Dixon-type water / fat separation MR imaging

Abstract

The invention relates to a method of Dixon-type MR imaging. The object (10) is subjected to at least two shots of an imaging sequence, each shot comprising an excitation RF pulse followed by a series of refocusing RF pulses, wherein at least one pair of phase encoding echoes is generated in each time interval between two successive refocusing RF pulses, i.e. a first echo at a first echo time and a second echo at a second echo time. In two respective shots of the imaging sequence, two sets of echo signal pairs are acquired using a bipolar pair of readout magnetic gradients, i.e. a first set and a second set. The bipolar pair of readout magnetic field gradients in the acquisition of the second set has a polarity opposite to the polarity of the bipolar pair of readout magnetic field gradients in the acquisition of the first set. Alternatively or additionally, the time course of the readout magnetic field gradients in the acquisition of the second set is inverted with respect to the time course of the readout magnetic field gradients in the acquisition of the first set. Alternatively or additionally, the acquisition of the first set and the acquisition of the second set differ from each other with respect to the gradient area of the magnetic field gradients in the readout direction (M) before and after the bipolar pair of readout magnetic field gradients, respectively. Finally, a MR image is reconstructed from the acquired first and second sets of echo signal pairs, whereby a signal contribution from water protons and a signal contribution from fat protons are separated. Furthermore, the invention relates to an MR device (1) and to a computer program to be run on the MR device (1).

Description

Technical Field

[0001] This invention relates to the field of magnetic resonance (MR) imaging. Specifically, it relates to a method for performing MR imaging on an object placed within the examination volume of an MR device. The invention also relates to an MR device and a computer program running on that device. Background Technology

[0002] Today, image-forming MR methods, which utilize the interaction between magnetic fields and nuclear spins to form two-dimensional or three-dimensional images, are widely used, especially in the field of medical diagnostics. This is because for the imaging of soft tissues, such image-forming MR methods are superior to other imaging methods in many ways, do not require ionizing radiation, and are generally non-invasive.

[0003] Typically, according to the MR method, the patient's body is positioned within a strong and uniform magnetic field B0, the direction of which simultaneously defines the axis of the coordinate system upon which the measurement is based (usually the z-axis). The magnetic field B0 generates different energy levels for individual nuclear spins depending on its strength, and these individual nuclear spins can be excited (spin resonance) by applying an electromagnetic field (RF field) at a defined frequency (Ramohr frequency) within the radio frequency range. From a macroscopic perspective, the distribution of individual nuclear spins produces a global magnetization, which can be deflected from equilibrium by applying an electromagnetic pulse (RF pulse) of appropriate frequency perpendicular to the z-axis, causing the magnetization to precess about the z-axis. This precession describes the surface of a cone, and the cone angle is called the flip angle. The amplitude of the flip angle depends on the intensity and duration of the applied electromagnetic pulse. In the case of a so-called 90° pulse, the spin deflects from the z-axis to the transverse plane (90° flip angle).

[0004] After the RF pulse terminates, the magnetization relaxes back to its original equilibrium state, in which magnetization in the z-direction is re-established with a first time constant T1 (spin lattice or longitudinal relaxation time), and magnetization in the direction perpendicular to the z-direction relaxes with a second time constant T2 (spin-spin or transverse relaxation time). This change in magnetization can be detected by means of a receiving RF coil arranged and oriented within the examination space of the MR device in a manner that measures the change in magnetization in the direction perpendicular to the z-axis. After applying, for example, a 90° pulse, the decay of transverse magnetization is accompanied by a transition of nuclear spins (caused by local magnetic field inhomogeneities) from an ordered state with the same signal phase to a state of uniform phase distribution (phase decoupling). Phase decoupling can be compensated by means of a refocusing pulse (e.g., a 180° pulse). This generates an echo signal in the receiving coil.

[0005] To achieve spatial resolution within the body, constant magnetic field gradients extending along three principal axes are superimposed on a uniform magnetic field B0, resulting in a linear spatial dependence of the Larmor frequency. The signal picked up in the receiving coil then contains components of different frequencies, which can be correlated to different locations within the body. The signal data obtained via the receiving coil corresponds to the spatial frequency domain and is referred to as k-space data. k-space data typically comprises data from multiple lines in k-space acquired using different phase encodings. Each k-space line is digitized by collecting a large number of samples. For example, the sample set from multiple lines in k-space is converted into an MR image using a Fourier transform.

[0006] In MR imaging, it is often desirable to obtain information about the relative contributions of water and fat to the total signal, so as to either suppress the contribution of one of them or analyze their contributions individually or together. These contributions can be calculated by combining information from two or more corresponding echoes acquired at different echo times (regarding excitation or spin echo refocusing). This can be considered chemical shift coding, in which an additional dimension (i.e., the chemical shift dimension) is defined and encoded by acquiring two or more MR images at slightly different echo times. For water / fat separation, these types of experiments are often referred to as Dixon-type measurements. Water / fat separation is achieved using Dixon MR imaging, or Dixon water / fat MR imaging, by calculating the contributions of water and fat based on two or more corresponding echoes acquired at different echo times. Typically, this separation is possible because of the known precession frequency difference of hydrogen in water and fat. In its simplest form, water and fat images are generated by addition or subtraction of in-phase and out-of-phase datasets.

[0007] Several Dixon-type MR imaging methods have been proposed in recent years. Besides different strategies for water / fat separation, the main characteristics of known techniques lie in the specific number of echoes (spots) they acquire and the constraints they impose on the echo times used. Conventional so-called two-point and three-point methods require in-phase and out-of-phase echo times when the water and fat signals are parallel and antiparallel in the complex plane, respectively. Three-point methods have been gradually extended to allow for flexible echo times. Therefore, they no longer restrict the angle or phase between the water and fat signals at echo times to certain values. In this way, they provide more degrees of freedom in imaging sequence design, and in particular, allow for trade-offs between the signal-to-noise ratio (SNR) gain from acquisition and the SNR loss during separation. To reduce scan time, sampling only two echoes instead of three is preferable. However, the constraint on echo time can actually make two-echo acquisition slower than three-echo acquisition. Eggers et al.'s article (Magnetic Resonance in Medicine, Vol. 65, pp. 96-107, 2011) proposed a flexible dual-echo Dixon-type MR imaging method that eliminates such limitations. By using this MR imaging method with more flexible echo times, it is no longer necessary to acquire in-phase and out-of-phase images; instead, in-phase and out-of-phase images can be optionally synthesized based on water and fat images.

[0008] Dixon-type MR imaging methods are often used in conjunction with fast (turbine) spin echo sequences employing a multiple repetition method. Typically, two or three interleaved measurements with shifted readout magnetic field gradients and acquisition windows are used. Figure 2 The figure depicts a schematic pulse sequence diagram of a conventional turbine spin echo (TSE) Dixon sequence. It shows the magnetic field gradient switching along the frequency encoding direction (M), phase encoding direction (P), and slice selection direction (S). Furthermore, it shows the RF excitation and refocusing pulses, and the time intervals (represented by ACQ) between them for acquiring echo signals. The figure covers the acquisition of the first three echo signals from one excitation of the imaging sequence. Double arrows indicate the shift of the readout magnetic field gradient (top) and the acquisition window ACQ (bottom) between multiple repetitions of one excitation with identical phase encoding. Based on the shift of the readout magnetic field gradient, different phase shifts are obtained for the signal contributions from water protons and fat protons, respectively, which is the basis of Dixon-type water / fat separation.

[0009] Compared to standard (non-Dixon) TSE sequences, Dixon TSE technology offers better fat suppression and multiple contrast in a single acquisition. However, the scan time increases because each excitation with the exact same phase encoding needs to be repeated multiple times. Furthermore, scan efficiency is reduced due to the dead time introduced to allow for readout of the magnetic field gradient and the shifting of the acquisition window. Alternatively, the echo interval increases, and longer or more echo trains are required. This results in less coverage and greater blurring in the reconstructed MR images, or again, longer scan times.

[0010] Alternatively, a multi-echo approach is employed, acquiring two or three echoes instead of one after each RF refocusing pulse. Ma et al.'s article (Magnetic Resonance in Medicine, Vol. 58, pp. 103-109, 2007) describes the three-echo DixonTSE technique, in which three echo signals are generated in each time interval between two successive refocusing RF pulses. A bipolar triplet with readout magnetic field gradients is used to acquire the echo signals to reduce the time interval between the three echo signals, and thus reduce the time interval between two successive refocusing RF pulses and the T2 attenuation on the echo train. Figure 3 An example of this method is shown. A pulse sequence diagram of a three-echo DixonTSE sequence is depicted, covering the acquisition of the first two echo signal triplets of an excitation. Three gradient recall echoes are acquired after each RF refocusing pulse. Tampered readout magnetic field gradients are present before and after the bipolar triplets of the readout magnetic field gradient. Vertical dashed lines indicate different echo times.

[0011] US 2016 / 0033605 A1 discloses a dual-echo DixonTSE technique in which only two echo signals are generated in each time interval between two successive refocused RF pulses. Furthermore, a bipolar pair for reading the magnetic field gradient is used to partially acquire the echo signals to reduce the time interval between the two echo signals and further reduce the time interval between the two successive refocused RF pulses and the T2 attenuation on the echo train. A turbulent readout magnetic field gradient precedes the bipolar pair for reading the magnetic field gradient, followed by a retrace and a turbulent readout magnetic field gradient.

[0012] For it to be effective, the multi-echo method must rely on, for example... Figure 3The bipolar readout magnetic field gradient is shown. This causes a biased resonance effect due to chemical shift or inhomogeneity of the main magnetic field B0, resulting in distortion in the opposite direction in the obtained single-echo images. Additionally, eddy currents cause phase errors between single-echo images. Furthermore, since multi-echo methods often require partial echo sampling (especially at higher main magnetic field strengths), correction becomes difficult or impossible, and image quality is adversely affected.

[0013] US Patent Application US 2016 / 0033605 relates to a dual-echo DixonTSE technique in which only two echoes are acquired during the time interval between successive refocusing pulses. Summary of the Invention

[0014] The object of this invention is to provide a method for Dixon water / fat separation that can be further improved by combining it with TSE collection.

[0015] According to the present invention, a method for performing MR imaging on an object placed in the examination volume of an MR device is disclosed. The method includes the following steps:

[0016] The object (10) is subjected to at least two excitations of an imaging sequence, each excitation comprising an excitation RF pulse followed by a series of refocusing RF pulses, wherein at least one pair of phase-coded echoes is generated in each time interval between two successive refocusing RF pulses, namely, a first echo at a first echo time and a second echo at a second echo time.

[0017] In each repetition interval, a first set of echo signal pairs from the object (10) is acquired during the first excitation of the imaging sequence using a bipolar pair with a readout magnetic field gradient.

[0018] In each repetition interval, a second set of echo signal pairs from the object (10) is acquired during the second excitation of the imaging sequence using a bipolar pair that reads out the magnetic field gradient.

[0019] Wherein, the bipolar pairs of readout magnetic field gradients acquired in the second set have polarities opposite to those of the bipolar pairs of readout magnetic field gradients acquired in the first set, and / or

[0020] The time process of the readout magnetic field gradient in the second set of acquisitions is reversed with respect to the time process of the readout magnetic field gradient in the first set of acquisitions, and / or

[0021] The areas of gradient of the first set of acquisitions and the second set of acquisitions are different from each other with respect to the magnetic field gradient in the readout direction (M) before and after the bipole pair of readout magnetic field gradients, respectively.

[0022] The MR image was reconstructed based on the first and second sets of the acquired echo signal pairs, thereby separating the signal contributions from water protons and the signal contributions from fatty protons.

[0023] According to the present invention, two instances (excitations) of a dual-echo TSE Dixon imaging sequence are used, each instance acquiring two echo signals in each interval between two refocused RF pulses using a bipolar pair of readout magnetic field gradients. Typically, each instance can include multiple excitations, and each excitation can acquire more than two echo signals. However, for simplicity, it is assumed that each instance includes one excitation, and thereafter it is assumed that each excitation acquires two echo signals. The phase encoding of the echo signals is selected such that each item in the first and second sets of acquisitions covers the required k-space region. Therefore, the present invention proposes to combine the above-described multi-repetition and multi-echo methods. By performing two instances of bipolar dual-echo acquisition (the bipolar pairs of readout magnetic field gradients have opposite polarities and / or the time process of the readout magnetic field gradient is reversed and / or the gradient area (the area under the time process of the magnetic field gradient) is shifted between perturbations, retracements, and / or phase loss of the readout magnetic field gradient before and after the bipolar pairs of readout magnetic field gradients), a high acquisition duty cycle is achieved, and high image quality is maintained by promoting the suppression of artifacts caused by bipolar acquisition. In other words, reconstruction involves suppressing or eliminating artifacts that may be caused by bipolar readout. More specifically, readout repetition at opposite polarities of the bipolar gradient lobes provides phase information due to eddy current effects. This phase information is employed in reconstruction to at least partially eliminate eddy current effects caused by gradient switching of the bipolar readout. The reversal of the time course of the readout field pulses due to different echo times improves the control of water / fat separation and can reduce or even eliminate water / fat separation artifacts and / or noise amplification. These aspects can be combined in reconstruction. The difference in gradient area of ​​the magnetic field gradient in the readout direction (M) before and after the bipolar pair improves the overall coverage of k-space, especially for partial echo acquisition. This improves the effective coverage of k-space and enhances the control of the conjugate symmetry problem in reconstruction.

[0024] In a preferred embodiment, only some or all of the echo signals from the first set and / or the second set are acquired. This means that the k-space is sampled only partially in the positive or negative readout direction. The two echoes can be sampled only partially toward the center of the repetition interval between two successive refocused RF pulses. This means, for example, sampling the first echo before it (i.e., sampling the first echo only partially in the positive readout direction) and sampling the second echo after it (i.e., sampling the second echo only partially in the negative readout direction). In this way, a smaller echo displacement difference can be obtained, which is particularly beneficial at higher main magnetic field strengths (3T or higher). Furthermore, the interval between two successive refocused RF pulses can be shortened to reduce scan time and T2 attenuation on the echo train.

[0025] In another preferred embodiment, reconstructing the MR images involves reconstructing single-echo images for each item in the first and second sets based on the acquired echo signal pairs, i.e., a first single-echo image belonging to the first echo time and a second single-echo image belonging to the second echo time. Eddy current-induced phase errors can then be eliminated by aligning the pixel-wise or voxel-wise phases of the first single-echo images in the first and second sets, respectively, and by aligning the pixel-wise or voxel-wise phases of the second single-echo images in the first and second sets, respectively. This is particularly useful when the bipolar pair of readout magnetic field gradients in the acquisition of the second set has the opposite polarity to that in the acquisition of the first set, since in this case the first and second echo times are identical for both sets. The thus corrected single-echo images can then be used for water / fat separation in a further image reconstruction process. This preferably involves: a first water / fat separation based on a first single-echo image from the first set and a single-echo image from the second set, thereby generating a first water image and a first fat image; and a second water / fat separation based on a second single-echo image from the first set and another single-echo image from the second set, thereby generating a second water image and a second fat image. Advantageously, the water / fat separation is performed based on single-echo images in which the readout magnetic field gradients have the same polarity during acquisition. In this way, fat shift and B0 distortion can be ignored in this step. Finally, the first water image and the second water image are combined into a final water image, and / or the first fat image and the second fat image are combined into a final fat image. For this purpose, for example, the fat image is registered with the water image, and the B0 image estimated by the water / fat separation is used to correct for B0 distortion.

[0026] Furthermore, reversing the time process of the readout magnetic field gradient in the first set of acquisitions to the time process of the readout magnetic field gradient in the second set of acquisitions, and / or shifting the gradient area before and after the bipolar pair of the readout magnetic field gradients due to turbulence, retracement, and / or phase loss of the readout magnetic field gradient, allows modification of the echo shift between the first and second set of acquisitions of the echo signal pair. In this way, the control of the inversion problem associated with water / fat separation can be improved, and water / fat separation artifacts and / or noise amplification can be reduced or even eliminated. This is particularly applicable when water / fat separation is performed in k-space to correct for fat displacement.

[0027] The method of the present invention described so far can be performed by means of an MR device comprising: at least one main magnet coil for generating a substantially uniform and static magnetic field B0 within an examination volume; a plurality of gradient coils for generating switching magnetic field gradients in different spatial directions within the examination volume; at least one RF coil for generating RF pulses within the examination volume and / or receiving MR signals from the body of an object positioned within the examination volume; a control unit for controlling the temporal succession of the RF pulses and the switching magnetic field gradients; and a reconstruction unit for reconstructing an MR image based on the received MR signals. The method of the present invention can be implemented by correspondingly programming the reconstruction unit and / or control unit of the MR device.

[0028] The method of the present invention can be advantageously performed on most MR devices currently used in clinical practice. Therefore, only a computer program is needed to control the MR device to perform the above-described method steps of the present invention. This computer program can reside on a data carrier or in a data network for download and installation in the control unit of the MR device. Attached Figure Description

[0029] The accompanying drawings disclose preferred embodiments of the invention. However, it should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of the invention. In the drawings:

[0030] Figure 1 An MR apparatus for performing the method of the present invention is shown;

[0031] Figure 2 A schematic (simplified) pulse sequence diagram of a conventional multi-repetition TSEDixon imaging sequence is shown;

[0032] Figure 3 A schematic (simplified) pulse sequence diagram of a conventional multi-echo TSEDixon imaging sequence using bipolar readout magnetic field gradients is shown.

[0033] Figure 4 A schematic (simplified) pulse sequence diagram according to a first embodiment of the present invention is shown;

[0034] Figure 5 A schematic (simplified) pulse sequence diagram according to a second embodiment of the present invention is shown; and

[0035] Figure 6 A schematic (simplified) pulse sequence diagram according to a third embodiment of the present invention is shown. Detailed Implementation

[0036] refer to Figure 1 The MR apparatus 1 is shown in block diagram form. The apparatus includes a superconducting or resistive main magnet coil 2, which creates a substantially uniform, time-constant main magnetic field B0 along the z-axis passing through the examination volume. The apparatus also includes a set 2' of (1st-order, 2nd-order, and (where applicable) 3rd-order) shimming coils, wherein the current flowing through the individual shimming coils of set 2' is controllable, thereby minimizing B0 deviation within the examination volume.

[0037] The magnetic resonance generation and manipulation system applies a series of RF pulses and switched magnetic field gradients to invert, excite, saturate, refocus, and spatially encode magnetic resonance, as well as otherwise encode magnetic resonance, to perform MR imaging.

[0038] More specifically, gradient pulse amplifier 3 applies current pulses along the x, y, and z axes of the examination volume to one of the selected whole-body gradient coils 4, 5, and 6. Digital RF frequency transmitter 7 transmits RF pulses or pulse packets to body RF coil 9 via transmit / receive switch 8 to transmit RF pulses into the examination volume. A typical MR imaging sequence consists of packets of short-duration RF pulse segments, which, together with any applied magnetic field gradient, achieve selective manipulation of the MRI. Specifically, the RF pulse selection is located in a portion of body 10 within the examination volume. The MR signal is also picked up by body RF coil 9.

[0039] To generate an MR image of a limited region of body 10, a set of local array RF coils 11, 12, and 13 is placed adjacent to the region selected for imaging. The array coils 11, 12, and 13 are capable of receiving MR signals generated by the body's RF coils.

[0040] The resulting MR signal is picked up by body RF coil 9 and / or array RF coils 11, 12, 13, and demodulated by receiver 14, preferably including a preamplifier (not shown). Receiver 14 is connected to RF coils 9, 11, 12, and 13 via transmit / receive switch 8.

[0041] The host computer 15 controls the shimming coil 2', gradient pulse amplifier 3, and transmitter 7 to generate the imaging sequence of the present invention. For a selected sequence, the receiver 14 rapidly and alternately receives signal data from one or more k-space lines after each RF excitation pulse. The data acquisition system 16 performs analog-to-digital conversion on the received signals and converts each k-space line into a digital format for further processing. In modern MR devices, the data acquisition system 16 is a separate computer dedicated to acquiring raw image data.

[0042] Ultimately, the raw digital image data is reconstructed into an image representation by reconstruction processor 17, which applies Fourier transform or other suitable reconstruction algorithms (e.g., SENSE). MR images can represent planar slices across the patient, arrays of parallel planar slices, three-dimensional volumes, etc. The image is then stored in an image memory, accessible in the image memory to convert the slices, projections, or other portions of the image representation into an appropriate format for visualization, for example via video monitor 18, which provides a human-readable display of the resulting MR image.

[0043] The host computer 15 and the reconstruction processor 17 are configured to perform the methods of the present invention described above and below using corresponding programs.

[0044] According to the present invention, two instances (excitations) of the dual-echo TSEDixon imaging sequence are used, each instance acquiring two echo signals in each interval between two refocused RF pulses using a bipolar pair with readout magnetic field gradients. Figure 4 A pulse sequence diagram of a dual-echo TSE sequence constituting the imaging sequence according to the present invention is shown. The diagram illustrates the magnetic field gradient switching along the frequency encoding direction (M), phase encoding direction (P), and slice selection direction (S). Furthermore, the diagram shows the RF excitation and refocusing pulses and the time intervals (denoted by ACQ) for acquiring echo signals between them. A pair of echo signals is acquired in each time interval between two successive refocusing RF pulses using a bipolar pair that reads out the magnetic field gradient. Figure 4 The first two pairs of echo signals from one excitation of the imaging sequence are covered. The polarity of the bipolar readout magnetic field gradient is obtained by performing two excitations of the imaging sequence (reversing the polarity between excitations). Figure 4 The double arrows (indicated in the image) indicate that a high acquisition duty cycle was achieved, and high image quality was maintained by facilitating the suppression of artifacts caused by bipolar acquisition (see the description above for details). Typically, the phase encoding in the two excitations does not need to be exactly the same. In this case, advanced parallel imaging and / or compressed sensing subsampling and reconstruction techniques can be advantageously applied.

[0045] Alternatively or additionally, the time process of reading out the magnetic field gradient can be reversed in the second excitation (i.e., the order or sequence of individual readout pulses of the magnetic field gradient). Figure 5 The diagram illustrates this situation. Figure 5 Pulse sequence diagram and Figure 4 They are basically the same, the difference being that, in such cases... Figure 5 In the second excitation shown, the time process of the readout magnetic field gradient in the frequency encoding direction (M) is related to Figure 4 The first excitation shown has reversed.

[0046] Alternatively or additionally, in addition to the bipolar pair, the area of ​​the readout magnetic field gradient can also vary between the first and second excitations while keeping the sum of these gradient areas constant. Figure 6 The diagram illustrates this situation. Figure 6 Pulse sequence diagram and Figure 4 The results are largely the same, except that the gradient area shifts from the turbulence-induced readout of the magnetic field gradient after the bipolar pair readout to the turbulence-induced readout before the bipolar pair readout. Additionally, optional, minor timing optimizations are performed to extend the acquisition window slightly.

Claims

1. A method for performing MR imaging on an object (10) placed in an examination volume of an MR device (1), the method comprising the following steps: The object (10) is subjected to at least two excitations of an imaging sequence, each excitation comprising an excitation RF pulse followed by a series of refocusing RF pulses, wherein at least one pair of phase-coded echoes is generated in each time interval between two successive refocusing RF pulses, namely, a first echo at a first echo time and a second echo at a second echo time. In each repetition interval, a first set of echo signal pairs from the object (10) is acquired during the first excitation of the imaging sequence using a bipolar pair with a readout magnetic field gradient. In each repetition interval, a second set of echo signal pairs from the object (10) is acquired during the second excitation of the imaging sequence using a bipolar pair with a readout magnetic field gradient. Wherein, the bipolar pairs of readout magnetic field gradients acquired in the second set have polarities opposite to those of the bipolar pairs of readout magnetic field gradients acquired in the first set, and / or The time process of the readout magnetic field gradient in the second set of acquisitions is reversed with respect to the time process of the readout magnetic field gradient in the first set of acquisitions, and / or The areas of gradient of the first set of acquisitions and the second set of acquisitions are different from each other with respect to the magnetic field gradient in the readout direction (M) before and after the bipole pair of readout magnetic field gradients, respectively. The system is configured to reconstruct MR images based on a first set and a second set of acquired echo signal pairs, thereby separating the signal contributions from water protons and from lipid protons, wherein the reconstruction includes suppressing or eliminating artifacts that can be caused by bipolar acquisition.

2. The method according to claim 1, wherein, Only some or all of the echo signals from the first set and / or the second set are collected.

3. The method according to claim 1 or 2, wherein, Reconstructing the MR images involves reconstructing a single-echo image for each item in the first set and the second set based on the acquired echo signal pairs, namely, a first single-echo image belonging to the first echo time and a second single-echo image belonging to the second echo time.

4. The method according to claim 3, wherein, Eddy current-induced phase errors are eliminated by aligning the pixel-wise or voxel-wise phases of the first single-echo images of the first set and the second set with the same pixel-wise or voxel-wise phases.

5. The method according to claim 3, wherein, Reconstructing the MR image involves: performing a first water / fat separation based on a first single-echo image from the first set and a single-echo image from the second set, thereby generating a first water image and a first fat image; A second water / fat separation is performed based on the second single-echo image of the first set and another single-echo image of the second set, thereby generating a second water image and a second fat image.

6. The method according to claim 5, wherein, Correct for fat displacement and / or B0 distortion.

7. The method according to claim 5, wherein, The first water image and the second water image are combined into a final water image, and / or the first fat image and the second fat image are combined into a final fat image.

8. An MR device, comprising: At least one main magnet coil (2) is used to generate a uniform static magnetic field B0 within the inspection volume; A plurality of gradient coils (4, 5, 6) for generating switching magnetic field gradients in different spatial directions within the examination volume; at least one RF coil (9) for generating RF pulses and / or receiving MR signals from an object (10) positioned in the examination volume; a control unit (15) for controlling the temporal succession of the RF pulses and the switching magnetic field gradients; and a reconstruction unit (17) for reconstructing MR images based on the received MR signals, wherein the MR device (1) is configured to perform the steps of the method according to any one of claims 1-7.

9. A computer program product comprising a computer program for running on an MR device, said computer program including instructions required to perform the method according to any one of claims 1-7.

Images