of an RF transmitter of the magnetic resonance imaging apparatus; The shimming method of claim 1 .
17. The step of acquiring the B1 projection data further includes acquiring the B1 projection data using a Bloch-Siegert shift method, a double angle method, a substantially flip angle method, a dual refocusing echo acquisition mode method, a phase sensitive method, or a saturation recovery method. The shimming method of claim 1 .
18. and the acquiring step further comprises performing a 2D spatial selection to select a portion within the patient volume along the determined direction of the projection. The shimming method of claim 1 .
19. The step of acquiring the B1 projection data further includes performing two or more readouts per excitation to acquire two or more projections per repetition time. The shimming method of claim 1 .
20. 1. A shimming apparatus for performing patient-specific B1 magnetic field shimming in a magnetic resonance imaging apparatus, comprising: acquiring patient information of a patient to be imaged by the magnetic resonance imaging apparatus; determining a direction of projection based on the acquired patient information; acquiring B1 projection data along the determined projection direction using the magnetic resonance imaging device; determining a set of B1 shimming parameters based on the acquired B1 projection data; A shimming apparatus comprising a processing circuit.
Citation Information
Patent Citations
Wireless local transmit coils and array with controllable load
EP2734855A2
Magnetic resonance imaging apparatus and magnetic resonance imaging method
US10126395B2
Optical image-based position tracking for magnetic resonance imaging applications
US20050054910A1
Method and system for b1 field mapping in magnetic resonance imaging
US20130134972A1
US2021/10161382
