Improved layer multiplexing process

Abstract

An inventive method for separating measurement data of an object under investigation, which are recorded simultaneously as collapsed from a plurality of layers of the object under investigation using an in-plane acceleration technique with an undersampled scanning pattern of k-space corresponding to the acceleration technique, into single-layer measurement data comprises the steps: - Acquiring reference measurement data in at least two segments such that a complete set of reference measurement data according to Nyquist is obtained, and that a sampling pattern used when acquiring the reference measurement data within a segment corresponds to an acceleration factor that also corresponds to the sampling pattern of the collapsed measurement data, - Recording the collapsed measurement data, - Generating separation data based on reference measurement data recorded in at least one of the at least two segments, - Separating the collapsed measurement data into single-layer measurement data using the generated separation data.

Description

[0001] The invention relates to an improved layer multiplexing method, in particular to an improved layer multiplexing method using an in-plane acceleration technique.

[0002] Magnetic resonance imaging (MR) is a well-known technique for generating images of the interior of an object under examination. Simply put, the object under examination is positioned in a magnetic resonance scanner in a comparatively strong, static, homogeneous basic magnetic field, also known as a B0 field, with field strengths ranging from 0.2 Tesla to 7 Tesla and more, so that its nuclear spins are oriented along the basic magnetic field. To trigger nuclear magnetic resonances, which can be measured as signals, high-frequency excitation pulses (RF pulses) are radiated into the object under examination. The triggered nuclear magnetic resonances are measured as so-called k-space data, and based on these, MR images are reconstructed or spectroscopy data is acquired. To spatially encode the measured data, rapidly switched magnetic gradient fields, or gradients for short, are superimposed on the basic magnetic field.A scheme used, which describes a temporal sequence of RF pulses to be delivered and gradients to be switched, is referred to as a pulse sequence (scheme), or simply a sequence. The recorded measurement data are digitized and stored as complex numerical values ​​in a k-space matrix. A corresponding MR image can be reconstructed from the k-space matrix, for example, using a multidimensional Fourier transform.

[0003] In recent years, the acceleration of MR measurements through targeted undersampling of k-space, for example, through so-called parallel acquisition techniques, has become established. These parallel acquisition techniques (ppa techniques), which can be used to shorten the acquisition times required to acquire the desired data by incomplete sampling (i.e., undersampling) of (2D) k-space according to Nyquist, include GRAPPA (“GeneRalized Autocalibrating Partially Parallel Acquisition”) and SENSE (“SENSitivity Encoding”). With parallel acquisition techniques, the measurement points in k-space not measured during subsampling are generally evenly distributed across the k-space to be measured according to Nyquist, so that, for example, every second k-space row is measured in each k-space plane (kx,ky) assigned to a slice. Such PPA techniques are therefore also referred to as “in-plane” acceleration techniques.

[0004] In parallel acquisition techniques, the "missing" k-space data is reconstructed using coil sensitivity data. This coil sensitivity data of the receive coils used to acquire the measurement data is determined from reference measurement data that completely samples at least one region of the k-space to be measured, usually the central region, according to the Nyquist condition. The coil sensitivity data determined from the reference measurement data for the reconstruction of missing k-space data is also referred to as the "kernel," or in the case of GRAPPA, the "GRAPPA kernel."

[0005] Any measurement data from which image data is not (or not only) reconstructed, but which is used as a reference to determine further information for a technique or algorithm used, is referred to herein as reference measurement data.

[0006] One of the fastest known MR imaging techniques is echo-planar imaging (EPI), in which an oscillating, i.e., bipolar readout gradient is applied after an RF excitation pulse. With each change in the polarization direction of the gradient, this refocuses the transverse magnetization as far as the T2* decay allows, thus generating a gradient echo. In other words, by switching the bipolar readout gradient after an RF excitation pulse within the free induction decay after excitation (FID), or if an additional RF refocusing pulse is applied after the RF excitation pulse, an echo train of rising and falling gradient echoes with alternating signs is generated within the spin echo thus generated. EPI pulse sequences can be used as so-called "single-shot" methods, in which all measurement data is used to generate an image of a subvolume, e.g.a slice of the object under investigation can be recorded after only one RF excitation pulse.

[0007] In addition to reducing acquisition time, parallel acquisition techniques offer further advantages depending on the MR imaging technique used. For example, in EPI acquisition techniques, the length of the readout train can be reduced, particularly in the context of functional MR acquisitions (fMRI, or functional magnetic resonance imaging), for determining diffusion properties (DWI, or diffusion-weighted imaging), or for perfusion measurements. This allows the echo time of the acquired echo signals to be reduced, resulting in a higher signal-to-noise ratio (SNR). Furthermore, a higher bandwidth can be achieved along the phase-encoding direction, thus reducing distortion. Furthermore, shortening the readout train length in this way can reduce the influence of T2* relaxation on the acquired signals, resulting in sharper images.

[0008] In clinical routine, subsampling factors in the range of 2-4 are typically used - in-plane subsampling is often combined with simultaneous multi-slice (SMS) acquisition.

[0009] In general, such methods, in which images of different slices are acquired simultaneously, can be characterized by the fact that, at least during part of the measurement, the transverse magnetization of at least two slices is deliberately used simultaneously for the imaging process ("multi-slice imaging," "slice multiplexing"). In contrast, in established "multi-slice imaging," the signal from at least two slices is recorded alternately, i.e., completely independently of each other, with a correspondingly longer measurement time.

[0010] Well-known methods of this kind, also known as simultaneous multi-slice (SMS) methods, include Hadamard coding, methods with simultaneous echo refocusing, methods with broadband data acquisition, and methods that use parallel imaging in the slice direction. The latter methods include, for example, the CAIPIRINHA technique, as described by Breuer et al. in "Controlled Aliasing in Parallel Imaging Results in Higher Acceleration (CAIPIRINHA) for Multi-Slice Imaging," Magnetic Resonance in Medicine 53, 2005, pp. 684-691, and the blipped CAIPIRINHA technique, as described by Setsompop et al. in “Blipped-Controlled Aliasing in Parallel Imaging for Simultaneous Multislice Echo Planar Imaging With Reduced g-Factor Penalty”, Magnetic Resonance in Medicine 67, 2012, pp. 1210-1224.

[0011] In particular, the latter layer multiplexing techniques use a so-called multi-band RF pulse to simultaneously excite or otherwise manipulate two or more slices, e.g., to refocus or saturate them. Such a multi-band RF pulse is, for example, a multiplex of individual RF pulses that would be used to manipulate the individual slices to be manipulated simultaneously. Multiplexing produces, for example, a baseband-modulated multi-band RF pulse by summing the pulse shapes of the individual RF pulses. The spatial coding of the recorded signals is essentially achieved by a common two-way gradient switching (two-dimensional gradient coding).

[0012] The resulting signals from all excited layers are recorded in a collapsed data set using several receiving antennas and must then be separated according to the individual layers.

[0013] The algorithms used to separate the recorded signals and assign them to the individual layers rely on the availability of additional information from which the spatial sensitivity of the receiving antennas (sensitivity data) is determined, which is usually derived from reference measurement data. Such sensitivity data is used – in addition to the spatial coding provided by the gradient fields – as additional spatial coding, which allows the desired signal assignment to be achieved by solving systems of equations, using iterative methods, or using trained networks.

[0014] With slice multiplexing methods, parallel acquisition techniques can be used to separate the measurement data acquired simultaneously and thus in a collapsed manner for different slices. Reference measurement data must be acquired for all affected slices. This is usually done as part of an additional reference measurement, which measures reference measurement data individually for each desired slice.

[0015] In order to separate the resulting signals from the different slices, a different phase is applied to each individual RF pulse prior to multiplexing. This can be done, for example, by adding a phase that is linear (e.g., with the k-space coordinate in the phase encoding direction (k y)). This allows each slice to be given a different phase gradient, shifting the slices relative to each other in the image space. This shift is controlled by the so-called field of view (FOV) shift factor. How an optimal FOV shift factor can be determined is described, for example, in DE102016218955.

[0016] In the CAIPIRINHA methods described in the aforementioned articles by Breuer et al. and Setsompop et al., additional phase shifts are imposed between the simultaneously excited slices by switching additional gradient blips or by additionally modulating the phases of the RF pulses of the multi-band RF pulses. These phase shifts generate interslice FoV shifts in the image space. These additional shifts improve the quality of the separation of the slice signals, e.g., using the slice-GRAPPA method mentioned in the article by Breuer et al., particularly when the coil sensitivities exhibit such small differences in the sensitivity profiles of the individual coils used that they would otherwise be insufficient for reliable slice separation. This reduces artifacts in the image data ultimately reconstructed from the measured data.The sensitivity data determined from the reference measurement data for separating collapsed measurement data from multiple layers into measurement data from individual layers is also referred to as “kernel”, in the case of slice-GRAPPA also as “slice-GRAPPA kernel”.

[0017] The effect of additional phase shifts on the sampling pattern of a two-dimensional (2D) slice-multiplexed measurement can be described as follows: Due to the additional phases imposed in the slice-multiplexed CAIPIRNHA method, the measurement points imparted with the additional phase are shifted in k-space in the kz direction. The magnitude of this shift in the kz direction depends on the imposed phase. This is also described, for example, in the article by Zahneisen et al.: "Three-Dimensional Fourier Encoding of Simultaneously Excited Slices: Generalized Acquisition and Reconstruction Framework," Magn. Reson. Med. 71, pp. 2071-2081 (2014).

[0018] The reference measurement data from which sensitivity data for separating the simultaneously acquired slices are obtained (slice separation reference measurement data) are usually measured additionally for each slice multiplexing measurement.

[0019] With EPI acquisition techniques, it should also be noted that, due to the changing polarity of the readout gradient used to acquire the echo signals as measurement data, even small asymmetries in the acquired measurement data can lead to inconsistencies in the k-space data and, consequently, to unwanted replicas (also called "(N / 2) ghosts") in the image data. Such asymmetries can, for example, be the result of dynamic interference fields (e.g., induced eddy currents when switching gradient fields).

[0020] This occurs because the measurement data obtained from the echo signals recorded using EPI are sorted into a raw data k-space matrix in such a way that the sorting direction changes from row to row of the raw data k-space matrix. If even slight deviations occur from row to row, e.g., due to delays in the gradient circuit or eddy currents, this leads to so-called (N / 2) ghosts. This means that with an image matrix of NxN points, the actual image is reimaged shifted by N / 2 in the positive and negative directions relative to the center of the image matrix, generally with different intensities.To correct such (N / 2) ghosts, it is known, for example, from US6043651 to record three navigator signals while switching a bipolar readout gradient. This can be used to correct zero- and first-order phase shifts between gradient echoes recorded with different polarity in the readout direction. This can correct such shifts. For this purpose, a correlation of the recorded navigator signals in image space is used to determine correction factors. These correction factors are used in a reconstruction of image data from the gradient echoes recorded as measurement data in a raw data k-space matrix to correct the aforementioned shifts in the raw data k-space matrix.

[0021] Another phase correction method, called DORK, for correcting shifts caused by temporal variations in a basic magnetic field applied during an EPI measurement, e.g., drift, is known, for example, from US9329254B2. In this method, an evolution of the gradient echoes acquired with one polarity is compared with an evolution of the gradient echoes acquired with the other polarity over consecutive acquisitions of raw data k-space matrices. Typically, such a DORK correction is averaged over an entire image volume.

[0022] As a rule, such known correction methods succeed in reducing asymmetries in the measurement data measured for imaging and, if these are also recorded using EPI, also in reference measurement data, to such an extent that the occurrence of residual ghosts in their image data (image ghosts) no longer plays a real role for clinical diagnostics.

[0023] However, such correction methods based on one-dimensional navigator signals can reach their limitations, especially when the interference fields causing inconsistencies exhibit spatial variations not only along the readout direction. While the intensity of image ghosts is still reduced, image artifacts may still remain, which can lead to limitations in clinical usability.

[0024] Using a "Dual Polarity GRAPPA" (DPG) technique, as described in the article by Hoge et al., "Dual-Polarity GRAPPA for Simultaneous Reconstruction and Ghost Correction of Echo Planar Imaging Data," Magn. Reason. Med. 76: pp. 32-44, 2016, it is possible to reduce the intensity of image ghosts to such an extent, even in complex interference field geometries, that the images are fully suitable for clinical diagnosis.

[0025] Such a DPG approach requires additional reference measurement data. Specifically, sets of fully Nyquist-acquired reference measurement data are used for each desired slice and both polarization directions. From these sets, complete reference measurement data sets are generated for each polarization direction. These sets are used to generate coil sensitivity data for both polarization directions. These data are used in the reconstruction of image data from measurement data, usually subsampled using an EPI acquisition technique, within the DPG method to largely avoid ghost artifacts, especially even under conditions with nonlinear phase variations.For example, before performing the actual DPG algorithm for reference measurement data, a fully sampled set of sorted reference measurement data with only one readout direction, i.e. only with positive polarity of the readout gradient or only with negative polarity of the readout gradient, can be combined for a DPG algorithm from two sets of reference measurement data recorded with inverted polarities of the readout gradients, e.g. by correspondingly assigning recorded k-space lines.

[0026] In its original form, an algorithm used at DPG simultaneously reduces image ghosts and supplements k-space data missing due to the GRAPPA technique. Modified variants can also reduce only the asymmetries of the k-space data acquired with positive or negative readout gradients, thus enabling image ghost reduction for fully sampled data for imaging, e.g., imaging using SMS technology, and / or for alternative reconstruction methods (e.g., DL-based supplementation of missing k-space data).

[0027] Several methods for acquiring reference measurement data for applying a DPG algorithm are known in the state of the art. What all methods have in common is that two sets of reference measurement data are required, each with inverted polarity of the readout gradients.

[0028] For example, it is known to acquire reference measurement data for a DPG algorithm using a single-shot EPI acquisition technique. In this technique, reference measurement data are acquired using an EPI acquisition technique after each excitation for each slice in which measurement data for imaging is acquired. The contrast of the reference measurement data may differ from the respective contrast of the measurement data acquired for imaging, e.g., because a shorter repetition time TR was selected for acquiring the reference measurement data than for acquiring the measurement data for imaging in order to reduce the measurement time required to acquire the reference measurement data. In a second measurement, reference measurement data are acquired with inverted readout gradients.

[0029] Furthermore, it is known to acquire reference measurement data for DPG algorithms using a multi-shot EPI acquisition technique, in which the reference measurement data for each slice for which measurement data for imaging is to be acquired are acquired using multiple excitations of a segmented EPI acquisition technique. Here, too, the contrast of the reference measurement data can differ from the respective contrast of the measurement data acquired for imaging, e.g., if a shorter repetition time TR is selected for the acquisition of the reference measurement data in order to reduce the overall measurement time. Again, in second measurements, reference measurement data are acquired with inverted readout gradients.

[0030] To acquire reference measurement data for a DPG algorithm using a multi-shot EPI acquisition technique, conventional segmentation can be used. In this method, the k-space to be acquired is divided into segments, from which reference measurement data are acquired after each excitation. The reference measurement data for each segment are acquired after a longer relaxation time, resulting in a relatively large time interval between the individual excitations, but with an identical flip angle. This means that the reference measurement data for all segments implicitly exhibit comparable contrast.

[0031] To acquire reference measurement data for a DPG algorithm using a multi-shot EPI acquisition technique, a so-called FLEET technique (FLEET: "fast low-angle excitation echo-planar technique") can also be used, as described, for example, in the article by Polimeni et al. "Reducing Sensitivity Losses Due to Respiration and Motion in Accelerated Echo Planar Imaging by Reordering the Autocalibration Data Acquisition," Magn. Reson. Med. 75, pp. 665-679, 2016. In this technique, the reference measurement data of the individual segments are acquired in short temporal sequence. The flip angle of the respective excitation is varied between the different segments to ensure a similar contrast.

[0032] Acquiring reference measurement data for a DPG algorithm already takes a considerable amount of time. If additional reference measurement data is required, e.g., for SMS technology, especially for separating measurement data recorded in a collapsed manner for multiple layers into measurement data for individual layers, the total time required to acquire all reference measurement data increases further. Furthermore, it is often necessary to use additional time-consuming, so-called dummy recordings, during which no data is recorded, between recordings of different reference measurement data, for example, to compensate for steady-state effects.

[0033] Fig. Figure 1 schematically shows an example of an EPI-SMS measurement sequence with prior acquisition of all required reference measurement data. For example, this could be a measurement with an SMS factor of 4 (i.e., four slices are acquired simultaneously) and a GRAPPA acceleration factor of 2 (i.e., every second k-space line is acquired), with a total of 64 slices to be acquired.

[0034] For a DPG algorithm, recording reference measurement data Rdpg, e.g., using a FLEET technique, already takes a time Tdpg that can be approximately 25 seconds. For each slice to be recorded, the reference measurement data Rdpg can be recorded, for example, in two segments each, in both polarization directions of the readout gradients.

[0035] In the example shown, the acquisition of the reference measurement data Rdpg is followed by a dummy scan Dsms, e.g., to prepare a subsequently desired steady state, which in turn is followed by the acquisition of reference measurement data Rsms to determine sensitivity data for separating measurement data acquired in a collapsed manner using an SMS technique from multiple slices into measurement data from the individual slices. Such a dummy scan Dsms prior to the acquisition of the reference measurement data Rsms is necessary, for example, if a repetition time TR(Rdpg) for the acquisition of the reference measurement data Rdpg for the DPG algorithm differs from a repetition time TR(Rsms) for the acquisition of the reference measurement data for separating the collapsed measurement data, which would lead to undesirable steady-state effects without the dummy scan Dsms, and is usually the case when using FLEET when acquiring the reference measurement data Rpdg.In the example shown, a total time Tsms is required to record the reference measurement data Rsms for the separation of collapsed measurement data, which can be, for example, about 11 seconds.

[0036] Before acquiring measurement data B1 for a first tuple of N (e.g., N=4) slices using the SMS technique used, further, e.g., two, dummy scans D1, D2 can be applied to prepare a steady state desired for acquiring measurement data B1. These dummy scans comprise RF pulses corresponding to those used during the subsequent acquisition of measurement data. Acquiring measurement data B1 for the first tuple is followed by all further, e.g., 15, acquisitions of measurement data from the remaining tuples of the total, e.g., 64, slices to be acquired. Acquiring measurement data takes approximately 3 seconds, for example, so that in the example given, the acquisition of measurement data would take 16*3 seconds = 48 seconds.The time required to record the reference measurement data Rdpg and Rsms reduces the actual reduction in the total measurement time achieved by the acceleration techniques used to record the measurement data (SMS and GRAPPA in the example). Even the time Tsms assumed in the example for (preparing and) recording the reference measurement data Rsms is significant.

[0037] The invention is based on the object of enabling a faster and / or better, e.g. in relation to a signal-to-noise ratio (SNR), recording of all reference measurement data required for an SMS-DP measurement.

[0038] The object is achieved by a method for separating measurement data of an examination object, which are recorded as collapsed from a plurality of slices of the examination object simultaneously and using an in-plane acceleration technique with a subsampled scanning pattern of k-space corresponding to the acceleration technique by means of magnetic resonance technology, into single-slice measurement data according to claim 1, a magnetic resonance system according to claim 12, a computer program according to claim 13, and an electronically readable data carrier according to claim 14.

[0039] A method according to the invention for separating measurement data of an examination object, which are recorded as collapsed from a plurality of slices of the examination object simultaneously and using an in-plane acceleration technique with a subsampled scanning pattern of the k-space corresponding to the acceleration technique by means of magnetic resonance technology, into single-slice measurement data comprises the steps: - Recording reference measurement data in at least two segments in such a way that a complete set of reference measurement data is available according to Nyquist, and that a sampling pattern used when recording the reference measurement data within a segment corresponds to an acceleration factor to which the sampling pattern of the collapsed recorded measurement data also corresponds, - Recording the collapsed measurement data, - generating separation data based on reference measurement data recorded in at least one of the at least two segments, - Separating the collapsed acquired measurement data into single-layer measurement data using the generated separation data.

[0040] Although reference measurement data for a reconstruction of missing k-space data in measurement data acquired by means of an in-plane technique completely sample the k-space at least in a central area according to Nyquist, and, in contrast, reference measurement data for the later separation of collapsed measurement data are acquired with a sampling pattern corresponding to the subsequent acquisition of measurement data by means of the SMS technique used, in particular with an in-plane acceleration corresponding to the subsequent acquisition of measurement data by means of an SMS technique used, an acquisition of reference measurement data for a reconstruction of missing k-space data in measurement data acquired by means of an in-plane technique can be carried out in segments as proposed according to the invention.

[0041] By recording reference measurement data in segments according to the invention with a scanning pattern that corresponds to an (in-plane) acceleration factor that also corresponds to a scanning pattern used when recording collapsed measurement data from multiple slices, further recording of reference measurement data for separating the collapsed measurement data can be eliminated, thereby reducing the overall time otherwise required for recording the required reference measurement data. Furthermore, with reference measurement data recorded according to the invention for reconstructing missing k-space data in measurement data recorded using an in-plane technique, the quality of separation of the collapsed measurement data recorded for a plurality of slices can be improved.

[0042] A magnetic resonance system according to the invention comprises a magnet unit, a gradient unit, a radio frequency unit and a control device designed to carry out a method according to the invention with a reference measurement data unit.

[0043] A computer program according to the invention implements a method according to the invention on a control device when it is executed on the control device. For example, the computer program comprises instructions that, when the program is executed by a control device, e.g., a control device of a magnetic resonance imaging system, cause this control device to execute a method according to the invention. The control device can be configured in the form of a computer.

[0044] The computer program can also be in the form of a computer program product which can be loaded directly into a memory of a control device, with program code means for carrying out a method according to the invention when the computer program product is executed in a computing unit of the computing system.

[0045] A computer-readable storage medium according to the invention comprises instructions which, when executed by a control device, e.g. a control device of a magnetic resonance system, cause the control device to carry out a method according to the invention.

[0046] The computer-readable storage medium can be designed as an electronically readable data carrier which comprises electronically readable control information stored thereon, which control information comprises at least one computer program according to the invention and is designed such that it carries out a method according to the invention when the data carrier is used in a control device of a magnetic resonance system.

[0047] The advantages and details given with regard to the method also apply analogously to the magnetic resonance system, the computer program product and the electronically readable data carrier.

[0048] Further advantages and details of the present invention will become apparent from the exemplary embodiments described below and from the drawings. The examples listed do not represent a limitation of the invention. They show: Fig. 1 schematically shows an example of a sequence of an EPI-SMS measurement with prior recording of all required reference measurement data, Fig. 2 a schematic flow diagram of a method according to the invention, Fig. 3-4 Examples of segments with which reference measurement data can be recorded according to the invention, based on their scanning patterns in k-space, Fig. 5-6 Examples in k-space for possible combination data sets, as they can be combined from segments according to the invention for the acquisition of reference measurement data, Fig. 7 a schematically illustrated magnetic resonance system according to the invention.

[0049] Fig. 2 is a schematic flow diagram of a method according to the invention for separating measurement data MD of an examination object, which are acquired as collapsed from a plurality of slices of the examination object simultaneously and using an in-plane acceleration technique with a subsampled scanning pattern of the k-space corresponding to the acceleration technique by means of magnetic resonance technology, into single-slice measurement data.

[0050] In this case, reference measurement data RMD are recorded in at least two segments S1...Sn in such a way that a total set of reference measurement data is available that is complete according to Nyquist, and that a sampling pattern used when recording the reference measurement data RMD within a segment S1...Sn corresponds to an acceleration factor that also corresponds to the sampling pattern of the collapsed recorded measurement data (block 201).

[0051] The reference measurement data RDM can, for example, be reference measurement data for a parallel acceleration technique, e.g., in-plane GRAPPA. As described above, such reference measurement data are acquired for each desired slice and are complete for each slice at least in a central region of k-space. This means that the reference measurement data RMD acquired in all used segments together completely fill the corresponding k-space and are suitable for determining coil sensitivity data. Thus, based on such reference measurement data RMD, measurement data not acquired within the framework of a parallel acceleration technique can be reconstructed.

[0052] In order to divide a recording of such a complete set of reference measurement data into segments such that a sampling pattern used when recording the reference measurement data RMD within a segment S1...Sn corresponds to an acceleration factor that also corresponds to the sampling pattern of the collapsed recorded measurement data, the recording of the reference measurement data can, for example, simply be divided into a number of segments that corresponds to the acceleration factor of the recording of the collapsed recorded measurement data, and a sampling pattern can be used in each segment that also corresponds to the acceleration factor of the sampling pattern of the recording of the collapsed recorded measurement data, whereby, however, the sampling pattern used samples different k-space data in each segment, so that by combining the k-space data sampled in the respective segments, the k-space as a whole is completely sampled according to Nyquist.

[0053] Fig. Figure 3 shows an example of sampling patterns in k-space for segments S1', S2', S3', along which reference measurement data can be acquired for a separation of measurement data acquired collapsed with an in-plane acceleration technique, with a sampling pattern corresponding to an acceleration factor of three.

[0054] In the example shown, k-space lines (running in the readout direction) along which reference measurement data RMD are recorded in the respective segment S1', S2', S3' are shown as solid lines, and k-space lines along which no reference measurement data RMD are recorded in the respective segment S1', S2', S3' are shown as dotted lines. As can be seen, according to the sampling pattern in the example shown, only every third k-space line is sampled in each segment S1', S2', S3', i.e., reference measurement data is only recorded along every third k-space line. The sampled k-space lines in the segments S1', S2', S3' are permuted such that each k-space line of the desired k-space is sampled once, resulting in a complete set of reference measurement data. The sampling patterns here thus correspond to an acceleration factor of three (PAT=3).The recording of the reference measurement data RMD in the segments S1', S2', S3' is carried out for each layer recorded in the measurement data MD to be separated and can be carried out in a manner known per se using a selected recording technique.

[0055] For other acceleration factors when recording the collapsed measurement data to be separated, an analogous procedure can be used.

[0056] The collapsed measurement data can be acquired using an EPI acquisition technique. EPI acquisition techniques offer the advantages mentioned above and are often used, for example, for diffusion imaging. Especially for MD measurement data acquired using an EPI acquisition technique that is to be separated, the reference measurement data can be used for a dual-polarity GRAPPA (DPG) technique, since this allows ghost artifacts to be eliminated or at least mitigated, as described above.

[0057] As already described above, such reference measurement data for DPG techniques must be recorded along the k-space lines of the desired k-space, as in parallel acceleration techniques, completely according to Nyquist, but here additionally also for both readout directions.

[0058] Fig. Figure 4 shows an example of sampling patterns in k-space for segments S1, S2, S3, S4, along which reference measurement data can be acquired for a separation of measurement data acquired collapsed with an EPI acquisition technique and with an in-plane acceleration technique, with a sampling pattern corresponding to an acceleration factor of two.

[0059] In the example shown, k-space lines along which reference measurement data RMD are recorded in the respective segment S1, S2, S3, S4 are shown as solid arrows (in the respective readout direction) and k-space lines along which no reference measurement data RMD are recorded in the respective segment S1, S2, S3, S4 are shown as dotted lines. As can be seen, according to the sampling pattern in the example shown, only every second k-space line is sampled in each segment S1, S2, S3, S4, i.e. reference measurement data is only recorded along every second k-space line. In the segments S1 and S2 as well as in segments S3 and S4, the same k-space data is sampled, but in different readout directions.The sampled k-space lines are permuted in segments S1 and S3, as well as in segments S2 and S4, such that each k-space line of the desired k-space is sampled once in each readout direction, resulting in a complete set of reference measurement data for each readout direction. The sampling patterns thus correspond to an acceleration factor of two (PAT=2). The acquisition of the reference measurement data RMD in segments S1, S2, S3, and S4 is performed for each slice acquired in the measurement data MD to be separated and can be performed in a conventional manner using a selected acquisition technique for DPG techniques.

[0060] For other acceleration factors when recording the collapsed measurement data to be separated, the same procedure can be followed.

[0061] The collapsed measurement data MD to be separated is acquired (block 203). The acquisition of the collapsed measurement data MD occurs, for example, directly after the acquisition of the reference measurement data RMD and can be performed using a known layer multiplexing method using an in-plane acceleration technique, such as GRAPPA, with a known acquisition technique, e.g., an SMS-EPI acquisition technique.

[0062] Separation data TD are generated on the basis of reference measurement data recorded in at least one of the at least two segments (block 205).

[0063] In a simple case, the separation data TD is based on the reference measurement data RMD of one of at least two segments in which the reference measurement data was recorded. By selecting the sampling patterns in the segments of the reference measurement data recording such that they correspond to an acceleration factor that also corresponds to a sampling pattern used when recording the measurement data MD to be separated, the effective echo spacing when recording a segment of reference measurement data is equal to the effective echo spacing of the collapsed measurement data MD. Thus, each segment already corresponds to a conventional, previously separately recorded set of reference measurement data Rsms (cf. Fig. 1), and is therefore suitable for generating separation data TD in a known manner. For example, the separation data TD can be based on the reference measurement data of one of the segments S1', S2', S3' of the Fig. 3 for recording the collapsed measurement data MD with an acceleration factor (PAT=3) or, for recording the collapsed measurement data MD with an acceleration factor (PAT=4), one of the segments S1, S2, S3, S4. The separate recording of reference measurement data Rsms can thus be omitted.

[0064] It is also conceivable that the separation data TD are based on reference measurement data RMD from at least two of the segments in which the reference measurement data RMD were recorded. The resulting higher number of recorded reference measurement data increases, for example, the SNR. Furthermore, such an inclusion of more recorded reference measurement data RMD than in a usual recording of reference measurement data Rsms can bring about a noticeable improvement in the separation data TD, especially if the otherwise usual recording of reference measurement data Rsms only includes reference measurement data along a few k-space rows or if measurement data from a large number of slices (e.g. >4 or >8) are recorded simultaneously in a collapsed manner. This applies in particular if separation data TD, e.g. in the form of slice-GRAPPA kernels, are calculated separately for odd and even k-space rows (as is described, for example, in the article by Setsompop et al., “Improving diffusion MRI using simultaneous multi-slice echo planar imaging” NeuroImage 63: pp. 569-580, 2012, particularly with reference to the . Fig. described) and therefore the effective number of k-space lines along which reference measurement data RMD are recorded is halved per slice-GRAPPA kernel.

[0065] In this case, reference measurement data of the at least two segments can be combined, for example, into a combination data set KD such that a sampling pattern of the combination data set KD continues to correspond to the acceleration factor to which the sampling pattern of the collapsed measurement data also corresponds (block 202). The separation data TD can then be generated based on the combination data set KD.

[0066] With such a combination of segments, the procedure can be analogous to the method described in DE102017209988B3.

[0067] Fig. Figure 5 shows an example in k-space of a possible combination data set, as it is created from reference measurement data of the segments S1, S2, S3, S4 of the Fig. 4 could be combined. For this purpose, the sampling patterns of the segments S1, S2, S3, S4 were joined together (in the phase encoding direction kp) in such a way that the k-space lines of the segments S1, S2, S3, S4 are parallel and superimposed in the kp direction with a distance corresponding to the desired sampling pattern, whereby the k-space of the combination Dante set KD is extended in the phase encoding direction.

[0068] When combining segments of the reference measurement data (RMD), care is taken to ensure that the sampling pattern is essentially preserved, so that it continues to correspond to the acceleration factor that the sampling pattern of the collapsed measurement data also corresponds to. Particular attention must be paid to the interfaces between the segments, which must be carefully selected and / or manipulated accordingly to ensure that the sampling pattern is retained as desired.

[0069] In the Fig. In the example shown in Figure 5 with an even number (here 10) of k-space lines per segment, the segments cannot be joined together without further manipulation without disturbing the scanning pattern. In the illustrated case, an interface V occurs between two segments where the scanning pattern is disturbed.

[0070] To eliminate this interference, when combining segments of the reference measurement data RMD, a subset of the reference measurement data RMD of one of the combined segments can be discarded, in particular in such a way that the desired sampling pattern is retained. In the example of Fig. 5, either the last k-space line of segment S2 or the first k-space line of segment S3 could be discarded at the interface V, whereby the position of segments S1 and S2 can be adapted to the position of segments S3 and S4 in the kp direction according to the desired scanning pattern.

[0071] Alternatively, when combining segments of the reference measurement data RMD, reference measurement data RMD of different segments, in particular reference measurement data in edge areas of the segments and thus at possible interfaces between the segments, can be offset against each other in such a way that the desired scanning pattern is retained. In the example of Fig. 5, the last k-space line of segment S2 could be offset against the first k-space line of segment S3 at the interface V, e.g. averaged and shifted to the same position in the kp direction.

[0072] Furthermore, when combining segments of the reference measurement data RMD, additional “empty” k-space lines, along which no reference measurement data RMD is recorded, can be inserted at interfaces between segments in order to compensate for possible disturbances of the desired scanning pattern so that the desired scanning pattern is maintained.

[0073] Fig. Figure 6 shows an example in k-space of a possible combination data set, such as could be combined from reference measurement data of segments S1*, S2*, S3*, S4*. Segments S1*, S2*, S3*, S4* each have an odd number of k-space lines, so that without further manipulation, it is possible (as shown) to combine segments S1*, S2*, S3*, S4* in such a way that the desired scanning pattern is also retained at the interfaces between segments S1*, S2*, S3*, S4*.

[0074] It is also conceivable that the separation data TD are based on reference measurement data RMD, which are determined using a sliding window technique applied to a combination data set KD. In this way, not all of the reference measurement data RMD of the combined segments are used as the basis for the separation data TD, but only a selected portion of them.

[0075] In Fig. 5 is an example of a sliding window SW that can be moved in the kp direction to a desired position to select the k-space lines to be used, along which the reference measurement data RMD to be used as the basis for the separation data TD were recorded.

[0076] The examples of Fig. 5 and Fig. 6 show combined data sets from segments of a reference measurement data acquisition for a DPG technique. Similarly, combination data sets can also be combined from segments of a reference measurement data acquisition for a parallel acceleration technique. Generally, no attention needs to be paid to the readout direction used, but attention must still be paid to any necessary manipulation of the reference measurement data at the segment interfaces. For example, the segments S1', S2', S3' of the Fig. 3 are arranged and manipulated one above the other in a manner analogous to the procedure described above in order to form a combination data set.

[0077] It is also conceivable that separation reference measurement data TRD for separating the collapsed measurement data MD into single-slice measurement data eMD are recorded in addition to the reference measurement data RMD (Block 201*). This can be done in a known manner, e.g., as with the reference measurement data Rsms from Fig. 1. The separation data TD can then also be based on the recorded separation reference measurement data TRD. In this way, the time required for recording the separation reference measurement data TRD (e.g. in Fig. 1, the time Tsms) is not saved, but the database on which the separation data TD are determined and the SNR are increased, which can increase the quality of the separation data TD.

[0078] In this case, the generation of the separation data TD can, for example, comprise an averaging of reference measurement data RMD with separation reference measurement data TRD and / or a comparison of the reference measurement data RMD with the separation reference measurement data TRD with regard to artifacts contained in the respective data can be included, based on the result of which it can be selected which data of the reference measurement data RMD and / or the separation reference measurement data TRD are used in the generation of the separation data TD, whereby, for example, movement artifacts that can be caused by a breathing movement, for example, can be taken into account.

[0079] The measurement data MD acquired in a collapsed and simultaneous manner from multiple slices are separated into single-slice measurement data eMD of the individual simultaneously acquired slices using the generated separation data TD (block 207). The separation of the collapsed measurement data MD into single-slice measurement data eMD can be performed, for example, using a slice-GRAPPA method.

[0080] Using the acquired reference measurement data RMD (complete according to Nyquist), further completed single-slice measurement data eMD* can be generated by reconstructing missing k-space data in the single-slice measurement data eMD (due to the in-plane acceleration technique used to acquire the collapsed measurement data MD) (Block 209). This can be done in a generally known manner, e.g., using a GRAPPA technique.

[0081] In this case, the generation of the separation data TD can, for example, comprise an averaging of reference measurement data RMD with separation reference measurement data TRD and / or a comparison of the reference measurement data RMD with the separation reference measurement data TRD with regard to artifacts contained in the respective data can be included, based on the result of which it can be selected which data of the reference measurement data RMD and / or the separation reference measurement data TRD are used in the generation of the separation data TD, whereby, for example, movement artifacts that can be caused by a breathing movement, for example, can be taken into account.

[0082] Analogous to the generation of separation data TD, if separation reference measurement data TRD has been recorded, the completion of the single-slice measurement data eMD to form completed single-slice measurement data eMD* can also include an averaging of reference measurement data RMD with separation reference measurement data TRD and / or a comparison of the reference measurement data RMD with the separation reference measurement data TRD with regard to artifacts contained in the respective data, based on the result of which it can be selected which data of the reference measurement data RMD and / or the separation reference measurement data TRD are used in the generation of the separation data TD, whereby, for example, movement artifacts that can be caused by a respiratory movement can be taken into account.

[0083] From completed single-slice measurement data eMD* and / or, e.g. using at least one appropriately trained function, from the single-slice measurement data eMD, image data BD of the respective slices can be reconstructed in a conventional manner (block 211).

[0084] With the method according to the invention, the acquisition of separation reference measurement data (TRD) can thus be eliminated, which saves the overall required acquisition time. Additionally or alternatively, the quality of the separation data can be significantly improved by including more reference measurement data than usual and / or by including separation reference measurement data (TRD) that has already been recorded.

[0085] Fig. 7 schematically illustrates a magnetic resonance system 1 according to the invention. This comprises a magnet unit 3 for generating the basic magnetic field, a gradient unit 5 for generating the gradient fields, a radio-frequency unit 7 for irradiating and receiving radio-frequency signals, and a control device 9 designed to carry out a method according to the invention.

[0086] In the Fig. In Figure 7, these subunits of the magnetic resonance system 1 are shown only in a rough schematic. In particular, the radio-frequency unit 7 can consist of several subunits, for example, several coils such as the schematically shown coils 7.1 and 7.2, or more coils, which can be configured either only for transmitting radio-frequency signals or only for receiving the triggered radio-frequency signals, or for both.

[0087] To examine an examination object U, for example, a patient or a phantom, it can be placed on a couch L in the magnetic resonance system 1 within its measurement volume. The schematically illustrated slices s1 and s2 represent exemplary slices of a target volume of the examination object that can be acquired individually or simultaneously (using a slice multiplexing method), from which echo signals can be recorded and acquired as measurement data.

[0088] The control device 9 serves to control the magnetic resonance system 1 and can, in particular, control the gradient unit 5 by means of a gradient controller 5' and the radio-frequency unit 7 by means of a radio-frequency transmit / receive controller 7'. The radio-frequency unit 7 can comprise several channels on which signals can be transmitted or received.

[0089] The radio-frequency unit 7, together with its radio-frequency transmit / receive control 7', is responsible for generating and transmitting (transmitting) a radio-frequency alternating field for manipulating the spins in a region to be manipulated (for example, in slices S to be measured) of the object U under investigation. The center frequency of the radio-frequency alternating field, also referred to as the B1 field, is generally set as close as possible to the resonant frequency of the spins to be manipulated. Deviations from the center frequency to the resonant frequency are referred to as off-resonance. To generate the B1 field, controlled currents are applied to the RF coils in the radio-frequency unit 7 by means of the radio-frequency transmit / receive control 7'.

[0090] Furthermore, the control device 9 comprises a reference measurement data unit 15, with which a recording of reference measurement data according to the invention can be controlled. The control device 9 is designed overall to carry out a method according to the invention.

[0091] A computing unit 13 included in the control device 9 is configured to perform all the computational operations required for the necessary measurements and determinations. Intermediate results and results required for this purpose or determined in this process can be stored in a memory unit S of the control device 9. The units shown are not necessarily to be understood as physically separate units, but merely represent a subdivision into meaningful units, which, however, can also be realized, for example, in fewer or even in only a single physical unit.

[0092] Via an input / output device I / O of the magnetic resonance system 1, control commands can be sent to the magnetic resonance system, e.g. by a user, and / or results of the control device 9, such as image data, can be displayed.

[0093] A method described herein may also be in the form of a computer program comprising instructions that execute the described method on a control device 9. Likewise, a computer-readable storage medium may be present comprising instructions that, when executed by a control device 9 of a magnetic resonance system 1, cause the system 1 to execute the described method.

[0094] Regardless of the grammatical gender of a particular term, persons with male, female or other gender identities are included. QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] DE 102016218955

[0015] US 6043651

[0020] US 9329254B2

[0021] DE 102017209988B3

[0066] Cited non-patent literature

[0000] Breuer et al. in “Controlled Aliasing in Parallel Imaging Results in Higher Acceleration (CAIPIRINHA) for Multi-Slice Imaging”, Magnetic Resonance in Medicine 53, 2005, pp. 684-691

[0010] Setsompop et al. in “Blipped-Controlled Aliasing in Parallel Imaging for Simultaneous Multislice Echo Planar Imaging With Reduced g-Factor Penalty”, Magnetic Resonance in Medicine 67, 2012, pp. 1210-1224

[0010] Zahneisen et al.: „Three-Dimensional Fourier Encoding of Simultaneously Excited Slices: Generalized Acquisition and Reconstruction Framework“, Magn. Reson. Med. 71, S. 2071-2081 (2014

[0017] von Hoge et al, „Dual-Polarity GRAPPA for Simultaneous Reconstruction and Ghost Correction of Echo Planar Imaging Data“, Magn. Reason. Med. 76: S. 32-44, 2016

[0024] Polimeni et al. „Reducing Sensitivity Losses Due to Respiration and Motion in Accelerated Echo Planar Imaging by Reordering the Autocalibration Data Acquisition“, Magn. Reson. Med. 75, S. 665-679, 2016

[0031] Setsompop et al., „Improving diffusion MRI using simultaneous multi-slice echo planar imaging“ NeuroImage 63: S. 569-580, 2012

[0064]

Claims

[1] Method for separating measurement data (MD) of an examination object (U), which are recorded as collapsed from a plurality of slices of the examination object (U) simultaneously and using an acceleration technique in the plane (in-plane technique) with a subsampled sampling pattern of the k-space corresponding to the acceleration technique by means of magnetic resonance technology, into single-slice measurement data (eMD), comprising the steps: - Recording reference measurement data (RMD) in at least two segments in such a way that a total of a Nyquist complete set of reference measurement data (RMD) is available, and that a sampling pattern used when recording the reference measurement data (RMD) within a segment corresponds to an acceleration factor to which the sampling pattern of the collapsed recorded measurement data (MD) also corresponds, - Recording the collapsed measurement data (MD), - generating separation data (TD) based on reference measurement data (RMD) recorded in at least one of the at least two segments, - Separating the collapsed acquired measurement data (MD) into single-slice measurement data (eMD) using the generated separation data (TD). [2] Method according to claim 1, wherein the reference measurement data (RMD) are reference measurement data for a parallel acceleration technique, e.g. in-plane GRAPPA. [3] Method according to claim 1, wherein the collapsed measurement data (MD) are acquired using an EPI acquisition technique and the reference measurement data (RMD) are reference measurement data for a dual polarity GRAPPA (DPG) technique. [4] Method according to one of the preceding claims, wherein the separation data (TD) are based on reference measurement data (RMD) of at least two of the segments in which the reference measurement data (RMD) were recorded. [5] Method according to claim 4, wherein reference measurement data (RMD) of the at least two segments are combined to form a combination data set (KD) in such a way that a sampling pattern of the combination data set (KD) further corresponds to the acceleration factor to which the sampling pattern of the collapsed recorded measurement data (MD) also corresponds, and wherein the separation data (TD) are generated on the basis of the combination data set (KD). [6] Method according to claim 5, wherein when combining segments of the reference measurement data (RMD), a subset of the reference measurement data (RMD) of one of the combined segments is discarded. [7] Method according to one of claims 5 or 6, wherein, when segments of the reference measurement data (RMD) are combined, reference measurement data (RMD) of different segments, in particular reference measurement data (RMD) in edge regions of the segments, are offset against one another, e.g. averaged. [8] Method according to one of claims 5 to 7, wherein the separation data (TD) are based on reference measurement data (RMD) determined by means of a sliding window technique applied to a combination data set (KD). [9] Method according to one of claims 1 to 3, wherein the separation data (TD) are based on the reference measurement data (RMD) of one of the at least two segments in which the reference measurement data (RMD) were recorded. [10] Method according to one of the preceding claims, wherein separation reference measurement data (TRD) for separating the collapsed acquired measurement data (MD) into single-slice measurement data (eMD) are further acquired, and the separation data (TD) are also based on the acquired separation reference measurement data (TRD). [11] The method of claim 10, wherein generating the separation data (TD) comprises averaging reference measurement data (RMD) with separation reference measurement data (TRD). [12] Magnetic resonance system (1) comprising a magnet unit (3), a gradient unit (5), a radio-frequency unit (7) and a control device (9) with a radio-frequency transmission / reception controller (7') and with a reference measurement data unit (15), wherein the control device (9) is designed to carry out a method according to one of claims 1 to 11 on the magnetic resonance system (1). [13] Computer program comprising instructions which, when the program is executed by a control device (9) of a magnetic resonance system (1), cause the system (1) to carry out the method according to one of claims 1 to 11. [14] Computer-readable storage medium, comprising instructions which, when executed by a control device (9) of a magnetic resonance system (1), cause the system (1) to carry out the method according to one of claims 1 to 11.

Images