METHODS FOR ACQUIRING A MAGNETIC RESONANCE IMAGE DATASET AND FOR GENERATING A MOTION-CORRECTED IMAGE DATASET

Abstract

A method for acquiring a magnetic resonance image dataset of an object includes using an imaging protocol in which a number of k-space lines are acquired in one shot. The imaging protocol includes a plurality of shots. A plurality of additional k-space lines are acquired in at least a subset of the shots, such that movement of the object is detected throughout the imaging protocol. A method for generating a motion-corrected magnetic resonance image dataset from the dataset thus acquired, a magnetic resonance imaging apparatus, and a computer program are also provided.

Claims

1. A method for acquiring a magnetic resonance image dataset of an object, the method comprising: using an imaging protocol in which spatial encoding is performed using phase encoding gradients along at least one spatial dimension, and frequency encoding gradients along another spatial dimension, wherein k-space is sampled during an acquisition in a plurality of k-space lines oriented along a frequency encoding direction, wherein a number of k-space lines of the plurality of k-space lines are acquired in one shot, the imaging protocol comprising a plurality of shots, wherein a plurality of additional k-space lines are acquired in at least a subset of shots of the plurality of shots, such that movement of the object is detected throughout the imaging protocol.

2. The method of claim 1, wherein a position in k-space of the plurality of additional k-space lines acquired in each shot of the plurality of shots or in each of the subset of shots is constant or is varied over the shots or the subset of shots.

3. The method of claim 1, wherein the plurality of additional k-space lines are acquired in each shot of the plurality of shots.

4. The method of claim 1, wherein 2 to 16 additional k-space lines are acquired in each shot of the plurality of shots or in each of the subset of shots.

5. The method of claim 4, wherein 4 to 8 additional k-space lines are acquired in each shot of the plurality of shots or in each of the subset of shots.

6. The method of claim 1, wherein the acquisition of the plurality of additional k-space lines takes up 0.5% to 5% of a total acquisition time of the imaging protocol.

7. The method of claim 6, wherein the acquisition of the plurality of additional k-space lines takes up 1% to 3% of the total acquisition time of the imaging protocol.

8. The method of claim 1, wherein the k-space is sampled in a sampling order in which the one or the number of phase encoding gradients are changed incrementally from one k-space line to the next, with the exception of the plurality of additional k-space lines.

9. The method of claim 8, wherein the k-space is sampled in a linear, spiral, or radial sampling order, with the exception of the plurality of additional k-space lines.

10. The method of claim 1, wherein the plurality of additional k-space lines are disposed in a central region of the k-space, which is equivalent to a magnetic resonance image of low resolution.

11. The method of claim 10, wherein the magnetic resonance image of low resolution has a pixel size of ≥3 mm.

12. The method of claim 11, wherein the magnetic resonance image of low resolution has a pixel size of ≥4 mm.

13. The method of claim 1, further comprising acquiring a low-resolution scout image of the object.

14. The method of claim 1, wherein the imaging protocol uses a parallel imaging technique, in which one or all phase encoding directions are subsampled by a predetermined acceleration factor, and the magnetic resonance image dataset is acquired using a multi-channel coil array.

15. A method for generating a motion-corrected magnetic resonance image dataset of an object, the method comprising: receiving k-space data acquired using an imaging protocol; and estimating the motion-corrected magnetic resonance image dataset and rigid-body motion parameters for each shot or each of a subset of shots, the estimating comprising minimizing a data consistency error between the k-space data acquired in the imaging protocol and a forward model described by an encoding matrix, wherein the encoding matrix includes effects of rigid-body motion for each shot and Fourier encoding.

16. The method of claim 15, wherein the encoding matrix further includes subsampling, coil sensitivities, or subsampling and coil sensitivities of a multi-channel coil array.

17. The method of claim 16, further comprising receiving a low-resolution scout image of the object, wherein minimizing the data consistency error comprises: estimating the rigid-body motion parameters for each shot using the low-resolution scout image and k-space lines acquired in at least a subset of shots; and estimating the motion-corrected image using the estimated rigid-body motion parameters.

18. The method of claim 17, wherein the k-space lines are not used in the estimating of the motion-corrected image.

19. A magnetic resonance imaging apparatus comprising: a radio frequency controller configured to drive a radio frequency (RF) coil comprising a multi-channel coil array; a gradient controller configured to control gradient coils; a controller configured to control the radio frequency controller and the gradient controller to execute the imaging protocol, the execution of the imaging protocol comprising: use of the imaging protocol, in which spatial encoding is performed using phase encoding gradients along at least one spatial dimension, and frequency encoding gradients along another spatial dimension, wherein k-space is sampled during an acquisition in a plurality of k-space lines oriented along a frequency encoding direction, wherein a number of k-space lines of the plurality of k-space lines are acquired in one shot, the imaging protocol comprising a plurality of shots, and wherein a plurality of additional k-space lines are acquired in at least a subset of shots of the plurality of shots, such that movement of the object is detected throughout the imaging protocol.

20. In a non-transitory computer-readable storage medium that stores instructions executable by one or more processor to acquire a magnetic resonance image dataset of an object, the instructions comprising: using an imaging protocol in which spatial encoding is performed using phase encoding gradients along at least one spatial dimension, and frequency encoding gradients along another spatial dimension, wherein k-space is sampled during an acquisition in a plurality of k-space lines oriented along a frequency encoding direction, wherein a number of k-space lines of the plurality of k-space lines are acquired in one shot, the imaging protocol comprising a plurality of shots, wherein a plurality of additional k-space lines are acquired in at least a subset of shots of the plurality of shots, such that movement of the object is detected throughout the imaging protocol.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings illustrate various example methods and other example embodiments of various aspects of the invention.

[0046] FIG. 1 is a schematic representation of a magnetic resonance imaging (MRI) apparatus according to an embodiment;

[0047] FIG. 2 is a sequence diagram of a Turbo (Fast) spin-echo imaging protocol;

[0048] FIG. 3 is a schematic representation of three-dimensional k-space;

[0049] FIG. 4 is a schematic representation of aspects of a retrospective motion correction technique;

[0050] FIG. 5 is a schematic plan view of the phase-encode plane;

[0051] FIG. 6 is an exemplary diagram regarding a motion error depending on a size of a calibration region;

[0052] FIG. 7 is an exemplary diagram regarding a motion error depending on a sampling density.

[0053] Similar elements are designated with same reference signs in the drawing.

DETAILED DESCRIPTION

[0054] FIG. 1 schematically shows an embodiment of a magnetic resonance (MR) apparatus 1. The MR apparatus 1 has an MR data acquisition scanner 2 with a basic field magnet 3 that generates the constant magnetic field, a gradient coil arrangement 5 that generates the gradient fields, one or a number of radiofrequency (RF) antennas 7 for radiating and receiving radio-frequency signals, and a control computer 9 configured to perform an embodiment of a method. In FIG. 1, such sub-units of the magnetic resonance apparatus 1 are only outlined schematically. The radio-frequency antennas 7 may include a multi-channel coil array including at least two coils (e.g., the schematically shown coils 7.1 and 7.2), which may be configured either to transmit and receive radio-frequency signals or only to receive the radio frequency signals (e.g., MR signals).

[0055] In order to acquire MR data from an examination object U (e.g., a patient or a phantom), the examination object U is introduced on a bed B into the measurement volume of the MR data acquisition scanner 2. A slab S is an example of a 3D slab of the examination object, from which MR data may be acquired using a method according to an embodiment. The control computer 9 centrally controls the magnetic resonance apparatus 1, and may control the gradient coil arrangement 5 with a gradient controller 5′ and the radio-frequency antenna 7 with a radiofrequency transmit/receive controller 7′. The radio-frequency antenna 7 has multiple channels corresponding to the multiple coils 7.1, 7.2 of the coil arrays, in which signals may be transmitted or received. The radio-frequency antenna 7 together with its radiofrequency transmit/receive controller 7′ is responsible for generating and radiating (e.g., transmitting) a radio-frequency alternating field for manipulating nuclear spins in a region to be examined (e.g., in the slab S) of the examination object U. The control computer 9 also has an imaging protocol processor 15 that determines a reordering pattern according to an embodiment. A control unit 13 of the control computer 9 is configured to execute all the controls and computation operations required for acquisitions. Intermediate results and final results required for this purpose or determined in the process may be stored in a memory 11 of the control computer 9. The units shown may not necessarily be considered as physically separate units, but simply represent a subdivision into functional units, which may also be implemented by fewer physical units, or just one. A user may enter control commands into the magnetic resonance apparatus 1 and/or view displayed results (e.g., image data) from the control computer 9 via an input/output interface E/A. A non-transitory data storage medium 26 may be loaded into the control computer 9, and may be encoded with programming instructions (e.g., program code) that cause the control computer 9, and the various functional units thereof described above, to implement any or all embodiments of the method, as described herein.

[0056] FIG. 2 shows a sequence diagram illustrating a Fast Spin-Echo sequence, in which a 90° pulse is followed by a train of 180° refocusing pulses, as shown on the line names “RF”. For illustration purpose, this a 2D sequence; therefore, the RF pulses are transmitted concurrently with a slice select gradient G.sub.ss. The train of refocusing RF pulses leads to an echo train e1, e2, e3, . . . show in the signal row. Each of these echoes is used to acquire one k-space line 12 in the two-dimensional k-space 10, where echo e1 corresponds to line L1, echo e2 corresponds to line L2, echo e3 corresponds to k-space line L3, etc. In order to distribute the k-space lines 12 around the two-dimensional k-space 10, phase encode gradients GPE are used, which are incrementally changed for each echo. During acquisition, a readout gradient GRO is applied. The echoes e1, e2, . . . , e8 have a signal intensity that diminishes over the echo train due to T2 relaxation. The contrast of the final image is determined by the echo that is acquired in the center of k-space (e.g., e5 corresponding to L5). The time after the 90° pulse of this echo determines the effective echo time TE.sub.eff.

[0057] If a Fast Spin-Echo imaging protocol is performed in 3D, the slice select gradient may be applied only once during the 90° pulse in order to select one slab S. The further slice select gradients are replaced by a further phase encode gradient in a direction orthogonal to the 2D phase encode and the slice select gradient, so that phase encoding is performed in two spatial directions, leading to the distribution of the k-space lines 12 across a volume, rather than a plane 10.

[0058] FIG. 3 illustrates such three-dimensional k-space 14 having directions k.sub.x in readout direction, and directions k.sub.y and k.sub.z in the phase encode plane 20. A k-space line acquired during one echo is illustrated at 12. The k-space volume 14 is divided into a calibration region 16 and a periphery 18. Since the full acquisition in readout direction does not cost additional imaging time, usually the calibration region 16 will extend along the full length of the volume in readout direction k.sub.x. However, in the phase encode plane 20, which in this illustration, is oriented in the plane of the paper, the calibration region 16 covers only about 1/9 of the total square phase encode plane 20. The illustrated k-space line 12 is in the periphery 18.

[0059] A retrospective motion correction technique will now be illustrated with reference to FIG. 4. The mathematical model used is an extension of SENSE parallel imaging, as described in the above-cited paper by K. P. Pruessmann et al., with rigid-body motion parameters included into the forward model. The encoding operator E.sub.θ for a given patient motion trajectory θ relates the motion free image x to the acquired multi-channel k-space data s. FIG. 4 illustrates the mathematical components that contribute to the encoding at each shot. For each shot i, a position of the subject is described by a new set of six rigid-body motion parameters θ.sub.i that describe the 3D position of the object. Accordingly, the multi-channel k-space data s.sub.i for a given shot i may be related to the 3D image volume x through image rotations R.sub.i, image translations T.sub.i, coil sensitivity maps C, Fourier operator F, and under-sampling mask M.sub.i by the following formula 1

s.sub.i=E.sub.θ.sub.ix=M.sub.iF C T.sub.θ.sub.iR.sub.θ.sub.ix  [1]

[0060] Using an ultra-fast low-resolution scout scan, the method according to an embodiment creates an efficient method for directly estimating the motion trajectory θ, thus completely avoiding time-consuming alternating optimization between the image vector (formula 2) and the motion vector (formula 3):

[{circumflex over (x)}]=argmin.sub.x∥E.sub.{circumflex over (θ)}x−s∥.sub.2  [2]

[{circumflex over (θ)}]=argmin.sub.θ∥E.sub.θ{circumflex over (x)}−s∥.sub.2  [3]

[0061] Prior art methods require repeated updates of the coupled optimization variables x and θ, using the formulas 2 and 3. This may lead to convergence issues as updates of x and θ will be computed on inaccurate information. Further, the reconstruction is computationally demanding, as repeated updates of x (e.g., millions of imaging voxels) are to be provided.

[0062] If, however, a low-resolution scout image is acquired, the scout {tilde over (x)} approximates the motion-free image volume {circumflex over (x)}, and each motion state may be determined independently by minimizing the data consistency error of the forward model:

[{circumflex over (θ)}.sub.i]=argmin.sub.θ.sub.i∥E.sub.θ.sub.i{tilde over (x)}−s.sub.i∥.sub.2  [4]

[0063] For the final image reconstruction, the individual motion states from each shot are combined, and the motion-mitigated image is obtained from solving a standard least-squares problem:

[{circumflex over (x)}]=argmin.sub.x∥E.sub.{circumflex over (θ)}x−s∥.sub.2  [5]

[0064] This strategy completely avoids the difficult non-linear and non-convex joint optimization that contains millions of unknowns, as the strategy only considers six rigid body parameters per motion optimization, and the strategy does not require computationally costly full or partial updates to the image.

[0065] This framework may also be extended to Wave-CAIPI encoding. This method exploits available information in modern multi-channel receivers and may provide up to R=9-fold speedup for many important clinical contrasts. The sinusoidal gradients in Wave-encoding lead to a spatially varying phase that is applied along the read-out in hybrid space. The notation from the encoding model of formula [1] may be used

E.sub.θ.sub.i=M.sub.iF.sub.y,zP.sub.yzF.sub.xCT.sub.θ.sub.iR.sub.θ.sub.i  [6]

where the Fourier transform has been modified to contain the Wave point-spread-function P.sub.yz.

[0066] FIG. 5 shows a schematic plan view of the phase-encode plane 20 with each k-space line represented by one circle 21. The filled circles 22, 24 illustrate all k-space lines acquired in one shot. According to the imaging protocol, a linear echo train 22 is acquired, thereby sampling all k-space lines along one column in k.sub.z direction. In the same echo train, a number of additional k-space lines 24 are acquired from within the calibration region 16, which has an extent of L.sub.y×L.sub.z=5×5 in the phase-encode plane. In this example, one out of 4 k-space lines within the calibration region are sampled by the additional k-space lines 24, corresponding to a sampling density of 25%. This is because, as shown in the further data, even the smaller sampling density has proven to be sufficient.

[0067] Looking at FIG. 6, the standard deviation Std of the motion error for translational movement 32 (in mm) or rotational movement 30 (in °) is illustrated depending on the size of the calibration region 16, in number of samples. This data was obtained using two motion-free MPRAGE scans that were combined to construct motion data sets that mimicked a single movement from the first motion state to the second motion state in the middle of the imaging protocol, and where the movement included translational and rotational motion. The MPRAGE scans were combined to reconstruct k-space data with various sizes of the calibration region, as well as sampling density in the calibration region (see FIG. 7). FIG. 7 shows the standard deviation of the motion error (e.g., translational 34 and rotational 36) for a fixed size of the calibration region of L.sub.y×L.sub.z=17×17, so that a maximum number of 289 additional k-space lines may be acquired within the calibration region. The overall size of the data set was 256×256 k-space lines in the phase-encode plane. As take, from FIG. 6, both the rotational error 30 as well as the translational error 32 increased sharply when the size of the calibration region was reduced to below 10×10. However, an increase beyond 20×20 and, for example, 30×30 gave little further improvement. Thus, the calibration region may cover between about 0.15% to 0.6% or 1.4% of k-space and still yield a significant improvement in motion correction. As taken form FIG. 7, the standard deviation of the motion error was remarkably independent from the sampling density, and even a small number of only 4 calibration samples within the calibration region provides good results. Thus, a sampling density of about 1% to 3.33% (e.g., 1.33% to 2%) corresponding to an acceleration factor of between about 30 to 100 (e.g., 50 to 75) may be used in the method of the present embodiments.

[0068] While the present disclosure has been described in detail with reference to certain embodiments, the present disclosure is not limited to those embodiments. In view of the present disclosure, many modifications and variations would present themselves, to those skilled in the art without departing from the scope of the various embodiments of the present disclosure, as described herein. The scope of the present disclosure is, therefore, indicated by the following claims rather than by the foregoing description. All changes, modifications, and variations coming within the meaning and range of equivalency of the claims are to be considered within the scope.

[0069] It is to be understood that the elements and features recited in the appended claims may be combined in different ways to produce new claims that likewise fall within the scope of the present disclosure. Thus, whereas the dependent claims appended below depend from only a single independent or dependent claim, it is to be understood that these dependent claims may, alternatively, be made to depend in the alternative from any preceding or following claim, whether independent or dependent, and that such new combinations are to be understood as forming a part of the present specification.