PROPELLER MR IMAGING WITH ARTEFACT SUPPRESSION

Abstract

The invention relates to a method of MR imaging of a body (10) of a patient. It is an object of the invention to provide a method that enables efficient compensation of image artefacts in combination with PROPELLER imaging. The invention proposes to combine k-space blades in image space, and not in k-space like in conventional PROPELLER imaging. Local image artefacts are detected and corrected in single-blade MR images. The artefact detection and correction in the image domain prior to combining the single-blade MR images into a final MR image results in an improved image quality by better suppression of local artefacts and, thus, an increased signal-to-noise. Moreover, the invention relates to a MR device (1) and to a computer program for a MR device (1).

Claims

1. A method of magnetic resonance (MR) imaging of a body of a patient placed in the examination volume of a MR device, the method comprising the steps of: a) generating MR signals by subjecting at least a portion of the body to a PROPELLER MR imaging sequence of at least one RF pulse and switched magnetic field gradients; b) acquiring the MR signals as a plurality of k-space subsets, each k-space subset covering a different portion of k-space, wherein at least a part of a central portion of k-space is acquired for each k-space subset; with the k-space subsets being k-space blades that are rotated about the centre of k-space, so that the total acquired data set of MR signals spans a circle in k-space, c) reconstructing single-subset MR images from each k-space subset; d) identifying image regions containing artefacts are in the single-subset MR images and deriving weighting factors from the spatial distribution of image artefacts in the single-subset images the weighting factors reducing weighting of the voxel values of the single-subset images in the image regions containing artefacts; and e) combining the single-subset MR images into a final MR image by weighted superposition using said weighting factors of the single-subset MR images.

2. A method of MR imaging of a body of a patient placed in the examination volume of a MR device, the method comprising the steps of: a) generating MR signals by subjecting at least a portion of the body to a PROPELLER MR imaging sequence of at least one RF pulse and switched magnetic field gradients; b) acquiring the MR signals as a plurality of k-space subsets, each k-space subset covering a different portion of k-space, wherein at least a part of a central portion of k-space is acquired for each k-space subset; with the k-space subsets being k-space blades that are rotated about the centre of k-space, so that the total acquired data set of MR signals spans a circle in k-space, d) reconstructing single-subset low resolution MR images from central k-space data of each k-space subset; e) identifying image regions containing artefacts in the single-subset low resolution MR images and deriving weighting factors from the spatial distribution of image artefacts in the single-subset images the weighting factors reducing weighting to the voxel values of the single-subset images in the image regions containing artefacts; f) combining the single-subset low resolution MR images into a low-resolution MR image by the weighted superposition of the single-subset MR images according to said weighting factors; g) combining the k-space subsets into a full k-space dataset; h) combining the full k-space data set with a k-space representation of the low-resolution MR image into a combined full k-space data set; and i) reconstructing a final image from the combined full k-space dataset.

3. The method of claim 1, wherein the image regions containing artefacts are identified by a consistency analysis of the single-subset MR images.

4. The method of claim 1, wherein the weighted superposition is computed by solving a linear inverse problem.

5. The method of claim 1, comprising the step of estimating and correcting motion-induced displacements and phase errors in the k-space subsets prior to reconstructing the single-subset MR images.

6. A magnetic resonance (MR) device for carrying out the method claimed in claim 1, which MR device includes at least one main magnet coil for generating a uniform, steady magnetic field B0 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 a body of a patient 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 for reconstructing MR images from the received MR signals, wherein the MR device is configured to perform the following steps: a) generating MR signals by subjecting at least a portion of the body to a PROPELLER MR imaging sequence of at least one RF pulse and switched magnetic field gradients; b) acquiring the MR signals as a plurality of k-space subsets, each k-space subset covering a different portion of k-space, wherein at least a part of a central portion of k-space is acquired for each k-space subset; with the k-space subsets being k-space blades that are rotated about the centre of k-space, so that the total acquired data set of MR signals spans a circle in k-space, c) reconstructing single-subset low resolution MR image from central k-space data of each k-space subset; d) identifying image regions containing artefacts are in the single-subset low resolution MR images and derive weighting factors from the spatial distribution of image artefacts in the single-subset images the weighting factors reducing weighting of the voxel values of the single-subset images in the image regions containing artefacts; e) combining the single-subset low resolution MR images into low-resolution MR image by the weighted superposition of the single-subset MR images according to said weighting factors; f) combining the k-space subsets into a full k-space dataset; g) combining the full k-space data set with a k-space representation of the low-resolution MR image into a combined full k-space data set; and h) reconstructing a final image from the combined full k-space dataset.

7. A computer program to be run on a magnetic resonance (MR) device, which computer program comprises instructions for: a) generating MR signals by subjecting at least a portion of the body to a PROPELLER MR imaging sequence of at least one RF pulse and switched magnetic field gradients; b) acquiring the MR signals as a plurality of k-space subsets, each k-space subset covering a different portion of k-space, wherein at least a part of a central portion of k-space is acquired for each k-space subset; with the k-space subsets being k-space blades that are rotated about the centre of k-space, so that the total acquired data set of MR signals spans a circle in k-space, c) reconstructing single-subset low resolution MR image from central k-space data of each k-space subset; d) identifying image regions containing artefacts are in the single-subset low resolution MR images and derive weighting factors from the spatial distribution of image artefacts in the single-subset images the weighting factors reducing weighting of the voxel values of the single-subset images in the image regions containing artefacts; e) combining the single-subset low resolution MR images into low-resolution MR image by the weighted superposition of the single-subset MR images according to said weighting factors; f) combining the k-space subsets into a full k-space dataset; g) combining the full k-space data set with a k-space representation of the low-resolution MR image into a combined full k-space data set; and h) reconstructing a final image from the combined full k-space dataset.

8. A magnetic resonance (MR) device for carrying out the method claimed in claim 1, which MR device includes at least one main magnet coil for generating a uniform, steady magnetic field B0 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 a body of a patient 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 for reconstructing MR images from the received MR signals, wherein the MR device is configured to perform the following steps: a) generating MR signals by subjecting at least a portion of the body to a PROPELLER MR imaging sequence of at least one RF pulse and switched magnetic field gradients; b) acquiring the MR signals as a plurality of k-space subsets, each k-space subset covering a different portion of k-space, wherein at least a part of a central portion of k-space is acquired for each k-space subset; with the k-space subsets being k-space blades that are rotated about the centre of k-space, so that the total acquired data set of MR signals spans a circle in k-space, c) reconstructing single-subset MR images from each k-space subset; d) identifying image regions containing artefacts are in the single-subset MR images and deriving weighting factors from the spatial distribution of image artefacts in the single-subset images the weighting factors reducing weighting of the voxel values of the single-subset images in the image regions containing artefacts; and e) combining the single-subset MR images into a final MR image by weighted superposition using said weighting factors of the single-subset MR images.

9. A computer program to be run on a magnetic resonance (MR) device, which computer program comprises instructions for: a) generating MR signals by subjecting at least a portion of the body to a PROPELLER MR imaging sequence of at least one RF pulse and switched magnetic field gradients; b) acquiring the MR signals as a plurality of k-space subsets, each k-space subset covering a different portion of k-space, wherein at least a part of a central portion of k-space is acquired for each k-space subset; with the k-space subsets being k-space blades that are rotated about the centre of k-space, so that the total acquired data set of MR signals spans a circle in k-space, c) reconstructing single-subset MR images from each k-space subset; d) identifying image regions containing artefacts are in the single-subset MR images and deriving weighting factors from the spatial distribution of image artefacts in the single-subset images the weighting factors reducing weighting of the voxel values of the single-subset images in the image regions containing artefacts; and e) combining the single-subset MR images into a final MR image by weighted superposition using said weighting factors of the single-subset MR images.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0022] 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:

[0023] FIG. 1 shows a MR device for carrying out the method of the invention;

[0024] FIG. 2 schematically illustrates the PROPELLER acquisition scheme of the invention;

[0025] FIG. 3 shows single-blade MR images containing local image artefacts;

[0026] FIG. 4 shows a block diagram illustrating one embodiment of the method of the invention;

[0027] FIG. 5 shows a diagram of k-space illustrating the keyhole approach of the invention;

[0028] FIG. 6 shows an example of a XI map for detecting SENSE artefacts.

DETAILED DESCRIPTION OF THE EMBODIMENTS

[0029] With reference to FIG. 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 B.sub.0 is created along a z-axis through an examination volume. The device further comprises a set of (1.sup.st, 2.sup.nd, and—where applicable—3.sup.rd order) shimming coils 2′, wherein the current flow through the individual shimming coils of the set 2′ is controllable for the purpose of minimizing B.sub.0 deviations within the examination volume.

[0030] 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.

[0031] More specifically, a gradient amplifier 3 applies current pulses or waveforms 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 body 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, together with any applied magnetic field gradients, achieve a selected manipulation of nuclear magnetic resonance signals. 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 body RF coil 9.

[0032] For generation of MR images of limited regions of the body 10 or for scan acceleration by means of parallel imaging, 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.

[0033] The resultant MR signals are picked up by the body 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.

[0034] A host computer 15 controls the shimming coils 2′ as well as 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 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.

[0035] Ultimately, the digital raw image data are reconstructed into an image representation by a reconstruction processor 17 which applies a Fourier transform or other appropriate reconstruction algorithms, such as SENSE or GRAPPA. 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.

[0036] FIG. 2 illustrates the k-space sampling of PROPELLER MR imaging according to the invention. As shown in the left illustration of FIG. 2, nine k-space subsets (blades) 21-29 are acquired. Each blade 21-29 covers a different portion of k-space, wherein a central circular portion 30 of k-space is acquired for each blade 21-29. The blades 21-29 are rotated about the center of k-space, so that the total acquired MR data set spans a circle in k-space. In the right illustration of FIG. 2 a single k-space blade 21 is shown which is acquired using SENSE. The orientation of the phase encoding direction and the readout direction relative to the blade orientation is maintained for all rotation angles of the k-space blades 21-29.

[0037] FIG. 3 shows examples of eight single-subset (single-blade) MR images (one MR image is reconstructed from each blade) containing image artefacts, as indicated by the arrows. The artefacts have a local character which means that the larger part of each single-blade MR image is correct. The artefacts are located at different positions in each single-blade MR image. Hence, for a single location in the anatomy a majority of the single-blade MR images will have the correct pixel values.

[0038] According to the invention, the single-blade MR images are combined into a final MR image in image space in order to account for the local character of the image artefacts. The single-blade MR images can be combined in image space by solving a linear inverse problem. The inverse problem can be formulated as:

min.sub.pΣ.sub.i=1.sup.N∥p.sub.blade,i−A.sub.ip∥.sup.2

[0039] Wherein N is the number of blades, p.sub.blade,i, is the vector containing the single-blade MR image pixel values, p is the vector containing the final MR image pixel values and A.sub.i is a sparse matrix reflecting the relation between the final MR image pixel values and the single-blade MR image pixel values. The A matrices can be derived using the knowledge of the k-space positions of each acquired blade. In other words, A.sub.i reflects the blade angulations and resolutions. The inverse problem is linear and, thus, convex which means that it has a unique solution and can be solved by any least squares algorithm. There are several ways of detecting the positions of the local artefacts in the single-blade MR images. Two possible techniques will be explained in detail below. Under the assumption that the information of possibly defective voxels is known for every single-blade MR image in the image domain, it can be easily incorporated into the inverse problem by extending it into a weighted inverse problem:

min.sub.pΣ.sub.i=1.sup.N∥W.sub.ip.sub.blade,i−W.sub.iA.sub.ip∥.sup.2

[0040] Wherein W is a diagonal weight matrix that assigns a low weight to those equations containing defective single-blade voxels.

[0041] In the afore-described embodiment, the final MR image p is directly computed from the complete single-subset MR images p.sub.blade,i. In an alternative embodiment, which is illustrated in FIGS. 4 and 5, a hybrid scheme of combining the blades is applied such that the computational effort is significantly reduced.

[0042] In step 41, the k-space blades are acquired as shown in FIG. 1. Motion-induced displacements and phase errors in the blades are detected and corrected in step 42 like in conventional PROPELLER imaging. Low-resolution single-blade MR images p.sub.blade,i are reconstructed only from the central k-space data (portion 30, see FIG. 1) of the blades in step 43. The motion-corrected low-resolution single-blade MR images p.sub.blade,i are regridded to a common grid. Once this is done, the inversion problem for weighted combination of the low-resolution single-blade MR images p.sub.blade,i into a low-resolution MR image p in step 44 can be written as:

min.sub.pΣ.sub.i=1.sup.N∥W.sub.ip.sub.blade,i−W.sub.iA.sub.ip∥.sup.2

[0043] This inverse problem can be solved per voxel. There is no coupling between individual voxels as W.sub.i is a diagonal matrix. The solution may be derived simply by computing the weighted average of the low-resolution single-blade MR images:

[00001]$p_k=\underset{i=1}{\overset{N}{.Math.}}.Math.\frac{w_{\mathrm{ik}}.Math.p_{\mathrm{blade},i}}{\underset{j=1}{\overset{N}{.Math.}}.Math.w_{\mathrm{jk}}}$

[0044] This will result in an artefact-free low resolution MR image p.sub.k. However, the final MR image should be a high-resolution MR image. In order to achieve this, the acquired k-space blades are combined in k-space in step 45, again like in conventional PROPELLER reconstruction. In step 46, a k-space representation of the low-resolution MR image p.sub.k (covering only the central portion of k-space) is combined with the full k-space data set generated in step 45. This way of combining the data corresponds to a key-hole technique as illustrated in FIG. 5. The central k-space portion 51 of the full k-space data as acquired, motion-corrected and combined in steps 41, 42, and 43 is replaced by the k-space representation of the low-resolution MR image computed in step 44. The peripheral k-space data 52 are preserved. The final high-resolution MR image is reconstructed from this combined k-space dataset. The result is a high-resolution MR image with a reduced artefact level and improved SNR.

[0045] A key feature of the scheme of the invention is the ability to detect the image regions within the single-blade MR images where artefacts are located. The image regions containing artefacts can be identified by a consistency analysis of the single-blade MR images. Two methods for detecting the defective image regions are described in the following.

[0046] The first option is to use a so-called XI map. A XI map is computed per single-blade MR image by projecting the reconstructed single-blade MR image back onto the folded image space (i.e. the image space to which the single-coil k-space blades are reconstructed prior to SENSE unfolding). Then the mean squared error of the difference between the projection and the folded single-coil/single-blade MR images m.sub.ij is computed:

XI.sub.blade,i=Σ.sub.j=1.sup.C∥m.sub.ij−S.sub.ijp.sub.blade,i∥.sup.2

[0047] Wherein C is the number of RF coils 11, 12, 13 used in the SENSE acquisition of the k-space blades, S.sub.ij is the SENSE encoding matrix of blade i. The XI map will “highlight” the image regions containing any inconsistencies, e.g. SENSE artefacts resulting from inaccurate coil sensitivity maps used in SENSE unfolding (see FIG. 6) or flow artefacts. This method works well in cases in which the number of coils that are sensitive in a given image region exceeds the effective acceleration factor, i.e. there is redundant image data. A benefit of this method is that the information of artefact positions is available at the resolution of the individual k-space blade, i.e. on the grid of the single-blade MR image which has a high resolution in the readout direction. A drawback of this method is that not all types of artefacts may be detected equally well.

[0048] FIG. 6 illustrates an example of a XI map and SENSE artefacts in a head scan. The left image is a SENSE reconstructed MR image containing SENSE artefacts (indicated by arrows). The right image is the corresponding XI map “highlighting” the locations of SENSE artefacts.

[0049] Another option is to use the low-resolution single-blade MR images (reconstructed from the centre portion 30 of k-space of each k-space blade). To determine which single-blade MR image contains defective voxels at a given image position, it should first be determined what the “true” voxel value must be at that position. It is known that in almost all cases the artefacts are located in different positions per single-blade MR image, which means that per image position the majority of the single-blade MR images have the correct voxel value. Hence finding the ‘true’ value can be achieved by solving the following simple problem:

min.sub.pΣ.sub.i=1.sup.N∥p.sub.blade,i−p∥.sup.1

[0050] This problem can be solved efficiently using a weighted least squares solving algorithm. The output will be the value ofp and a matrix of weights denoting which single-blade MR image contains a defective voxel value indicating an image artifact. These weights may be compared to the XI maps (see above) or may be directly used in the weighted combination of the single-blade MR images. A benefit of this method is that all artefacts are detectable in principle. A drawback is that the information is available only at low resolution. The consequence is that possibly more down-weighting will be applied during the combination of the single-blade MR images resulting in a certain amount of blurring in the final MR image.

[0051] If there are many defective voxels in the single-blade MR images, the weighted inverse problem (see above) may become ill-conditioned. To ensure that the solution represents the true anatomy, additional regularization may be needed for stabilizing the problem. This can be formulated, for example, as:

min.sub.pΣ.sub.i=1.sup.N∥W.sub.ip.sub.blade,i−W.sub.iA.sub.ip∥.sup.2+W.sub.reg∇p∥.sup.2

[0052] Wherein W.sub.reg is a weight matrix based on the knowledge of the image regions containing artefacts. ∇p is the set of spatial derivates of the solution p. If an image region is corrupted in one of the single-blade MR images, the weight is made non-zero. This effects that the solution is of lower resolution in those image regions where information is missing (because of artefacts in the single-blade MR images). In other words, the artefact level is reduced at the cost of local blurring.