Cardiac imaging

Abstract

This invention relates to cardiac imaging, specifically a method for generating MR images of a patient at an MR imaging system. The method includes sampling a k-space using a 3D k-space acquisition scheme having two phase-encoding directions and one readout direction. The sampling scheme includes applying a sequence of inversion pulses to the patient to invert the patient's magnetization, wherein sampling of the k-space is performed using uniform undersampling of the k-space and n distinct EPI readout excitations along the readout direction after each inversion pulse, where n => 5. An MR image is reconstructed based on the sampled k-space.

Description

Technical Field

[0001] This invention relates to a method for generating MR images of a patient using a 3D acquisition scheme in k-space, a magnetic resonance imaging (MRI) system, a computer program including program code, and a carrier including the computer program. The invention particularly relates to cardiac imaging. Background Technology

[0002] Cardiac imaging using MRI has become a valuable diagnostic tool for assessing cardiovascular disease. MRI provides high-resolution images of the heart and surrounding structures, allowing for detailed evaluation of cardiac anatomy, function, and pathology. Cardiac MRI imaging offers high spatial resolution as well as excellent soft tissue contrast and functional assessment. Late gadolinium enhancement (LGE) MRI is a key component of cardiac MRI examinations. It is used for the detection and classification of myocardial infarction. However, 3D LGE acquisition can be time-consuming, which may be difficult for some patients to tolerate. A recent proposal to address this issue is to generate 3D LGE images in a single breath-hold using compression sensing (CS), as discussed by Roman Johannes Gertz et al. in "Compressed SENSE accelerated 3D single-breathhold late gadolinium enhancement cardiovascular magnetic resonance with isotropic resolution: clinical evaluation" published in *Frontiers in Cardiovascular Medicine* (November 2023, Vol. 10). This is achieved by acquiring highly undersampled images and reducing scan time. However, CS reconstruction may suffer from residual blur artifacts, especially at higher undersampling factors.

[0003] Therefore, it is necessary to avoid the above-mentioned drawbacks and provide an imaging method with reduced artifacts and acceptable image acquisition time. Summary of the Invention

[0004] This need is satisfied by a feature of one aspect of the invention. Other aspects are described in some embodiments of the invention.

[0005] According to a first aspect, a method for generating MR images of a patient at an MR imaging system is provided. The method includes sampling a k-space using a three-dimensional acquisition scheme having two phase-encoding directions and one readout direction. The sampling includes applying a sequence of inversion pulses to the patient to invert the patient's magnetization, and performing sampling of the k-space using uniform undersampling of the k-space and n distinct EPI readout excitations along the readout direction after each inversion pulse, where n => 5. The MR image is then reconstructed based on the sampled k-space.

[0006] The combination of uniform undersampling, inverted pulses, and EPI readouts after each inverted pulse provides highly efficient sampling with reduced acquisition time and acceptable image quality.

[0007] Preferably, the MR images are cardiac MR images, and a three-dimensional acquisition scheme is used in conjunction with cardiac gating, wherein n readout excitations are acquired within a single heartbeat. This imaging method provides particularly good results in cardiac imaging and allows MR images to be acquired within a single breath-hold.

[0008] Preferably, k-space sampling is performed after the contrast agent has been applied to the patient and the concentration of the contrast agent has reached equilibrium in the patient. This is also known as late contrast-enhanced imaging or late gadolinium-enhanced LGE imaging, which is particularly suitable for the detection of specific lesions in the heart, such as the detection and classification of myocardial infarction. The proposed image acquisition scheme is particularly useful in this application area.

[0009] Preferably, the k-space is divided into n regions in one direction of the phase encoding direction, and an EPI readout excitation from n excitations is acquired from each region within one heartbeat. As a result, uniform undersampling is obtained, which facilitates image reconstruction.

[0010] Alternatively, it is possible to acquire S excitations in each of n regions in k-space, where S ≥ 3, and using the CAIPIRINHA imaging technique, offset the k-space trajectories of the excitations within a region relative to each other in one phase-encoding direction of the phase-encoding direction. Here, S is preferably in the range of 6 to 10. Utilizing this image acquisition and reconstruction technique, a 3D whole-heart volume can be acquired with satisfactory image quality and reduced artifacts during a single breath-hold. Preferably, the reconstructed MR images are performed using a k-space interpolation-based technique, which includes GRAPPA (Generalized Automatic Calibration Partial Parallel Acquisition) using a 3D calibration kernel.

[0011] Furthermore, the n EPI readout excitations can be distributed in k-space, such that the center of k-space is sampled using earlier excitations from the n excitations, and the region outside the center is sampled using later excitations from the n excitations. The center of k-space determines the image contrast, as it is acquired earlier at locations with greater magnetization and therefore stronger MR signals. Thus, image contrast is improved when the center of k-space is sampled earlier than the outer portions of k-space.

[0012] Preferably, after each inversion pulse, one of the n excitations is acquired in each of the n regions, and n is in the range of 5 to 10, more preferably in the range of 6 to 8.

[0013] Preferably, the inversion time is selected based on the MR signal of the myocardium, and more preferably, the inversion time is selected such that the MR signal of the myocardium is close to zero at the time of image acquisition, thereby obtaining good contrast in the image of the area surrounding the myocardium.

[0014] In addition, an MR imaging system is provided, which includes a processing unit and a memory configured to store program code that, when executed by the processing unit, causes the processing unit to perform the image acquisition steps identified above.

[0015] In addition, a computer program including program code is provided for execution by the processing unit of the MR imaging system, wherein the execution of the program code causes the processing unit to perform the methods described above or further detailed below.

[0016] Finally, a carrier including a computer program is provided, wherein the carrier is one of electrical signals, optical signals, radio signals, and computer-readable storage media.

[0017] It should be understood that, without departing from the scope of the invention, the features described above and those explained below can be used not only in the corresponding combinations shown, but also in other combinations or individually. Unless otherwise expressly stated, the features of the above aspects and the embodiments described below can be combined with each other in other embodiments. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of an MIR imaging system, which can be used to acquire 3D whole heart volume in a single breath-hold.

[0019] Figure 2 A schematic diagram of a sampling scheme for acquiring a 3D volume of the heart is shown.

[0020] Figure 3 It shows in Figure 2 A schematic diagram of the sampling mode used to sample the k-space during the sampling scheme.

[0021] Figure 4 An example schematic diagram of a flowchart is shown, illustrating the steps performed during image acquisition or k-space sampling. Detailed Implementation

[0022] In the following, embodiments of the present invention will be described in detail with reference to the accompanying drawings. It should be understood that the following description of the embodiments should not be considered limiting. The scope of the present invention is not limited to the embodiments or drawings described below, which are merely illustrative.

[0023] The accompanying drawings should be considered schematic representations, and the elements shown are not necessarily shown to scale. Rather, the various elements are shown such that their function and general purpose will become apparent to those skilled in the art. Any connection or coupling between functional blocks, devices, or components of the physical or functional units shown in the drawings and described below may be achieved through indirect connection or coupling. Coupling between components may be established through wired or wireless connections. Functional blocks may be implemented using hardware, software, firmware, or a combination thereof.

[0024] The following section will discuss an MR imaging method that provides a sufficiently short dataset of 3D images of the heart so that it can be acquired in a single breath-hold without suffering from artificial image appearance and residual blur.

[0025] Figure 1 A schematic diagram of an MRI (Magnetic Resonance Imaging) system 40 is shown, which includes a magnet 30 that generates a polarization field B0. A subject 33 lying on an examination table 32 is moved to the center of the MRI system 40, where an MR signal following RF excitation can be detected by a receiving coil 31, which may include different coil segments, each associated with a corresponding detection channel. By applying RF pulses and a magnetic field gradient, the nuclear spins in the subject 33 (particularly the portion located in the receiving coil 31) are excited, and the current induced by relaxation is detected. How MR images are generated and how MR signals are detected using RF pulse sequences and magnetic field gradient sequences are known in the art, and therefore detailed descriptions are omitted.

[0026] The MRI system includes a control unit 20 for controlling the MRI system. The control unit 20 includes a gradient control unit 24 for controlling and switching magnetic field gradients, and an RF control unit 25 for controlling and generating RF pulses for imaging sequences. An image sequence control unit 26 is provided, which controls the applied magnetic field gradient and the sequence of RF pulses, thereby controlling the gradient control unit 24 and the RF control unit 25. The computer program required to operate the MRI system and the imaging sequences required to generate MR images can be stored in a memory 27 along with the generated MR images. The generated MR images can be displayed on a display 28, where an input unit 29 is provided for the user of the MRI system to control the functions of the MRI system. A processing unit 23 can coordinate... Figure 1 The operation of the different functional units shown may include one or more processors that can execute instructions stored in memory 27. The memory includes program code to be executed by processing unit 23.

[0027] Figure 2 The structure of a sequence diagram that can be used to generate a 3D dataset during a single breath-hold is shown. The imaging sequence uses a segmented 3D EPI sequence prepared by inversion recovery and employs a skip-type CAIPIRINHA undersampling mode. An inversion pulse 50 is applied, followed by n excitations with shorter excitation pulses 51 and 52 to excite magnetization. After each excitation, echo-plane imaging EPI-style data acquisition and gradient switching along the readout direction (as shown by gradient switches 61 and 62) are applied. Gradient switching along the readout direction (here, Gx) generates a gradient echo that is acquired during the switching of Gx. In the phase encoding directions Gy and Gz, the CAIPIRINHA undersampling mode is used, with gradient switches 71, 72, 73, and 74 included in one phase encoding direction and gradient switches 81 to 84 included in the other phase encoding direction.

[0028] Also refer to Figure 3 The sampling k-space pattern is shown in more detail. The sampling pattern uses CAIPIRINHA, and in this scheme, a set of several parameters parameterizes the sampling mask. Parameter R... y The undersampling factor describes the phase coding direction, parameter R. z Describes the undersampling factor in the partition coding direction, and Describe the partition offset between adjacent samples respectively.

[0029] exist Figure 3 In the diagram, the numbers indicate the order of the sampling points in the k-space, where all points marked "1" belong to the sampling after the first pulse in n inverted pulses, and all points marked "2" belong to the sampling after the first pulse in n inverted pulses. Figure 2The second data acquisition follows the second inversion pulse 52. The region containing the markers includes samples, where other empty regions are not sampled, and the readout direction extends perpendicular to the image plane. Furthermore, the k-space is divided into several n segments (here, S1 to S5), and a segmentation factor is described that determines how the sampling mask is segmented into EPI excitations. In the example shown, each segment S1 to S5 contains three different data acquisitions; however, it should be understood that this is for illustrative purposes only. Therefore, in Figure 3 In the example shown, S equals 3 and 3 data acquisition chains are acquired in each segment. Image acquisition is then further segmented between different cardiac cycles (here, the RR interval), thus introducing another parameter n, which is the number of shots performed after each inversion pulse. Figure 3 In the example shown, the number n is 5, and the segmentation divides the k-space into n regions, here five regions. An excitation is acquired from each region after each inversion pulse, starting from the center region and moving outwards in a center-pattern (as shown in numbers 2, 3, 4, and 5). Therefore, within each inversion pulse, the excitation indicated by region 110 is acquired, thus region 110 includes n excitations acquired within one heartbeat. Therefore, image acquisition is determined by five parameters R. y R z , S and n describe the order in which k-space points are acquired. Image acquisition was tested in patients using a 16-channel coil and a 32-channel spinal coil. The imaging sequences were acquired using the following parameters:

[0030] Lateral orientation, 2.5 mm isotropic resolution, providing 128 x 128 x 56 matrix and 320 x 320 x 140 mm matrix. 3 The field of view was [not specified]. The echo interval was 0.68 ms, the receiver bandwidth was 1964 Hz / px, a 20° flip angle was used, and a 1-2-1 binomial water excitation pulse was used for fat suppression. (R [not specified]) y R z , The parameters of (S, n) are (2, 2, 1, 8, 7), providing an echo train length of 16, an acquisition window of 119 ms, and a total undersampling factor of 4. A complete 3D cardiac image can be acquired over 19 heartbeats (e.g., one RR interval).

[0031] for Figure 2 and Figure 3The LGE acquisition shown uses a selected inversion time to suppress signals from viable myocardium. Additionally, a trigger delay can be selected so that the acquisition window for acquiring MR signals after the inversion pulse is placed during the quiescent phase of the heart, such as diastole. Image reconstruction can be performed using iterative sensing (SENSE) reconstruction, where the stopping criterion is that the residual falls below a predefined threshold, which is fixed to a certain threshold, such as 10. -6 Or in 10 -8 Or 10 -4 Any other value between these ranges. The coil sensitivity map used for reconstruction can be generated from an external low-resolution free-breathing gradient echo reference scan, which can preferably be normalized using a body coil.

[0032] Figure 4 The above combination is summarized. Figure 2 and Figure 3 Some steps for image acquisition are discussed. In step 210, 3D k-space sampling is performed using two phase-encoded directions and one readout direction, during which a sequence of inversion pulses is applied to invert magnetization, and after each inversion pulse, n different EPI readout excitations are used in the readout direction. In step 220, the MR image is then reconstructed based on the sampled k-space.

[0033] In summary, the above method enables rapid image acquisition, which allows for 3D single-breathe-hold image acquisition without requiring a very high undersampling factor.

Claims

1. A method for generating MR images of a patient at an MR imaging system (40), comprising: - The k-space is sampled using a k-space 3D acquisition scheme with two phase encoding directions and one readout direction, including: - Apply a reverse pulse sequence (50) to the patient to reverse the patient's magnetization. - Sampling of the k-space is performed using uniform undersampling of the k-space and n distinct EPI readout excitations along the readout direction after each inversion pulse, where n => 5. - Reconstruct the MR image based on the sampled k-space.

2. The method according to claim 1, wherein, The MR image is an MR image of the heart, and the 3D acquisition scheme is used in conjunction with cardiac gating, wherein the n readout excitations are acquired within one heartbeat.

3. The method according to claim 2, wherein, The k-space is divided into n regions (S1 to S5) in one of the phase encoding directions, and an EPI readout excitation from each of the n excitations is acquired in one heartbeat.

4. The method according to claim 3, wherein, In each of the n regions of the k-space, S excitations are acquired, where S => 3, and the k-space trajectories of the excitations in a region are offset relative to each other in a phase encoding direction of the phase encoding direction using the CAIPIRINHA imaging technique.

5. The method according to claim 4, wherein, S is in the range of 6 to 10.

6. The method according to claim 4 or 5, wherein, The MR image is reconstructed using a k-space interpolation technique, which includes GRAPPA with a 3D calibration kernel.

7. The method according to any of the preceding claims, wherein, After the contrast agent has been administered to the patient and the concentration of the contrast agent has reached equilibrium in the patient, sampling of the k-space is performed.

8. The method according to any of the preceding claims, wherein the n EPI readout excitations are distributed in a k-space such that the center of the k-space is sampled using the earlier excitations from the n excitations, and the region outside the center is sampled using the later excitations from the n excitations.

9. The method according to claims 3 and 8, wherein, After each inversion pulse, one of the n excitations is acquired in each of the n regions.

10. The method according to any of the preceding claims, wherein, n is between 5 and 10.

11. The method according to any one of claims 2 to 10, wherein, The inversion time is selected based on the MR signal of the myocardium, and preferably, the inversion time of the inversion pulse sequence (50) is selected such that the MR signal of the myocardium is substantially zero during MR signal acquisition.

12. An MR imaging system (40) includes a processing unit and a memory configured to store program code, which, when executed by the processing unit, causes the processing unit to perform the following steps: - The k-space is sampled using a k-space 3D acquisition scheme with two phase encoding directions and one readout direction, including: - Apply a reverse pulse sequence to the patient to reverse the patient's magnetization. - Sampling of the k-space is performed using uniform undersampling of the k-space and n distinct EPI readout excitations along the readout direction after each inversion pulse, where n => 5. - Reconstruct MR images based on the sampled k-space.

13. The MR imaging system according to claim 12, wherein, The processing unit is configured to perform the method according to any one of claims 2 to 11.

14. A computer program comprising program code to be executed by a processing unit of an MR imaging system, wherein, The execution of the program code causes the processing unit to perform the method according to any one of claims 1 to 11.

15. A carrier comprising the computer program according to claim 14, wherein, The carrier is one of electrical signals, optical signals, radio signals, and computer-readable storage media.

Images