FIELD OF THE INVENTION

The invention relates to the field of magnetic resonance (MR) imaging. It concerns a method of MR imaging of an object. The invention also relates to a MR device and to a computer program to be run on a MR device.

BACKGROUND OF THE INVENTION

Image-forming MR methods which utilize the interaction between magnetic fields and nuclear spins in order to form two-dimensional or three-dimensional images are widely used nowadays, notably in the field of medical diagnostics, because for imaging of soft tissue they are superior to other imaging methods in many respects, do not require ionizing radiation and are usually not invasive.

According to the MR method in general, the object, for example the body of the patient to be examined, is arranged in a strong, uniform magnetic field whose direction at the same time defines an axis (normally the z-axis) of the co-ordinate system on which the measurement is based. The magnetic field produces different energy levels for the individual nuclear spins in dependence on the magnetic field strength which can be excited (spin resonance) by application of an electromagnetic alternating field (RF field) of defined frequency (so-called Larmor frequency, or MR frequency). From a macroscopic point of view the distribution of the individual nuclear spins produces an overall magnetization which can be deflected out of the state of equilibrium by application of an electromagnetic pulse of appropriate frequency (RF pulse), so that the magnetization performs a precessional motion about the z-axis. The precessional motion describes a surface of a cone whose angle of aperture is referred to as flip angle. The magnitude of the flip angle is dependent on the strength and the duration of the applied electromagnetic pulse. In the case of a so-called 90° pulse, the spins are deflected from the z axis to the transverse plane (flip angle 90°).

After termination of the RF pulse, the magnetization relaxes back to the original state of equilibrium, in which the magnetization in the z direction is built up again with a first time constant Ti (spin lattice or longitudinal relaxation time), and the

magnetization in the direction perpendicular to the z direction relaxes with a second time constant T ₂ (spin-spin or transverse relaxation time). The variation of the magnetization can be detected by means of receiving RF coils which are arranged and oriented within an examination volume of the MR device in such a manner that the variation of the

magnetization is measured in the direction perpendicular to the z-axis. The decay of the transverse magnetization is accompanied, after application of, for example, a 90° pulse, by a transition of the nuclear spins (induced by local magnetic field inhomogeneities) from an ordered state with the same phase to a state in which all phase angles are uniformly distributed (dephasing). The dephasing can be compensated by means of a refocusing pulse (for example a 180° pulse). This produces an echo signal (spin echo) in the receiving coils.

To realize spatial resolution in the body, constant magnetic field gradients extending along the three main axes are superposed on the uniform magnetic field, leading to a linear spatial dependency of the spin resonance frequency. The signal picked up in the receiving coils then contains components of different frequencies which can be associated with different locations in the body. The signal data obtained via the receiving coils corresponds to the spatial frequency domain and is called k-space data. The k-space data usually includes multiple lines acquired with different phase encoding. Each line is digitized by collecting a number of samples. A set of k-space data is converted to a MR image by means of an image reconstruction algorithm.

The so-called 'Turbo Spin Echo' (TSE) sequence is a well-known multi-echo imaging sequence. One 'shot' of the TSE sequence comprises an initial RF pulse for excitation of magnetic resonance followed by a plurality of rapidly applied (typically 180°) refocusing RF pulses generating a train of differently phase-encoded spin echo signals. The echo signals are acquired, wherein each echo signal represents a k-space line, i.e. a one- dimensional sample of k-space, the position of the k-space line being determined by the applied frequency encoding and phase encoding switched magnetic field gradients of the sequence. The so-called 'Turbo Factor' (TF) is the number of echoes acquired after each excitation. Typically, a plurality of shots of the multi-echo sequence is applied for completely sampling k-space in order to be able to reconstruct a MR image from the acquired k-space profiles, k-space is typically segmented such that each segment is attributed to one echo index number in the train of echoes generated by each shot of the sequence. In other words, each shot typically acquires k-space profiles from all segments of k-space, with the number of segments in k-space being equal to TF.

TSE sequences are applied today in almost all applications. Despite its wide use it is known to be very sensitive to motion due to the segmented k-space distributed acquisition. Small movements of the imaged object, for example of subcutaneous fat, can introduce ghosts overlapping with the part of interest disturbing the overall diagnostic value of a scan.

The above-mentioned k-space segmentation principles and problems do apply to both spin echo (TSE) and gradient echo sequences (TFE) and 'Echo Planar Imaging' - EPI - sequences.

Deriving motion information from correlations between sets of phase- encoding lines is known from the US-patent US 8 155 389.

The US-patent application US2008/061779 concerns a magnetic resonance data acquisition for three-dimensional MR imaging. This data acquisition employs k- lines (in 3D k-space) that are distributed of partial acquisitions. That is, each partial acquisition includes a number of kx-pro files (in the frequency encoding direction). Respective k _x -pro files have different (k _y ,k _z ) two-dimensional phase-encodings. Notably, in the (k _y ,k _z ) plane a radial scanning approach is employed. An even distribution of k-pro files is employed (over the two-dimensional phase encoding plane) in the central and peripheral k-space over the entire measuring time.

SUMMARY OF THE INVENTION

From the foregoing it is readily appreciated that there is a need for an improved MR imaging technique. It is consequently an object of the invention to enable fast MR imaging which is more robust with respect to motion.

In accordance with the invention, a method of MR imaging of an object placed in an examination volume of a MR device is disclosed. The method comprises the steps of:

generating echo signals by subjecting the object to an imaging sequence, - acquiring the echo signals, each echo signal being attributed to a k-space line, wherein a number of k-space lines, which are adjacently arranged in a part of k-space, are repeatedly sampled, with said number of k-space lines being sampled in a different sequential order per repetition, and

reconstructing a MR image from the acquired echo signals.

The k-space scanning approach is applied to the two-dimensional k-space (k-plane) spanned by the read-out (frequency encoding direction (k _x )and the phase-encoding direction

(k _y ). According to the invention, a number of k-space lines from a part of k-space are sampled repeatedly. This means that the same k-space lines of said number are sampled two or more times. The repeated sampling is performed for the purpose of signal averaging, for example. At different repetitions, the spin condition or magnetisation condition is selected to be different. For example, averaginig is done over repetitions at different, notably opposite phases of the magnetisaton. In Dixon acquisition, different echo times are selected in the different repetitions. In phase-sensitive inversion acquisition, different repetitions are made with and withoug application of a spin inversion, respectively. The respective repeatedly sampled k-space lines are sampled with a different sequential order per repetition in order to compensate for motion artefacts. The sequential order defines the order in which the k-space lines are successively sampled. The sequential order is determined by the attribution of the k- space position (i.e. the phase-encoding) of each k-space line to the position of the respective echo signal in the succession of acquired echo signals. This means, in other words, that the (for example random) sequential order in which the k-space lines are sampled by the imaging sequence is varied from repetition to repetition, which also implies that the temporal succession in which said number of k-space lines are acquired generally differs from the sequential arrangement of the respective k-space lines in the respective part of k-space. This approach achieves a particularly effective averaging-out of motion artefacts.

The imaging sequence applied according to the invention may be a single-echo spin echo (SE) or gradient echo (GE) sequence. In an alternative embodiment, the imaging sequence may be a multi-echo imaging sequence, wherein the object is subjected to a number of shots of the multi-echo imaging sequence, a train of echo signals being generated by each shot. In this embodiment, the sequential order in which said number k-space lines are acquired is determined by the attribution of the k-space position of each k-space line to an echo index number indicating the number of the respective echo signal in the train of echo signals generated by each shot.

In a further preferred embodiment of the invention, k-space is divided into a central k-space part and one or more peripheral k-space parts, wherein the number of signal averages in the central k-space part is higher than the number of signal averages in the peripheral k-space parts. According to this embodiment, the invention proposes a motion- insensitive partial averaging approach, in which the central part of k-space is sampled with a higher number of acquisitions for the purpose of signal averaging in order to compensate for motion artefacts. Different signal averaging is applied in the central and peripheral k-space parts, respectively. Motion artefacts are efficiently averaged out in the reconstructed MR image by applying a higher number of signal averages in the central k-space part than in the peripheral k-space parts since the central k-space data mainly determine the overall image contrast. It is noted that signal averaging may be performed according to the invention in k-space or, for example by using an iterative MR image reconstruction technique, in image space.

Preferably, at least some of the repeatedly sampled k-space lines (of the central and/or peripheral k-space parts) are sampled in a random order. The randomization distributes motion-induced discontinuities of the acquired k-space data. This avoids a regular k-space spacing of motion-corrupted k-space data and, consequently, ghosting in the reconstructed MR image.

According to yet another preferred embodiment of the invention, at least a part of k-space is sampled with variable density. This means, in other words, that the acquired k- space lines are irregularly distributed in the phase-encoding direction. The variable k-space density adds to reduce ghosting by avoiding a regular distribution of motion-corrupted k- space samples in k-space.

Moreover, at least a part of k-space, preferably the peripheral k-space parts, may be incompletely sampled, which means that the sampling density in the respective k- space regions is below the Nyquist limit with regard to the field-of-view to be imaged.

This may be employed, for example, in order to reduce scan time by motion- optimized variable density compressed sensing. The theory of CS is known to have a great potential for significant signal data reduction. In CS theory, a signal data set which has a sparse representation in a transform domain can be recovered from undersampled

measurements by application of a suitable regularisation algorithm. The possibility of undersampling leads to a significantly reduced acquisition time. As a mathematical framework for signal sampling and reconstruction, CS prescribes the conditions under which a signal data set can be reconstructed exactly or at least with high image quality even in cases in which the k-space sampling density is far below the Nyquist criterion, and it also provides the methods for such reconstruction. In most existing CS-based MR acquisition and reconstruction schemes the basic CS formulation is used which exploits only the prerequisite that the MR signal data is sparse in a transform domain.

Further, the method of the invention may be combined with the per se known 'partial matrix' technique. 'Partial matrix' is a scan time reduction method that takes advantage of the complex conjugate property of k-space. Since k-space profiles from negative k-space positions are substantially identical to complex conjugate k-space profiles from corresponding positive k-space positions, only a peripheral positive or negative part may be sampled in accordance with the invention in order to replicate a sampling of the entire k-space. This results in a significant reduction in scan time while full resolution is maintained.

According to a preferred embodiment of the invention, the multi-echo imaging sequence of the invention is a spin echo sequence, of which one shot comprises:

- at least one RF pulse for excitation of magnetic resonance,

a plurality of refocusing RF pulses,

a plurality of switched magnetic field gradients for frequency-encoding of each echo signal, and

a plurality of switched magnetic field gradients for phase-encoding of each echo signal. In other words, the method of the invention may be applied in combination with per-se known 'turbo' spin echo sequences, such as TSE or spin echo-type EPI.

Alternatively, the multi-echo imaging sequence is a gradient echo sequence, of which one shot comprises:

at least one RF pulse for excitation of magnetic resonance,

- a plurality of switched magnetic field gradients for refocusing magnetic resonance and for frequency-encoding of each echo signal, and

a plurality of switched magnetic field gradients for phase-encoding of each echo signal. In other words, the method of the invention may be applied in combination with per-se known 'turbo' gradient echo sequences, such as TFE or EPI.

The method of the invention is particularly well-suited to be combined with

Cartesian k-space sampling schemes. However, the method may as well be applied in combination with non-Cartesian sampling. For all types of k-space sampling, the term 'central k-space part' within the meaning of the invention refers to k-space profiles close to and encompassing the k-space origin.

According to yet another preferred embodiment, motion information is derived from the echo signals acquired from the repeatedly sampled k-space lines. For example, echo signals acquired from the same k-space positions at different points in time may be correlated in order to extract motion information. The motion information may be employed for corresponding motion compensation of the acquired signal data by appropriate post- processing. Alternatively, if the detected motion is found to be too severe to be compensated for, it may be decided automatically to discard the motion-corrupted k-space profiles. Thus, motion artefacts are either averaged out according to the invention or completely avoided by correcting or discarding the corrupted data. As known to those skilled in the field of MR imaging, it is often desired to obtain information about the relative contribution of different chemical species, such as water and fat, to the overall signal, either to suppress the contribution of some of them or to separately or jointly analyze the contribution of all of them. These contributions can be calculated if information from two or more corresponding echoes, acquired at different echo times, is combined. This may be considered as chemical shift encoding, in which an additional dimension, the chemical shift dimension, is defined and encoded by acquiring two or more echo signal data sets at slightly different echo times. In particular for water-fat separation, these types of experiments are often referred to as Dixon-type of measurements. By means of Dixon imaging or Dixon water/fat imaging, a water/fat separation can be achieved by calculating contributions of water and fat from two or more corresponding echoes, acquired at different echo times. In general such a separation is possible because there is a known precessional frequency difference of hydrogen in fat and water. In its simplest form, water and fat images are generated by either addition or subtraction of the 'in phase' and Out of phase' datasets.

In a preferred embodiment, the method of the invention is combined with Dixon imaging. Thus, in other words, the imaging sequence is a Dixon sequence, wherein signal contributions from water and fat to the acquired echo signals are separated. It is an important insight of the invention that the particularly motion-robust acquisition scheme of the invention is well-suited for Dixon imaging, wherein the low motion artefacts in the reconstructed MR image leads to a high diagnostic image quality. In the case of Dixon imaging, the repeated sampling of identical k-space lines may advantageously be performed with a different echo time per repetition.

Most preferably, the echo signals are acquired at each echo time of the Dixon sequence with a different order of the k-space lines. A different randomisation of the acquisition order results in an optimized averaging-out of motion artefacts.

This applies in particular in combination with the reconstruction of a so-called modulus-added weighted fat and water image, wherein the signal contributions from water and fat are combined in the step of reconstructing the MR image. For reconstructing a modulus-added weighted fat and water image, the nth root of the weighted sum of the modulus of the fat contribution raised to the power n and the modulus of the water contribution raised to the power n is computed for each image pixel or voxel.

More generally, the method of the invention may be combined with any type of imaging technique in which k-space lines are repeatedly acquired and at least one imaging parameter is varied from repetition to repetition. Hence, the method of the invention may be combined advantageously with inversion recovery techniques (IR/PSIR), diffusion- weighted imaging, magnetization transfer contrast (MTC), chemical exchange saturation transfer (CEST) and so forth.

In a further aspect of the invention, phase-sensitive magnetic resonance images are reconstructed in that phase-information of the acquired magnetic resonance signal is preserved in the reconstrution. That is, pixels of the reconstructed magnetic resonance images are complex valued.

The method of the invention described thus far can be carried out by means of a MR device including at least one main magnet coil for generating a uniform static magnetic field within an examination volume, a number of gradient coils for generating switched 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 for receiving MR signals from an object positioned in the examination volume, a control unit for controlling the temporal succession of RF pulses and switched magnetic field gradients, and a reconstruction unit. The method of the invention can be implemented, for example, by a corresponding programming of the reconstruction unit and/or the control unit of the MR device.

The method of the invention can be advantageously carried out in most MR devices in clinical use at present. To this end it is merely necessary to utilize a computer program by which the MR device is controlled such that it performs the above-explained method steps of the invention. The computer program may be present either on a data carrier or be present in a data network so as to be downloaded for installation in the control unit of the MR device.

BRIEF DESCRIPTION OF THE DRAWINGS

The enclosed drawings disclose preferred embodiments of the present invention. It should be understood, however, that the drawings are designed for the purpose of illustration only and not as a definition of the limits of the invention. In the drawings:

Figure 1 shows a MR device for carrying out the method of the invention;

Figure 2 shows a diagram of k- space schematically illustrating a first embodiment of the sampling scheme of the invention;

Figure 3 shows a diagram of k- space schematically illustrating a second embodiment of the sampling scheme of the invention; Figure 4 shows a diagram of k- space schematically illustrating a third embodiment of the sampling scheme of the invention;

Figure 5 shows a diagram of k- space schematically illustrating a fourth embodiment of the sampling scheme of the invention;

Figure 6 shows a diagram of k-space schematically illustrating a fourth embodiment of the sampling scheme of the invention;

Figure 7 shows schematically the in-phase, out-of-phase, water and fat images with motion ghost contributions;

Figure 8 shows a diagram of k-space schematically illustrating sampiong of in- phase and out-of-phase in a Dixon measurement.

DETAILED DESCRIPTION OF THE EMBODIMENTS

With reference to Figure 1 , a MR device 1 is shown. The device comprises superconducting or resistive main magnet coils 2 such that a substantially uniform, temporally constant main magnetic field is created along a z-axis through an examination volume.

A magnetic resonance generation and manipulation system applies a series of RF pulses and switched magnetic field gradients to invert or excite nuclear magnetic spins, induce magnetic resonance, refocus magnetic resonance, manipulate magnetic resonance, spatially and otherwise encode the magnetic resonance, saturate spins, and the like to perform MR imaging.

More specifically, a gradient pulse amplifier 3 applies current pulses to selected ones of whole-body gradient coils 4, 5 and 6 along x, y and z-axes of the

examination volume. A digital RF frequency transmitter 7 transmits RF pulses or pulse packets, via a send-/receive switch 8, to a whole-body volume RF coil 9 to transmit RF pulses into the examination volume. A typical MR imaging sequence is composed of a packet of RF pulse segments of short duration which taken together with each other and any applied magnetic field gradients achieve a selected manipulation of nuclear magnetic resonance. The RF pulses are used to saturate, excite resonance, invert magnetization, refocus resonance, or manipulate resonance and select a portion of a body 10 positioned in the examination volume. The MR signals are also picked up by the whole-body volume RF coil 9.

For generation of MR images of limited regions of the body 10, a set of local array RF coils 11, 12, 13 are placed contiguous to the region selected for imaging. The array coils 11, 12, 13 can be used to receive MR signals induced by body-coil RF transmissions. The resultant MR signals are picked up by the whole body volume RF coil 9 and/or by the array RF coils 11, 12, 13 and demodulated by a receiver 14 preferably including a preamplifier (not shown). The receiver 14 is connected to the RF coils 9, 11, 12 and 13 via send-/receive switch 8.

A host computer 15 controls the gradient pulse amplifier 3 and the transmitter

7 to generate any of a plurality of MR imaging sequences, such as echo planar imaging (EPI), echo volume imaging, gradient and spin echo imaging, fast spin echo (TSE) imaging, and the like. For the selected sequence, the receiver 14 receives a single or a plurality of MR data lines in rapid succession following each RF excitation pulse. A data acquisition system 16 performs analog-to-digital conversion of the received signals and converts each MR data line to a digital format suitable for further processing. In modern MR devices the data acquisition system 16 is a separate computer which is specialized in acquisition of raw image data.

Ultimately, the digital raw image data is reconstructed into an image representation by a reconstruction processor 17 which applies a Fourier transform or other appropriate reconstruction algorithms. The MR image may represent a planar slice through the patient, an array of parallel planar slices, a three-dimensional volume, or the like. The image is then stored in an image memory where it may be accessed for converting slices, projections, or other portions of the image representation into appropriate format for visualization, for example via a video monitor 18 which provides a man-readable display of the resultant MR image.

With continuing reference to Figure 1 and with further reference to Figures 2- 5, embodiments of the motion-robust partial averaging approach of the invention are explained.

According to the invention, the body 10 is subjected to a number of shots of a multi-echo imaging sequence, such as, for example, a TSE sequence. A train of echo signals is generated by each shot of the sequence. The echo signals are acquired, wherein each echo signal is attributed to a k-space line. The k-space lines are represented in Figures 2-5 as thin horizontal lines extending parallel to the k _x axis. Signal averaging is performed by repeatedly sampling at least some of the k-space lines. Repeated sampling means that identical k-space lines are sampled two or more times during the signal acquisition process for reconstructing the final MR image, wherein a signal to noise improvement is achieved by taking the average of the several echo signals acquired from the same k-space positions respectively. The number of averages is also referred to as the number of acquisitions (NSA). The approximate amount of improvement in signal to noise (SNR) is calculated as the square root of the number of averages.

As shown in Figure 2, k-space 20 is divided into a central k-space part 21 , a first peripheral k-space part 22, and a second peripheral k-space part 23. In the depicted embodiment, the central k-space part 21 is sampled by each shot of the TSE sequence, first peripheral k-space part 22 is sampled by every 2 ^nd shot, while second peripheral k-space part 23 is sampled by every 4 ^th shot. According to the proposed partial averaging approach, a variable repeated sampling of k-space positions is performed, wherein central k-space 21 is acquired more often compared to the peripheral parts 22, 23. The central k-space part 21 with its limited set of k-space lines is sampled very frequently. Motion information may be derived from the echo signals acquired from the repeatedly sampled k-space lines of central part 21. For example, cross correlations of the k-space data acquired during different averaging steps may be computed. On this basis, an automated motion compensation may be performed. In case the detected motion is found to be too severe, the corresponding echo signal data may be rejected. In case the position of the imaged object is found to have irreversibly changed due to motion, the scan parameters may be updated automatically according to the position shift. The intermediate peripheral k-space part 22 with a larger set of k-space lines is scanned less frequently to average out motion, compensate for artefacts and signal decay effects like T ₂ relaxation, and to provide at least a certain improvement of SNR. The second peripheral part 23 is acquired only once. It may be acquired even less than once depending on whether, for example CS and/or partial matrix techniques are used.

Finally, a MR image is reconstructed from the acquired echo signals. This general approach can be applied to all types of sequences, like FFE, SE, TFE, TSE and EPI sequences.

A further aspect of the invention, which is illustrated in Figures 3-5, is that k- space sampling is performed with variable density, with a denser sampling of k-space in the central k-space part 21 and a sparser sampling of the peripheral k-space parts 22, 23. As illustrated in Figures 3-5, the acquired k-space lines are irregularly distributed in the phase- encoding direction k _y . The variable k-space density effects a reduction of ghosting by avoiding a regular distribution of motion-corrupted k-space samples in k-space 20. This variable density acquisition is especially advantageous in turbo sequences like TSE or TFE, wherein the image contrast is preserved and is essentially independent of the choice of the number of acquisitions in the peripheral k-space parts 22, 23. As in the embodiment shown in Figure 2, the central k-space part is updated more often, whereas the peripheral k-space parts 22, 23 are updated with a decreased frequency. Consequently, the most relevant, signal dominating k-space lines from central k-space 21 are updated within a short time period (for example within each shot of the multi-echo imaging sequence). In addition, the size of the k- space parts 21, 22, 23 increases to outer k-space, which results in a further distribution of motion induced k-space discontinuities over k-space 20.

Figure 4 illustrates the k-space lines acquired during a first and second acquisition step designated as ACQl and ACQ2 respectively. In this embodiment, the positions of the k-space lines in the peripheral k-space part 22 are shifted in the second acquisition step ACQ2 in the k _y -direction. The k _y positions of the k-space lines in the first acquisition step ACQl are indicated as dotted lines in the right diagram of Figure 4 for the purpose of illustration. In the central k-space part 21, in contrast, the k _y positions of the k- space lines scanned during the first and second acquisition steps ACQl and ACQ2 are maintained, resulting in a correspondingly repeated sampling of the k-space lines in the central k-space part 21, while each k-space line in the peripheral k-space part 22 is scanned only once. During each individual acquisition step ACQl, ACQ2, the peripheral k-space part 22 is incompletely sampled, which means that the sampling density in the k _y direction is below the Nyquist threshold. The combination of the k-space lines sampled in both acquisition steps ACQ 1 , ACQ2 results in a completely sampled data set.

The main aspect of the invention is the repeated sampling of the respective k- space lines in a different order in each repetition, wherein the sampling order defines the temporal order in which the k-space lines are sampled. Preferably, a different random order is applied per acquisition step. The technique of the invention, which is applied in combination with all embodiments of the invention illustrated in Figures 2-5, averages out motion artefacts and distributes motion-induced discontinuities of the acquired k-space data resulting in an irregular k-space spacing of motion-corrupted k-space data. Ghosting in the

reconstructed MR image is avoided in this way. For example, with reference to Figure 4, the (preferably random) temporal order in which the k-space lines are sampled from the central k-space part 21 during acquisition step ACQ2 is different from the temporal sampling order applied in the central k-space part 21 during acquisition step ACQl .

In the embodiment illustrated in Figure 5, a random distribution of the k-space sampling grid is applied in the peripheral k-space part 22, wherein those k-space lines are sampled during acquisition step ACQ2 that have been missed out during acquisition step ACQl . The combination of both acquisitions ACQl, ACQ2 results in a fully sampled data set. Like in the embodiment of Figure 4, central k-space part 21 is fully sampled during each acquisition step ACQl, ACQ2 such that repeated sampling is achieved. The random undersam ling of the peripheral k- space part 22 during each acquisition step ACQ1, ACQ2 results in a yet further improved distribution of motion-corrupted k-space samples over k- space and, consequently, in a further improved motion robustness.

In the embodiment shown in Figure 6, again k-space 20 is divided into a central k-space part 21 and a peripheral k-space part 22. Central k-space part 21 is repeatedly sampled by correspondingly repeated application of a single-echo spin echo (SE) sequence in the depicted embodiment. The central k-space part 21 comprises a limited number of adjacently arranged k-space lines that are designated as Kl, K2, K3, K4, K5. Each k-space line Kl, K2, K3, K4, K5 corresponds to an echo signal generated by the SE sequence. The k- space lines Kl, K2, K3, K4, K5 are repeatedly sampled with a different sequential order per repetition according to the invention. The sequential order may be varied as follows during, for example, four repetitions:

Repetition no. K1, K2, K3, K4, K5;

Repetition no. K2, K1, K4, K3, K5;

Repetition no. K2, K4, K1, K5, K3;

Repetition no. K3, K5, K2, K4, Kl .

Signal averaging is performed on this basis to average out motion, compensate for artefacts and signal decay effects like T ₂ relaxation, and to provide an improvement of SNR. The peripheral k-space part 22 is scanned only once. Finally, a MR image is reconstructed from the acquired echo signals.

In Dixon-type measurements, motion ghosts are typically encoded in the water or fat image dependent on the spectral source as shown for the simple in-phase out-phase Dixon example of Figure 7, (W = IP + OP, F = IP - OP). Here water motion leads to a motion ghost) with the full amplitude in the water image and not in the fat image. The left portion of Figure 7 relates to uncorrected motion, the right portion of Figure relates to correlated motion.

In case motion is uncorrected as shown for the simple in-phase out-phase Dixon example below left the motion ghost is present in both water and fat images, nevertheless the artefact is reduced by a factor of 2. This averaging effect can be used to effectively reduce motion artefacts.

In DIXON TSE the different profile ordering per Dixon acquisition can be achieved by a different, for example, randomized ordering of shots as illustrated below. Per shot a profile is acquired per k-space segment (in the example of Figure 8) 7 profiles out of the 7 k-space segments are sampled). The unequidistant or random sampling of profiles (blue lines) in a shot is varied for the Dixon acquisitions e. g. for IP and OP acquisitions. In case of motion this different sampling per Dixon acquisition leads to a motion ghost that is uncorrelated in the IP and OP acquisition. Applying Dixon reconstruction this motion ghost is averaged out.

According to one aspect of the invention k-space parts are randomly ordered with a different sequential, preferably random, order per Dixon acquisition.

For turbo sequences, like TSE or TFE, profiles acquired over shots and hence positioned close in k-space within k-space segments are randomly distributed and this random order changes with the Dixon acquisition.

For non-turbo sequences like FFE, k-space is subdivided in k-space segments with a random subsegment ordering. The random subsegment ordering is changed per Dixon acquisition.

The averaging effect is enhanced the more Dixon acquisitions are used. Hence in case of Dixon quantification where typically six (6 )echoes are used the averaging effect is optimized.