阅读说明：本技术 使用具有软运动门控的3d径向或螺旋采集的mr成像 (MR imaging using 3D radial or helical acquisition with soft motion gating ) 是由 G·贝克 C·克里希纳 S·纳加拉杰 J·M·佩特斯 于 2020-03-13 设计创作，主要内容包括：本发明涉及对对象(10)进行MR成像的方法。本发明的目的是使用3D径向或螺旋采集方案实现MR成像,从而在存在运动的情况下提供增强的图像质量。所述方法包括以下步骤：通过使所述对象(10)经受包括RF脉冲和切换的磁场梯度的成像序列来生成MR信号；使用具有对k空间的中心部分(26)的过采样的3D径向或螺旋采集方案来采集所述MR信号；在采集所述MR信号期间检测所述对象(10)的运动引起的移位(d)和/或变形,并且将所采集的MR信号中的每一个分配给运动状态；根据在k空间的所述中心部分(26)中加权的所述MR信号来重建MR图像,其中,较强的加权(W、30)被应用于在较频繁的运动状态中采集的MR信号,而较弱的加权(W、31、32)被应用于在较不频繁的运动状态中采集的MR信号。此外,本发明涉及一种MR设备(1)和一种用于MR设备(1)的计算机程序。(The invention relates to a method of MR imaging of an object (10). It is an object of the present invention to enable MR imaging using a 3D radial or helical acquisition scheme, providing enhanced image quality in the presence of motion. The method comprises the following steps: generating MR signals by subjecting the object (10) to an imaging sequence comprising RF pulses and switched magnetic field gradients; acquiring the MR signals using a 3D radial or helical acquisition scheme with oversampling of a central portion (26) of k-space; detecting motion-induced shifts (d) and/or deformations of the object (10) during the acquisition of the MR signals and assigning each of the acquired MR signals to a motion state; reconstructing an MR image from the MR signals weighted in the central portion (26) of k-space, wherein stronger weights (W, 30) are applied to MR signals acquired in more frequent motion states and weaker weights (W, 31, 32) are applied to MR signals acquired in less frequent motion states. Furthermore, the invention relates to a MR device (1) and to a computer program for a MR device (1).)

1. Method of MR imaging of an object (10) positioned in an examination volume of a MR device (1), comprising the steps of:

generating MR signals by subjecting the object (10) to an imaging sequence comprising RF pulses and switched magnetic field gradients;

acquiring the MR signals using a 3D radial or helical acquisition scheme with oversampling of a central portion (26) of k-space;

detecting motion-induced shifts (d) and/or deformations of the object (10) during the acquisition of the MR signals and assigning each of the acquired MR signals to a motion state; and is

MR images are reconstructed from MR signals weighted in the central portion (26) of k-space, wherein stronger weights (W, 30) are applied to MR signals acquired in more frequent motion states and weaker weights (W, 31, 32) are applied to MR signals acquired in less frequent motion states.

2. The method of claim 1, wherein the MR signals are acquired using a star-stacked or helical-stacked acquisition scheme, or a 3D KOOSH sphere or a 3D Helix acquisition scheme.

3. The method according to claim 1 or 2, wherein each of the motion states corresponds to one of a plurality of consecutive ranges of motion-induced displacement (d) and/or deformation of the object (10).

4. Method of claim 3, wherein the frequency of occurrence of each motion state is determined based on a histogram reflecting the number of MR signals acquired for each motion state.

5. Method of any one of claims 1-4, wherein more weighting (W) is applied to MR signals in the central portion (26) of k-space and less or no weighting is applied to MR signals in the peripheral portion (27) of k-space.

6. Method of any one of claims 1-5, wherein the weighting (W) applied to the MR signals during reconstruction of the MR image is derived from a user-specified gating percentage.

7. Method of any one of claims 1-6, wherein the MR signals are acquired in parallel via a plurality of RF receive coils (11, 12, 13) having different spatial sensitivity profiles.

8. The method of any one of claims 1-7, wherein the MR image is reconstructed using a compressed sensing or parallel image reconstruction algorithm like SENSE.

9. Method of any one of claims 1-8, wherein a time series of MR images is reconstructed from the acquired MR signals.

10. Method of any one of claims 1-9, wherein the MR signals are acquired as radial or helical k-space profiles, wherein the rotation angle of the k-space profiles is incremented according to the golden angle scheme during acquisition of successive k-space profiles.

11. MR device comprising at least one main magnet coil (2) for generating a homogeneous, stable magnetic field B in an examination volume, a plurality of gradient coils (4, 5, 6), at least one RF coil (9), a control unit (15) and a reconstruction unit (17)₀A plurality 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 subject (10) positioned in the examination volume, a control unit for controlling a temporal succession of RF pulses and switched magnetic field gradients, a reconstruction unit for reconstructing an MR image from the received MR signals, wherein the MR device (1) is arranged for performing the steps of:

generating MR signals by subjecting the object (10) to an imaging sequence comprising RF pulses and switched magnetic field gradients;

acquiring the MR signals using a 3D radial or helical acquisition scheme with oversampling of a central portion (26) of k-space;

detecting motion-induced shifts (d) and/or deformations of the object (10) during the acquisition of the MR signals and assigning each of the MR signals to a motion state; and is

MR images are reconstructed from MR signals weighted in the central portion (26) of k-space, wherein stronger weights (W, 30) are applied to MR signals acquired in more frequent motion states and weaker weights (W, 31, 32) are applied to MR signals acquired in less frequent motion states.

12. Computer program to be run on a MR device, the computer program comprising instructions for:

generating an imaging sequence comprising RF pulses and switched magnetic field gradients;

acquiring MR signals using a 3D radial or helical acquisition scheme with oversampling of a central portion (26) of k-space;

detecting motion-induced shifts (d) and/or deformations from an object (10) during acquisition of the MR signals and assigning each of the MR signals to a motion state; and is

Reconstructing an MR image from the MR signals weighted in the central portion (26) of k-space, wherein stronger weights (W, 30) are applied to MR signals acquired in more frequent motion states and weaker weights (W, 31, 32) are applied to MR signals acquired in less frequent motion states.

Technical Field

The present invention relates to the field of Magnetic Resonance (MR) imaging. The invention relates to a method of MR imaging of an object placed in an examination volume of a MR device. The invention also relates to an MR device and a computer program to be run on an MR device.

Background

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

Generally, according to the MR method, the body of the patient to be examined is arranged in a strong and homogeneous magnetic field B₀Middle, magnetic field B₀While defining the measurement-related axis (typically the z-axis) of the coordinate system. Magnetic field B₀Different energy levels are generated for individual nuclear spins depending on the magnetic field strength, which can be excited (spin resonance) by applying an electromagnetic alternating field (RF field) of a defined frequency, the so-called larmor frequency or MR frequency. From a macroscopic point of view, the distribution of the individual nuclear spins produces a global magnetization which can be induced by applying electromagnetic pulses (RF pulses) of suitable frequencyThe deviation from the equilibrium state occurs while the corresponding magnetic field B1 of the RF pulse extends perpendicular to the z-axis, so that the magnetization performs a precession around the z-axis. Precession describes the surface of a cone whose aperture angle is known as the flip angle. The magnitude of the flip angle depends on the strength and duration of the applied electromagnetic pulse. In the case of a so-called 90 ° pulse, the magnetization is deflected from the z-axis into the transverse plane (flip angle 90 °).

After termination of the RF pulse, the magnetization relaxes back to the original equilibrium state in which the magnetization in the z-direction is re-established with a first time constant T1 (spin lattice or longitudinal relaxation time) and the magnetization in the direction perpendicular to the z-direction relaxes with a second, shorter time constant T2 (spin-spin or transverse relaxation time). The transverse magnetization and its change can be detected by means of a receiving RF coil which is arranged and oriented within the examination volume of the MR device such that the change in magnetization is measured in a direction perpendicular to the z-axis. The decay of the transverse magnetization is accompanied by a phase loss after RF excitation due to local magnetic field inhomogeneities, which facilitates the transition from an ordered state with the same signal phase to a state in which all phase angles are uniformly distributed. The dephasing can be compensated by means of refocusing RF pulses (e.g. 180 deg. pulses). This produces an echo signal (spin echo) in the receive coil.

In order to achieve spatial resolution in the body, a time-varying magnetic field gradient extending along the three main axes is superimposed on the homogeneous magnetic field B₀Resulting in a linear spatial dependence of the spin resonance frequency. The signals picked up in the receiving coils then contain components of different frequencies, which can be associated with different locations in the body. The signal data obtained via the receiving coils correspond to the spatial frequency domain and are referred to as k-space data. k-space data typically comprises a plurality of lines of acquisitions of different phase encodings. Each line is digitized by collecting a large number of samples. A set of k-space data is converted into an MR image by means of a fourier transformation.

In the known so-called three-dimensional (3D) star-stacked (stack-of-stages) acquisition schemes (see, for example, WO2013/159044a1), a number of spatially non-selective or plate-selective approaches are appliedSexual RF excitation, one or more MR signals (e.g., gradient echo signals) are acquired after each excitation, where each MR signal represents one k-space profile. MR signals are acquired as radial k-space contours from a plurality of parallel slices. The slices are arranged at different positions in k-space along the slice direction. In the slice direction (e.g. k)_zDirection) is performed, while the MR signal is along the surrounding center (k) within each individual slice_x＝k_y0) radial "spoke" acquisition of rotation. This forms a cylindrical k-space consisting of stacked disks ("star stack"). Technically, this is achieved by simultaneously generating magnetic field gradients in the in-plane direction of the slice and modulating their amplitude. Different schemes can be used to select the temporal order of the k-space contour acquisition steps. For example, all phase encoding steps along the slice direction can be acquired sequentially before acquiring the k-space profiles at different angular positions (rotation angles). This may ensure that the period of cartesian sampling is kept short, which results in high data consistency within the slice stack and preserves the general motion robustness of radial sampling for the star stacking method. The cartesian phase encoding step may be performed from the central slice towards the k-space periphery (from the center outwards), or from-k_z，maxTo + k_z，maxIs performed in linear order. For angular ordering, the imaging sequence can use equidistant angular sampling with multiple interleaving, or also a so-called golden angular scheme. In the equidistant scheme, the angular distance (i.e. the increment of the rotation angle of the radial k-space profile) is 180 °/n depending on Δ Φ_{General assembly}Is calculated, wherein n_{General assembly}Is the total number of spokes. It may be beneficial to use multiple interlaces (or "rotations") to acquire the spokes, since the interlaces reduce temporal coherence in k-space. Thus, motion inconsistencies may spread out in k-space and artifacts are attenuated. In the golden angle approach, the rotation angle of the k-space profile increases by 111.25 ° each time, which corresponds to 180 ° times the golden ratio. Thus, the spokes of a subsequent sample always add supplemental information while filling the largest gap within the set of spokes of the previous sample. Thus, any set of subsequently acquired spokesThe strips will cover k-space substantially uniformly, which for example enables reconstruction of temporal sub-frames and makes the golden angle approach well suited for dynamic (4D) imaging studies.

Similarly, in a helical stack acquisition scheme, also known, one or more MR signals representing a helical k-space profile are acquired after each non-selective or slab-selective RF excitation. Like in the star-stacked approach, the slices are also arranged at different positions in k-space along the slice direction in which the standard cartesian phase encoding is performed, while within each individual slice along the source point at the center of k-space (k-space)_x＝k_y0) acquires MR signals.

The aforementioned 3D radial star stacking and helical stacking schemes offer several promising advantages for clinical 3D and 4D MR imaging, such as high motion robustness and benign aliasing artifacts, in particular in combination with the golden angular distribution of the k-space profile.

However, despite this inherent motion robustness, the acquired MR images may still be compromised by motion-induced signal fluctuations, as long as no additional measures for motion compensation are applied.

Motion compensation methods for star stacked imaging are known in the art.

For example, gating techniques have been developed that only accept MR signal data acquired within a certain predefined respiratory gating window. To cope with potential drift problems, multi-gated window methods have been proposed that use individual motion states (binning) instead of one predefined gated window (called PAWS, see US 7039451B 1). Each of the motion states corresponds to one of a plurality of consecutive ranges of motion-induced displacement of the body under examination. The final MR image in PAWS is reconstructed from the MR signal data belonging to the motion state for which the complete set of MR signal samples was first acquired.

Disclosure of Invention

As can be readily appreciated from the foregoing, there is a need for improved motion compensation in 3D or 4D radial or helical MR imaging techniques. It is therefore an object of the present invention to enable MR imaging using a 3D radial or helical acquisition scheme that provides enhanced image quality in the presence of 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 following steps:

generating MR signals by subjecting the subject to an imaging sequence comprising RF pulses and switched magnetic field gradients;

acquiring the MR signals using a 3D radial or helical acquisition scheme with oversampling of a central portion of k-space;

detecting motion-induced shifts and/or deformations of the object during the acquisition of the MR signals and assigning each of the acquired MR signals to a motion state;

MR images are reconstructed from MR signals weighted in the central portion of k-space, wherein stronger weighting is applied to MR signals acquired in more frequent motion states and weaker weighting is applied to MR signals acquired in less frequent motion states.

According to the invention, a three-dimensional radial or helical acquisition (preferably a star-stacked or helical stacked acquisition) is performed with oversampling of the central portion of k-space. The radial or helical k-space sampling density is higher at k-space positions closer to the k-space center (k 0), while the sampling density is lower in slices located further away from the k-space center. Due to the oversampling, the sampling density is higher in the central part of k-space than required by the nyquist criterion. The increased density of k-space samples near the center of k-space has resulted in a reduced level of streak artifacts, while the total scan time can be kept at a minimum. According to the invention, the oversampling of the central k-space portion achieves a weighting of the MR signals in the central k-space for the purpose of a more further reduction of motion artifacts. It is noted that this can be achieved by a relatively strong weighting of the MR signals in the central k-space compared to the weighting of the MR signals in the peripheral k-space.

The gist of the invention is to detect the motion of an object under examination (patient), for example by means of an external motion sensor (respiratory belt, camera, etc.) or by means of intrinsic motion detection, for example based on the correlation of k-space contours acquired in a time series. During the acquisition of the MR signals, a displacement and/or deformation of the object is determined in such a way that each of the acquired MR signals can be assigned to a motion state. Each of the motion states may preferably be defined as one of a plurality of consecutive ranges corresponding to motion-induced displacement and/or deformation of the object. Motion detection can also be done based on elastic registration of low resolution images reconstructed from MR data from the central oversampled region of k-space. The invention proposes a 3D soft-gating method by weighting the MR signals in the oversampled central part of k-space, wherein a stronger weighting is applied to MR signals acquired in more frequent motion states and a weaker weighting is applied to MR signals acquired in less frequent motion states. In this way, MR signals acquired in the more common motion states of the object under examination contribute more to the reconstructed MR image, whereas MR signals assigned to more distant motion outliers have a weaker contribution. In other words, MR signals acquired in case the object under examination exhibits its most frequently assumed position are given a stronger weight, whereas MR signals acquired from objects in less frequently exhibited positions are suppressed in the reconstructed MR image. Preferably, the applied weighting factors vary smoothly according to the detected shift/deformation ("soft-gating"). The result of this method is that MR images reconstructed from 3D radial or helical acquisitions have a significantly reduced level of artifacts in the presence of motion of the object under examination.

In a practical embodiment, the frequency of occurrence of each motion state, which is the basis for the MR signal weighting in the reconstruction step, is derived from a histogram which is established during or after the MR signal acquisition and reflects the number of MR signals acquired per motion state. The weighting factors for reconstruction can be easily derived from the histogram, taking into account the user-specified gating percentage. The gating percentage defines the proportion of the MR signal that is suppressed by weighting as a global parameter that can be tuned by the user as desired. When determining the weighting factors, the conformance to the nyquist criterion must be considered for the central part of k-space in order to avoid aliasing artifacts. This can be achieved by deriving a variable oversampling rate from a user-specified gating percentage.

In a preferred embodiment, more weighting (i.e. a wider range of weighting factors) is applied to the MR signals in the central portion of k-space than in the peripheral portion of k-space. Less pronounced weighting in the peripheral k-space has the effect that streak artifacts from k-space sub-sampling can be avoided.

According to another preferred embodiment, the invention is applied in connection with parallel imaging. MR signals are acquired in parallel via a plurality of RF receive coils having different spatial sensitivity profiles. The MR images are correspondingly reconstructed using a parallel image reconstruction algorithm, such as (non-cartesian) SENSE. The present invention can also be combined with compressive sensing (see "Compressed sensing MRI" by m.lustig et al, IEEE signal processing magazine,2008, vol.25, No.2, pages 72-82). If it is combined with SENSE or compressive sensing, more pronounced weighting can be applied according to the invention to the MR signals in the peripheral k-space portion. These sparse sampling embodiments of the present invention may be implemented by artificial intelligence techniques. Streak artifacts can be avoided even if significant portions of the acquired MR signals assigned to less frequent motion states of the examined object are suppressed.

According to a further preferred embodiment, a time series of MR images is reconstructed from the acquired MR signals. In a 4D dynamic radial scan, the proposed soft-gating method of the present invention can be applied at each time frame (i.e., each dynamic scan). The advantages are not only that an improved image quality compared to e.g. navigator gating techniques is provided within the same scan time, but the method of the invention also leaves the equidistant dynamic scan time unaffected by e.g. the breathing pattern of the patient under examination. Furthermore, the 4D aspect may include deriving the time information using a contour sharing principle, 3D high pass filter (KWIC) with a weighting principle. The histogram and weighting may generally vary according to dynamics. The weighting may be constant in the central plateau, which is effective for 4D scanning to reduce dynamic flicker effects.

In order to optimize the k-space distribution of the acquired MR signals, the angular ordering of the radial or helical k-space profiles may be selected according to a golden angle approach. In the golden angle approach, as described above, the rotation angle of the k-space profile is 111.25 ° per increment from acquisition to acquisition, which corresponds to 180 ° multiplied by the golden ratio. Thus, a subsequently sampled radial or spiral k-space profile always adds supplementary information while filling the largest gap within the previously sampled profile set. Thus, the contours of any one set of subsequent acquisitions substantially uniformly cover k-space.

The distribution of rotation angles may also be adapted to the Anisotropic field of view (see Wu et al, "Anisotropic field-of-view support for gold angle radial imaging", Magn Reson Med.,76, 229-. Other methods for optimizing the sampling order, such as CENTRA ordering (see WO2016202707a1) or rotating star stacking (see Zhou et al, "Golden-ratio rotated stack-of-stands acquisition for improved volumetric MRI", magn.

The presently described inventive method can be performed by means of an MR device comprising: at least one main magnet coil for generating a uniform, steady magnetic field B in an examination volume, a plurality of gradient coils, at least one body RF coil, a control unit and a reconstruction unit₀A plurality of gradient coils for generating switched magnetic field gradients in different spatial directions within the examination volume, at least one body 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 a temporal succession of RF pulses and switched magnetic field gradients, and a reconstruction unit for reconstructing an MR image from the received MR signals. The method of the invention can be implemented by a corresponding programming of the reconstruction unit and/or the control unit of the MR device.

The method of the invention can advantageously be performed on most MR apparatuses currently in clinical use. For this purpose, it is merely necessary to make the MR device execute the above-described method steps of the invention by means of a computer program controlling the MR device. The computer program may be present on a data carrier or may be present in a data network for downloading to be installed in the control unit of the MR device.

Drawings

The accompanying drawings disclose preferred embodiments of the present invention. It is to be understood, however, that the drawings are designed solely for the purposes of illustration and not as a definition of the limits of the invention. In the drawings:

FIG. 1 shows a block diagram of an MR apparatus for carrying out the method of the invention;

FIG. 2 shows a schematic diagram of k-space schematically illustrating an embodiment of a k-space sampling scheme of the present invention;

FIG. 3 illustrates the determination of weighting factors applied in the soft motion gating scheme of the present invention;

fig. 4 shows two MR images, fig. 4a shows an MR image conventionally acquired by a 3D radial scan, and fig. 4b shows the same MR image acquired and reconstructed according to the invention.

Detailed Description

Referring to fig. 1, an MR device 1 is shown in block diagram. The device comprises superconducting or resistive main magnet coils 2 such that a substantially uniform, temporally constant main magnetic field B is created along a z-axis through an examination volume₀. The apparatus further comprises a set of (first, second and, where applicable, third) shim coils 2 ', wherein the current flowing through the individual shim coils of the set 2' is controllable so as to bring B within the examination volume₀The deviation is minimized.

The 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 magnetic resonance, saturate spins, etc., to perform MR imaging.

More specifically, the gradient amplifiers 3 apply current pulses or waveforms to selected ones of the whole-body gradient coils 4, 5 and 6 along the x-axis, y-axis and z-axis of the examination volume. A digital RF frequency transmitter 7 transmits RF pulses or pulse packets via a transmit/receive switch 8 to a body RF coil 9, thereby transmitting RF pulses into the examination volume. A typical MR imaging sequence comprises a packet of short duration RF pulse segments which, together with any applied magnetic field gradients, enable selected manipulation of the nuclear magnetic resonance signals. The RF pulses are used to saturate, excite resonance, invert magnetization, refocus resonance, or manipulate resonance and select the portion of the body 10 that is positioned in the examination volume. The MR signals are also picked up by the body RF coil 9.

For generating MR images of a limited region of the body 10 or for scan acceleration by means of parallel imaging, a set of local array RF coils 11, 12, 13 is placed adjacent to the region selected for imaging. The array coils 11, 12, 13 can be used for receiving MR signals caused by body coil RF transmissions.

The resulting MR signals are picked up by the body RF coil 9 and/or the array RF coils 11, 12, 13 and demodulated by a receiver 14 which preferably comprises a preamplifier (not shown). The receiver 14 is connected to the RF coils 9, 11, 12, and 13 via the transmission/reception switch 8.

The host computer 15 controls the shim 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 fast field echo (TFE) or fast spin echo (TSE) for 3D radial or helical imaging. For the selected sequence, the receiver 14 receives a single or multiple MR signal profiles in rapid succession after each RF excitation pulse. The 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 dedicated to acquiring raw image data.

Finally, the digital raw image data is reconstructed by a reconstruction processor 17 into an image representation, and the reconstruction processor 17 applies a fourier transform or other suitable reconstruction algorithm, e.g., SENSE. The MR images represent a three-dimensional volume. The images are then stored in an image memory where the images can be accessed to convert slices, projections or other portions of the image representation into a suitable format for visualization, for example, via a video monitor 18, the video monitor 18 providing a human-readable display of the resulting MR images.

The host computer 15 is programmed to execute the method of the present invention described above and below.

As shown in fig. 2a, MR signals are acquired according to a 3D star-stack acquisition scheme (e.g., using a conventional TFE imaging sequence). One or more MR signals are acquired after each of a plurality of spatially non-selective or slab-selective RF excitations, wherein each MR signal represents a k-space profile. The MR signals are acquired as radial k-space profiles from a plurality (in the embodiment of fig. 2a, five) of parallel slices 21, 22, 23, 24, 25. The slice is along the slice direction k_zAre arranged at different locations. At k_zPerforming Cartesian phase encoding in a direction while following a surrounding center (k) within each individual slice_x＝k_y0) rotating radial "spokes" acquire MR signals. This results in a cylindrical k-space coverage consisting of stacked disks. And for the angle sequencing of the spokes, adopting a golden angle scheme. The angle of the spokes Δ Φ is 111.25 ° each time. The radial density of the k-space profile of each slice, i.e. the number of spokes acquired, varies depending on the slice position, wherein the radial density is higher at more central k-space positions and lower at more peripheral k-space positions. This is achieved in the embodiment of fig. 2a as follows: in a first step, a plurality of spokes are acquired from only the central k-space slice. In the next step, the same number of spokes is acquired from the central three slices, and in the third step, the same number of spokes is again acquired from all five slices. The continuous acquisition of the phase encoding step along the slice direction is performed before sampling the k-space profile at different gold angular positions, which is crucial to ensure high data consistency and general motion robustness. In this way, the k-space center (k)_zNear 0) is more densely sampled than the k-space periphery. The radial density of the k-space contour (spokes) is varied in such a way that the Nyquist criterion for a given FOV is around the center of k-space (k)_x＝k_y＝k_z0) ellipsoid 26 (see fig. 2b)Is satisfied. The present invention minimizes the total scan time of k-space samples that meet the nyquist criterion in the central portion 26 of k-space. According to the invention, oversampling is provided in the image energy dominated central portion 26 of k-space. The radial sampling density gradually decreases from the central slices 22, 23, 24 to a lower radial sampling density in the peripheral slices 21, 25. Outside the ellipsoid 26, i.e. in the peripheral part 27 of k-space, the radial k-space density may even be below the nyquist threshold without significantly affecting the image quality. The star-stacked acquisition scheme may be implemented in practice as a 3D central radial star stack or a 3D elliptical variable density radial star stack. Due to the higher radial sampling density around the center of k-space, streak artifacts have been reduced by the described k-space sampling scheme within the minimum scan time.

According to the invention, the movement of the body 10 under examination is detected, for example, by means of a conventional breathing belt. During the acquisition of the MR signals, a displacement of the examined anatomy is determined in such a way that each of the acquired MR signals can be assigned to a motion (respiration) state. Each of the motion states is defined as one of a plurality of consecutive ranges corresponding to a displacement caused by respiratory motion.

On this basis, a 3D soft-gating method is implemented by weighting the MR signals in the oversampled central portion 26 of k-space, wherein a stronger weighting is applied to MR signals acquired in more frequent motion states and a weaker weighting is applied to MR signals acquired in less frequent motion states. The MR signals acquired in case the patient presents the position it most frequently assumes during breathing are given a stronger weight, whereas the MR signals acquired in the less frequently presented positions are suppressed in the reconstructed MR image.

The frequency of occurrence of each motion state as a basis for MR signal weighting is derived from a histogram as illustrated in the lower graph of fig. 3. The histogram is established during or after MR signal acquisition. Which reflects the number of MR signals acquired per motion state. In this schematic diagram, the frequency F is depicted as a function of the detected shift d assigned to the respective motion state. The weighting factor W shown in the upper diagram of fig. 3 is derived from the histogram, wherein the user specified gating percentage is taken into account. The gating percentage defines the proportion of the MR signal that is suppressed by weighting as a global parameter (where image noise and artifact levels are balanced) that can be tuned by the user as desired. When determining the weighting factors, the conformity with the nyquist criterion should be considered for the central part 27 of k-space in order to avoid aliasing artifacts. As can be seen in the upper graph of fig. 3, the largest weight (arrow 30) is applied to the most frequently occurring shift d. The weight W decreases towards less frequently occurring shifts (arrow 31). A minimum weighting is applied to the MR signal due to the rare outliers of shift d (arrow 32). The weighting factor W varies smoothly according to the detected shift d. In the depicted embodiment, the weighting factor is a linear function of the shift d, where different slopes are assigned to different ranges of the shift d. Any other shape of the curve w (d) is of course possible.

The result of this soft-gating approach is that the MR images reconstructed from the 3D radial or helical acquisition have significantly reduced levels of artifacts in the presence of motion of the patient's body 10. This can be seen in fig. 4, which fig. 4 shows a slice MR image acquired from the thorax region using a 3D radial acquisition method. The MR image shown in fig. 4a has been conventionally acquired and reconstructed, whereas the MR image of fig. 4b has been reconstructed using the soft-gating method of the present invention. The MR image of fig. 4a shows significant motion artifacts (white arrows indicate streak artifacts). These artifacts are not present in the MR image of fig. 4 b.