Method and apparatus for operating a magnetic resonance imaging system

Abstract

A method for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination object is described, wherein k-space data is detected by underscanning within the scope of accelerated echo-planar imaging. The method includes multiple repetitions, each of which includes: a first scanning mode for recording k-space data for Nyquist ghost correction or for generating a magnetic field map; a second scanning mode for subsequent accelerated echo-planar recording; and a common excitation mode for both recordings. The first scanning mode of a proper subset of the multiple repetitions is modified in such a way that, based on the k-space data detected by the modified first scanning mode, k-space data missing due to the underscanning and / or artifacts arising in image space due to the underscanning are supplemented. A control device for a magnetic resonance imaging system is also described. In addition, a magnetic resonance imaging system is described.

Description

Technical Field

[0001] Regardless of the grammatical part of speech of a particular term, people with male, female, or other gender identities are included.

[0002] The present invention relates to a method for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination object, wherein k-space data are detected during accelerated echo-planar imaging by underscanning according to a pulse sequence mode. The present invention also relates to a control device for a magnetic resonance imaging system. Furthermore, the present invention relates to a magnetic resonance imaging system. Background Art

[0003] Imaging systems based on magnetic resonance measurements, particularly nuclear spin methods, known as magnetic resonance tomography (MRI), have been successfully established and proven through widespread application. In this type of image acquisition, a static basic magnetic field B0, used for initial orientation and homogenization of the magnetic dipoles to be examined, is typically superimposed with rapidly switching magnetic fields (so-called gradient fields) to achieve spatial resolution of the imaging signal. To determine the material properties of the object being imaged, dephasing or relaxation times are determined after the magnetization is deflected from the initial orientation, allowing identification of relaxation mechanisms or relaxation times typical of different materials. The deflection is typically performed using multiple radio frequency pulses (the abbreviation HF stands for radio frequency), also called excitation pulses. Spatial resolution is based on the time-determined manipulation of the deflected magnetization using gradient fields in a so-called measurement or control sequence. These sequences determine the precise temporal order of the RF pulses, the changes in the gradient fields (via the switching sequence of the gradient pulses), and the acquisition of the measured values.

[0004] Typically, an intermediate step is used to correlate the measured magnetization (from which the aforementioned material properties can be derived) with the position coordinates of the measured magnetization in the spatial space of the examination subject. In this intermediate step, the acquired magnetic resonance raw data (also referred to as k-space data) are stored at readout points in so-called "k-space," where the k-space coordinates are encoded as a function of the gradient field. The magnitude of the magnetization at a specific location in the examination subject (particularly the magnitude of the transverse magnetization in a plane transverse to the aforementioned basic magnetic field) can be determined from the data at the readout points by means of a Fourier transform, which calculates the signal intensity of the signal in the spatial space from the signal intensity (magnification magnitude) associated with a specific frequency (spatial frequency) or phase.

[0005] Magnetic resonance tomography (MRI) is a relatively slow imaging method because data is recorded sequentially along lines in Fourier space, or k-space, and spin relaxation of the excited spins requires a certain amount of time. Recording images in two-dimensional slices is significantly less error-prone than recording in three dimensions, as the number of encoding steps is reduced. Therefore, many applications use image volumes with stacks of two-dimensional slices rather than a single three-dimensional recording. However, due to the long relaxation times of spins, image recording times are very long, which, for example, reduces comfort for the patient being examined. Even briefly leaving the MRI scanner or simply changing their position during recording is not possible, as this would disrupt the image recording process due to the change in position, and the entire process would have to be restarted. Therefore, accelerating the recording of two-dimensional slice stacks is an important goal.

[0006] For example, parallel image recording techniques are used to speed up image acquisition. Some of these techniques may produce artifacts due to underscanning. These artifacts can be eliminated by applying reconstruction algorithms.

[0007] To accelerate the acquisition of magnetic resonance raw data, echo planar imaging (EPI or EPI pulse sequence) is often used. Here, multiple echoes (also called echo trains or echo trains) are detected per repetition by switching the frequency gradient (also called readout gradient) back and forth. This allows images to be acquired with a small number of excitations or repetitions.

[0008] Echoplanar imaging sequences are often used in the form of multi-slice imaging sequences in order to accelerate imaging and, for example, also enable the recording of moving objects. Despite the presence of the described echo trains, echoplanar imaging sequences typically include a large number of repetitions. A repetition is understood to be the pulse sequence segment between two excitation radiofrequency pulses. In particular, in so-called BOLD imaging (BOLD = "blood oxygen level correlation," this type of imaging illustrates the oxygen content in red blood cells and, for example, can show a patient's brain activity) and extended diffusion protocols such as MDDW (MDDW = multi-directional diffusion weighting), recordings with at least 60 repetitions are not uncommon.

[0009] Typically, during so-called navigator recordings, three lines are recorded at each repetition directly after the excitation pulse to correct for Nyquist ghosts and / or B0 drift. This results in a very high temporal resolution for correcting the aforementioned artifacts over the aforementioned large number of repetitions.

[0010] As already mentioned, EPI recordings are performed using parallel imaging techniques such as GRAPPA (GRAPPA=Generated Autocalibrating Partially Parallel Acquisitions) or SMS (Simultaneous Multi-Slice) in order to image-scan multiple slices simultaneously.

[0011] In addition, there are extended correction methods, such as "Dual Polarity" (see Hoge et al., "Dual-Polarity GRAPPA for Simultaneous Reconstruction and Ghost Correction of Echo Planar Imaging Data", Magnetic Resonance in Medicine 2016 July; 76(1): 32-44. Doi: 10.1002 / mrm.25839. Epub 2015 July 24) for particularly precise correction of ghost artifacts or spatially resolved eddy current correction methods (see Chen et al., "Single-Shot and Segmented EPI Ghost Artifacts Removal with Two-Dimensional Phase Correction", Proc. Intl. Soc. Magn. Reson. Med. 8 (2000)). These specific methods require navigation data for the correction methods mentioned and, in addition, reference data for supplementing the under-scanned k-space data. The required number of repetitions of the pulse sequence is therefore increased, and thus also the measurement time or the time for such an EPI image recording method.

[0012] In "Field-Mapping-Embedded EPI for Geometric Distortion Correction", Proc. Intl. Soc. Mag. Reson. Med. 30 (2022) 1099 by Chi et al., a method is described in which the existing three lines of the navigation recording for detecting Nyquist ghost correction data are modified and continued for the duration of the entire recording so that a further k-space segment to be recorded is encoded in each repetition. The detected k-space segments are combined into a matching k-space and a B0 field map for correcting geometric distortion can be obtained therefrom for the duration of the recording. The polarity of the readout gradient for detecting the correction data is reflected in the kx direction (readout direction) so that k-space data in three monopolar k-spaces are obtained for subsequent image-based calculation of the B0 field map. In order to generate the B0 field map, a reference scan is additionally required to apply a parallel imaging method. The method is illustrated in Figures 1 and 2.

[0013] US Pat. No. 10,162,037 B2 and US Pat. No. 11,280,870 B2 describe methods in which the recording pattern of three lines of a navigator recording is modified in such a way that slice-specific ghost correction data are obtained for SMS imaging. Summary of the Invention

[0014] It is therefore an object of the present invention to propose a time-saving accelerated echo-planar imaging method with correction of the so-called Nyquist effect and with a reduced recording time compared to conventional methods.

[0015] This object is achieved by a method according to the invention for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination object, wherein k-space data are detected during accelerated echo-planar imaging by underscanning according to a pulse sequence mode, by a control device according to the invention for a magnetic resonance imaging system and by a magnetic resonance imaging system according to the invention.

[0016] The method according to the present invention involves scanning k-space according to a pulse sequence pattern. The pulse sequence pattern has multiple repetitions, each of which includes: a first scanning pattern for recording k-space data to perform Nyquist ghost correction and / or to generate a magnetic field map for magnetic field correction, preferably for B0 magnetic field correction, or to generate a so-called B0 field map; a subsequent second scanning pattern for accelerating echo-planar recording; and an excitation pattern common to both recordings. The magnetic field map indicates deviations from a target value of the field strength of the magnetic field generated during magnetic resonance imaging. These deviations result in phase deviations in the measurement signals, which must be taken into account during image reconstruction. A scanning pattern is to be understood as a specific method by which k-space is scanned using gradient circuits, wherein radio frequency signals are received as echo signals or measurement signals. In particular, such a scanning pattern corresponds to a specific k-space trajectory and a specific order of gradient pulses or a specific gradient pulse sequence or gradient pulse subsequence. K-space data is understood to be a measurement signal associated with a specific point in k-space. The unmodified first scanning mode is uncoded, i.e., scanning is performed without phase coding, so that preferably the "zeroth" k-space line is scanned or detected multiple times, particularly preferably three times, in alternating directions. An accelerated echo-planar acquisition is to be understood as one in which time is saved in the acquisition of k-space data by undersampling.

[0017] The first scanning pattern is modified in a proper subset of the multiple repetitions such that, based on k-space data acquired using the modified first scanning pattern, signal information missing due to underscanning is supplemented. In particular, the k-space data acquired using the modified first scanning pattern should be used to supplement the k-space data missing due to underscanning and / or to correct artifacts occurring in image space due to underscanning.

[0018] A subset of a plurality of repetitions that does not include at least one of the repetitions is to be understood as a true subset. Since modifications are not performed in all repetitions, there are also multiple repetitions, or more precisely, at least one repetition of the plurality of repetitions, in which no modification occurs and which is therefore used to acquire k-space data for Nyquist correction. The modified scanning mode is designed so that additional k-space data is detected thereby, which forms the basis for supplementing signal information missing due to accelerated imaging. As already mentioned, this missing signal information includes in particular missing k-space data in under-scanned k-space regions. However, the missing signal information can also include artifacts in image space caused by the missing k-space data. Therefore, in order to acquire a portion of k-space, the modified first scanning mode is preferably encoded compared to the unmodified first scanning mode, i.e., a phase encoding gradient is switched in order to acquire phase encoding lines with different phases.

[0019] As will be explained in more detail below, in order to calculate the missing k-space data based on the additional k-space data, information about the measurement performance of the magnetic resonance system is obtained, with the aid of which the missing k-space data can be reconstructed. In particular, in so-called multi-coil systems, with which accelerated imaging methods are typically performed, so-called autocalibration can be used to obtain information about the weighting or weighting with which the signals of the different coils at a point in k-space must be added in order to calculate the signals of adjacent, possibly unexamined, points in k-space.

[0020] In other words, based on the weighted and undersampled k-space data, it is possible to infer the signal strength in the unsampled subregion of k-space. To this end, a so-called kernel is preferably calculated based on the information determined during the autocalibration, which includes the aforementioned weights. If this kernel is applied to the undersampled k-space data, signal information missing due to the undersampling can be determined.

[0021] Alternatively, the missing signal information can also be supplemented in image space by ascertaining coil sensitivity values which allow the calculation of artifact-free image data based on artifact-laden image data obtained by undersampling with the aid of different coils.

[0022] When reconstructing missing signal information in image space, as is done in the SENSE method, k-space data acquired using a modified first scanning mode is used to determine coil sensitivity values. Initially, artifact-laden image data with a restricted image field (FOV) are reconstructed for each coil based on the underscanned k-space data. This image data typically exhibits folding of prominent image components. Ultimately, artifact-free image data for the complete image field (FOV) can be calculated based on the coil sensitivity values and the artifact-laden image data with a restricted image field.

[0023] That is, the navigation data or k-space data used for Nyquist correction or B0 magnetic field correction are acquired unencoded or without phase encoding, whereas they are encoded, preferably phase encoded, when the k-space data is acquired according to the modified first scanning mode. This phase encoding allows the aforementioned automatic calibration, which in particular includes scanning a subregion of k-space with an increased resolution compared to an underscan, preferably with a full scan. Based on the k-space data scanned with higher resolution, missing k-space data or missing signal information can be determined, wherein the missing signal information can be reconstructed not only in k-space, such as in the GRAPPA method, but also through calculations in image space, such as in the SENSE method. The image space data and the k-space data can be converted into each other by a Fourier transform.

[0024] In the common excitation process, a first slice selection gradient pulse is preferably generated along the slice selection direction. Furthermore, an excitation radio frequency pulse is generated, which includes N excitation frequencies for exciting one or more, preferably N, slices to be excited simultaneously, where N is an integer natural number. As will be explained in detail below, N typically has a value of 2 or a maximum of 8.

[0025] As already mentioned, the readout process for detecting k-space data generally includes alternating switching of rephasing gradient pulses along the readout direction, wherein, for example, further gradient pulses are additionally generated along the phase direction and / or along the slice selection direction and radio frequency signals are received to obtain magnetic resonance raw data or k-space data.

[0026] The method according to the present invention utilizes the recognition that the temporal resolution for determining so-called navigator data for Nyquist ghost correction or B0 field correction does not need to be too high, since the temporal evolution of interferences generally occurs slowly. Therefore, some pulse sequence segments that are typically used to determine the correction data for Nyquist ghost correction or B0 field correction can be used to obtain k-space data as reference data. This reference data is used to correct artifacts arising from underscanning.

[0027] If compensation is made in k-space, such as undersampling in the GRAPPA method, the reference data is used to calculate the k-space data missing for the actual imaging. Since the acceleration of the echo-planar acquisition is based on the undersampling of k-space, the k-space data missing for the complete image information due to the undersampling must be estimated or approximately calculated based on the reference data.

[0028] Advantageously, unlike conventional processing methods, a separate recording process is no longer necessary to acquire reference data, thereby reducing the overall imaging time. Furthermore, there is no need to increase the duration of the repetitions or echo trains, since the recording of the reference data simply replaces the recording of the navigator data in some repetitions, so that the echo times and repetition times of the individual repetitions preferably remain the same or can remain unchanged, regardless of whether the navigator data, data for correcting Nyquist ghosting effects, or reference data are acquired using the first scanning mode.

[0029] That is, compared to conventional EPI pulse sequences in which reference data are acquired separately, the image acquisition duration or the total time for acquiring k-space data is advantageously reduced.

[0030] The present invention further relates to a control sequence for controlling a magnetic resonance imaging system. The control sequence has a control signal corresponding to a pulse sequence pattern, the pulse sequence pattern having a plurality of pulse sequence segments or repetitions, each of which has an excitation segment and a readout segment.

[0031] The repetition includes the aforementioned common excitation pattern as a first pulse sequence subsegment, the also mentioned first scanning pattern for recording a magnetic field map for Nyquist ghost correction or for generating a magnetic field correction, in particular for B0 field correction, as a second pulse sequence subsegment, and the already mentioned subsequent second scanning pattern for accelerated echo-planar recording as a third pulse sequence subsegment. The first scanning pattern of the plurality of repeated proper subsets is modified so that, based on the k-space data acquired with the aid of the modified scanning pattern, missing k-space data due to undersampling and / or artifacts arising in image space due to undersampling are corrected. The control sequence according to the present invention shares the advantages of the method according to the present invention for controlling a magnetic resonance imaging system for generating magnetic resonance imaging data of an examination object, wherein the k-space data are acquired with the aid of undersampling during accelerated echo-planar imaging.

[0032] The control device according to the invention for a magnetic resonance imaging system is configured to control the magnetic resonance imaging system by using the method according to the invention for operating a magnetic resonance imaging system to generate magnetic resonance imaging data of an examination object, wherein k-space data are detected by means of underscanning during accelerated echo planar imaging. The control device according to the invention shares the advantages of the method according to the invention for operating a magnetic resonance imaging system to generate magnetic resonance imaging data of an examination object, wherein k-space data are detected by means of underscanning according to a pulse sequence mode during accelerated echo planar imaging.

[0033] The magnetic resonance imaging system according to the invention comprises a control device according to the invention. The magnetic resonance imaging system according to the invention shares the advantages of the method according to the invention for operating a magnetic resonance imaging system to generate magnetic resonance imaging data of an examination object, wherein the k-space data are detected in an aspect of accelerated echo-planar imaging by means of undersampling according to a pulse sequence mode.

[0034] Most of the aforementioned components of the method for operating a magnetic resonance imaging system according to the present invention can be implemented completely or partially as software modules in a processor of a corresponding computing system, for example, in the form of a control device for a magnetic resonance tomography system or a computer for controlling such a system. A software implementation, if possible, has the advantage that even previously used computing systems can be easily retrofitted via a software update to operate in the manner according to the present invention. This object is also achieved by a corresponding computer program product having a computer program that can be directly loaded into a computing system and includes program segments that, when executed in the computing system, implement the steps of the method for operating a magnetic resonance imaging system according to the present invention. In addition to the computer program, such a computer program product can optionally include additional components, such as documentation and / or additional components, as well as hardware components, such as a hardware key (dongle, etc.) for using the software.

[0035] For transport to a computing system or control device and / or for storage there or in a computing system or control device, a computer-readable medium such as a memory stick, hard disk or other transportable or permanently mounted data carrier on which program segments of a computer program readable and executable by a computing system are stored can be used. For this purpose, a computing system can, for example, have one or more cooperating microprocessors or the like.

[0036] The following descriptions respectively contain particularly advantageous embodiments and improvements of the present invention. In addition, within the scope of the present invention, various features of different embodiments can also be combined to form new embodiments.

[0037] In a preferred embodiment of the method for operating a magnetic resonance imaging system according to the present invention, when modifying the first scanning mode, the duration of the repetitions affected by the modification, and preferably also the duration of the second pulse sequence segment that reproduces the first scanning mode, remains unchanged. Advantageously, a stationary state of the spins with respect to their magnetization deflection, also referred to as a "steady state," is maintained. This "steady state" is particularly necessary in pulse sequences in which the repetition time or echo time is shorter than the relaxation time of the excited spins. Advantageously, image artifacts caused by a disruption of the steady state are avoided.

[0038] If, in a variant of the method according to the invention for operating a magnetic resonance imaging system, a change in the duration of the relevant repetitions is caused by modifying the first scanning mode, it may be necessary or advantageous to compensate for this change in duration by so-called dummy pulses or preparation pulses in order to achieve correct detection of the k-space data or the radio frequency resonance signal.

[0039] For example, in this modification of the first scanning mode, the number of k-space lines scanned per repetition in the modified first scanning mode is increased as the duration changes compared to the unmodified first scanning mode, so that an enlarged k-space region can be scanned in this way to obtain reference data.

[0040] Particularly preferably, all pulse sequence segments or repetitions of a plurality of repetitions of an EPI pulse sequence used for imaging comprise the first scanning pattern or a modified first scanning pattern. Advantageously, a maximum amount of k-space data is obtained for correcting artifacts and supplementing under-scanned k-space data.

[0041] More particularly preferably, only one of these repetitions, preferably the first repetition, has the first scanning mode, and the remaining or all remaining repetitions have a modified first scanning mode, in order to maximize the collection of k-space data used to supplement the undersampled k-space data. In this variant, the resulting image quality is further improved due to the wider data base for ascertaining the missing k-space data due to the undersampling. In this variant, only the first repetition is used to record uncoded k-space data, preferably navigator data or corresponding k-space data for correcting Nyquist ghosting correction and / or B0 field correction.

[0042] However, it is also very advantageous to use a plurality of repetitions distributed over the entire pulse sequence pattern for acquiring uncoded k-space data, preferably navigator data, for correcting Nyquist ghost correction and / or B0 field correction in order to update the correction factors used for such corrections. Advantageously, the correction factors can be adapted to dynamic changes that influence the appearance of artifacts, such as Nyquist ghost structures.

[0043] Preferably, a first scanning pattern in an EPI pulse sequence in different repetitions is replaced by a differently modified scanning pattern.

[0044] In this case, a first scanning mode modified differently in different repetitions is used to scan k-space. For example, individual gradients for scanning different k-space regions must be varied. Furthermore, as will be explained in more detail below, it may also be necessary to calculate missing spatial data using multiple differently structured cores acquired using different scanning modes, for example, when different acquisition techniques are to be combined or when specific artifacts need to be corrected particularly precisely.

[0045] In a preferred variant of the method for operating a magnetic resonance imaging system according to the present invention, the modified first scanning pattern and the number of repetitions included in the first subset are selected such that at least the k-space subregion required to completely fill in the missing k-space data is completely scanned. To determine the k-space data initially missing during underscanning, a so-called kernel is preferably calculated based on k-space data from the completely scanned k-space region. In this case, this is the region of k-space scanned using the modified first scanning pattern.

[0046] If the first scanning mode is designed for scanning with three phase-space encoding lines and this number does not change when modified, then a typical value for generating the core to supplement the k-space data in the underscanned image acquisition method is 24 phase-space encoding lines, for which 8 repetitions are required when the GRAPPA scanning mode is selected, or 16 repetitions are required when the SMS scanning mode is selected. If scanning with M phase-space encoding lines is selected to correct for Nyquist ghosting effects (M being a natural number), correspondingly 24 / M modified repetitions are required for the GRAPPA scanning mode and 48 / M modified repetitions are required for the SMS scanning mode.

[0047] Therefore, in the method according to the present invention, the modified first scanning pattern preferably includes scanning at least three k-space lines along the readout direction with different phase encodings. Advantageously, three k-space lines are sufficient for correction in three spatial directions with N / 2 Nyquist ghost correction. When modifying the first scanning pattern, in the advantageous variant, the three gradient pulses along the readout direction do not need to be modified. It should be noted that the first scanning pattern can also include scanning only one or two k-space lines, or more than three k-space lines, along the readout direction. The scanning of k-space lines according to the first unmodified scanning pattern should specifically include variants in which the same k-space line may be scanned multiple times in both directions, particularly preferably without encoding. If the number of k-space lines scanned in the modified first scanning pattern is to differ from the number of k-space lines scanned in the unmodified first scanning pattern, then the optional, still additional, modifications mentioned above, are useful as compensation so that the echo time and repetition time of the modified repetitions do not change, or change only slightly, compared to the unmodified repetitions.

[0048] Particularly preferably, the modified first scanning mode of the method according to the invention includes scanning preferably at least three k-space lines along the readout direction with at least partially different slice edge codings. Such different slice codings are particularly effective in image acquisition methods using simultaneous slice imaging techniques, in particular for SMS, GRAPPA+SMS, and bipolar GRAPPA, for the reasons already mentioned, in order to improve the image quality.

[0049] More particularly preferably, similarly structured scan patterns are used for the modified first scan pattern and the second scan pattern, in other words, the same type of scan pattern is used. If the reference data, i.e., the k-space data, based on which the underscanned k-space data are to be supplemented, are generated, and if the same type of scan pattern is used for scanning the underscanned k-space data itself, interference effects and geometric distortions that may occur in different scan patterns for obtaining the reference data and the underscanned k-space data can be avoided or reduced. Similar scan patterns are defined in this case by the fact that for two temporally sequentially scanned k-space lines, the spacing in k-space between the reference data and the underscanned data recorded in one repetition is the same, or in other words, the reference data in one repetition have the same underscanning as the underscanned (imaging) data. However, the reference data then, of course, again generate the fully scanned k-space in a manner consisting of all repetitions.

[0050] In a particularly effective embodiment of the method according to the present invention for operating a magnetic resonance imaging system, k-space data or reference data detected by a modified first scanning mode are updated at different time points during accelerated echo-planar imaging. Advantageously, the data basis for supplementing the under-scanned k-space data can be adapted to dynamic effects, in particular interference influences such as phase changes or patient movement. This adaptation advantageously allows the same image quality to be maintained throughout the entire imaging process. To this end, a complete scan of the reference data up to a first time point can be performed and repeated one or more times at subsequent time points. In graphical terms, a moving average of the reference data can be determined based on multiple reference data sets, and the k-space data to be supplemented can be determined based on this moving average. Alternatively, if the multiple reference data sets are not averaged until all reference data sets have been recorded, or if the missing k-space data are not reconstructed until all reference data sets have been recorded, it is also possible to determine "stable" or constant reference data after multiple reference data recordings.

[0051] If so-called kernels are derived based on reference data in order to determine the missing k-space data for accelerated imaging, multiple such updated kernels can be generated during the acquisition or imaging process using a moving average. This has the advantage of maintaining a consistently high image quality for the calculation of the missing k-space data, even when the aforementioned dynamic interference effects occur.

[0052] Alternatively, a single stable core can also be obtained at the end of the acquisition by scanning or acquiring a portion of the k-space required for the reference data, such as the k-space center, or the entire set of reference data multiple times during the measurement and subsequently calculating the core based on the ascertained reference data.

[0053] In a particularly practical embodiment of the method according to the invention, the modified first scanning pattern includes at least four consecutive phase encoding gradients and at least three consecutive readout gradients with alternating polarization, the readout gradients being temporally shifted along with the phase encoding gradients. Advantageously, at least three different k-space lines are detected in each repetition for calculating the k-space data to compensate for undersampling, thereby accelerating image acquisition.

[0054] Preferably, the second to fourth phase encoding gradients comprise a positive polarization.In this variant, the scanning direction along the phase encoding direction is maintained during a repetition.

[0055] If k-space data are to be acquired simultaneously in multiple slices, the slice selection gradients are preferably switched in temporally simultaneously via the second to fourth phase encoding gradients. Advantageously, the acquisition time of the magnetic resonance imaging can be reduced by the simultaneous acquisition.

[0056] Very particularly preferably, the slice selection gradient has alternating polarity.The scanning mode is particularly suitable for SMS, GRAPPA+, SMS or bipolar GRAPPA imaging methods.

[0057] If an imaging method with underscanning in the phase encoding direction is selected, such as the GRAPPA mode or a combination of the GRAPPA mode and the SMS mode, the amplitude of the modified phase encoding gradient of the first scanning mode is preferably selected as a function of the underscanning factor of the second scanning mode. Preferably, the amplitude is selected such that, for an underscanning factor with the value F (F being a natural number), F-1 lines are skipped.

[0058] In a preferred embodiment of the method according to the invention for operating a magnetic resonance imaging system, the following special imaging technique is used for the accelerated echo planar recording:

[0059] -GRAPPA,

[0060] -SMS,

[0061] -SENSE,

[0062] - A combination of GRAPPA and SMS,

[0063] Compressed SENSING-based or SENSE-based imaging methods with the aid of deep learning reconstruction methods, which likewise require coil sensitivity information for implementation.

[0064] In Σ-net: Systematic Evaluation of Iterative Deep Neural Networks for Fast Parallel MR Image Reconstruction, Kerstin Hammernik et al., https: / / arxiv.org / abs / 1912.09278, compressed SENSING or SENSE-based imaging and deep learning reconstruction methods are described.

[0065] Advantageously, during imaging using the GRAPPA method, a Nyquist ghost correction line is first determined using a first scanning mode, and by applying a modified first scanning mode, reference lines for GRAPPA calibration, i.e., for calculating the kernel to supplement the under-scanned k-space data, are additionally determined. By modifying the phase encoding gradient (Gy gradient), a subsegment of k-space to be acquired is selected, and the image is further moved within the navigation system formed by the modified first scanning mode using additional phase blips or smaller phase encoding gradients. With the described method, a k-space with 24 phase encoding lines, which is highly suitable for GRAPPA calibration, can be obtained using the modified first scanning mode with eight repetitions, with three phase encoding lines detected per repetition. In addition to calibrating the GRAPPA kernel, the acquired k-space data can also be used to calibrate the SENSE method and, in particular, to obtain coil sensitivity maps for deep learning reconstruction methods.

[0066] Advantageously, during imaging using the SMS method using the first scanning mode, Nyquist ghost correction lines can be obtained, and reference lines for SMS calibration, i.e., for calculating the core to compensate for the under-scanned k-space data, can be obtained using the modified first scanning mode. Advantageously, reference lines for SMS core calibration are obtained in this manner.

[0067] In a special embodiment of the method based on SMS scanning technology, reference data for the Slice GRAPPA core calculation is determined. Slice GRAPPA is a conventional reconstruction method for SMS recordings, see, for example, https: / / onlinelibrary.wiley.com / doi / full / 10.1002 / mrm.23097. The recorded data points are selected or arranged so that slice blips, such as data shifts along the slice encoding direction (kz direction), can be processed. This processing method is described more precisely in US Patent No. 10557903 B2.

[0068] If adjacent k-space points along the ky and kz directions are represented by successive integers 1, 2, 3, etc., then when three phase encoding lines are detected using the first scanning mode, the data points kz=1 / ky=1, kz=2 / ky=2, kz=1 / ky=3 are recorded in the first repetition while modifying the first scanning mode. This is achieved by alternating spikes along the kz direction and unipolar spikes along the ky direction. The spikes are gradient pulses, which represent short trajectory segments. In subsequent repetitions, supplementary k-space segments kz=2 / ky=1, kz=1 / ky=2, kz=2 / ky=3 are recorded. In this case, the polarity of the kz spikes is rotated accordingly. Subsequently, further subsegments are recorded. Alternatively, it is also possible to record "in rows" along the kz direction, i.e., to perform a single movement along the kz axis before the first echo and then record ky=1, 2, 3 with fixed kz values. As described in US Pat. No. 10557903 B2, slice-specific calibration data can then be obtained by Fourier transformation along the kz axis. To obtain a core with 24 k spatial lines, 16 repetitions are required when 3 phase encoding lines are detected per repetition using the first scanning mode.

[0069] Advantageously, during imaging using the combined GRAPPA+SMS method, the Nyquist ghost correction line can be obtained by the first scanning mode, and the reference line for GRAPPA+SMS calibration, ie, for calculating the kernel to supplement the under-scanned k-space data, can be obtained by the modified first scanning mode.

[0070] Advantageously, reference lines are obtained for combined GRAPPA+SMS calibration. This allows reference data to be obtained for both GRAPPA and SMS. Compared to the variant for pure SMS imaging, this variant is modified as follows: the spacing of the phase encoding lines in the ky direction, i.e., in the phase encoding direction, corresponds to the undersampling factor, for example, 2. To calibrate the so-called Slice-GRAPPA core, the detected k-space data is rearranged according to DE 10 2017 209 988 B3 so that the undersampling acquisition mode is maintained. In contrast, to calibrate the conventional GRAPPA core, the phase encoding lines are received as blocks, as already described above.

[0071] In the various underscanning methods described above, other arrangements of the k-space points acquired in the repetitions are also conceivable, as long as a completely scanned k-space region is ultimately achieved.

[0072] For the GRAPPA+SMS calibration, for two cores with 24 k-space lines each, 32 repetitions are already required when detecting 3 phase encoding lines per repetition using the first scanning mode.

[0073] If, for example, in a bipolar GRAPPA imaging method, the second scanning pattern comprises a scanning with bipolarity, then the modified first scanning pattern comprises at least two scanning patterns with mutually inverted readout polarities for each scanned phase. For this purpose, preferably two successive repetitions can be used, which have differently modified first scanning patterns with mutually inverted readout polarities of the phase encoding gradients and otherwise identical gradient patterns.

[0074] This means that, compared to the standard GRAPPA method, in this first sub-variant the number of repetitions required for determining the reference data is increased by a factor of two.

[0075] In a second sub-variant of the bipolar GRAPPA imaging method, k-space data are acquired using a modified first scanning mode with inverted readout polarities by scanning each phase multiple times, preferably three times, with alternating polarities. This means that in this sub-variant, only phase encoding lines are acquired during each repetition of recording reference data, tripling the number of repetitions required for recording reference data compared to the unmodified first scanning mode, where three phase encoding lines are acquired during each repetition to detect Nyquist ghost correction data.

[0076] If an imaging method for simultaneously reading out a plurality of slices, such as the aforementioned imaging method, is used as the accelerated EPI imaging method in the method according to the invention, then in a particularly efficiently implementable embodiment of the invention the number N of slices to be read out simultaneously is equal to two.

[0077] Alternatively, the number N of slices to be read simultaneously can also be equal to 3. In principle, it is optimal to read as many slices as possible simultaneously. However, the number of slices to be read simultaneously is limited by the fact that the energy delivered to the patient per pulse is proportional to the number of slices N. However, the permissible energy input per unit time is limited. Furthermore, the more difficult the separation in image reconstruction, the more slices are read simultaneously, which increases the noise in the separated images. This increase appears nonlinear. Therefore, in practice, only a small number of slices can be read simultaneously. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] The present invention will be described in more detail below based on an embodiment with reference to the accompanying drawings.

[0079] FIG. 1 shows a diagram illustrating the conventional working manner of Nyquist ghost correction,

[0080] 2 shows a diagram illustrating a conventional method for determining a B0 field map, wherein additional k-space segments to be acquired are encoded with the aid of phase encoding gradients in each repetition and are subsequently combined to form a suitable k-space.

[0081] FIG. 3 shows a diagram illustrating a pulse sequence for carrying out the imaging method illustrated in FIG. 2 ,

[0082] Figure 4 A mode for acquiring reference lines for GRAPPA calibration in a method for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination object is shown according to one exemplary embodiment of the present invention, wherein k-space data are detected by means of underscanning during accelerated echo planar imaging.

[0083] Figure 5 The following diagram is shown, which shows a method for implementing the echo planar imaging method in Figure 4 The pulse sequence of the scanning mode is illustrated in FIG.

[0084] Figure 6 An alternative mode for acquiring reference lines for GRAPPA calibration in a method for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination object according to one embodiment of the present invention is shown, wherein k-space data are detected by means of underscanning during accelerated echo planar imaging.

[0085] Figure 7 The following diagram is shown, which shows a method for implementing the echo planar imaging method in Figure 6 The pulse sequence of the scanning mode is illustrated in FIG.

[0086] Figure 8 A mode for acquiring reference lines for SMS calibration in a method for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination object according to one embodiment of the present invention is shown, wherein k-space data are detected by means of underscanning during accelerated echo planar imaging.

[0087] Figure 9 The following diagram is shown, which shows a method for implementing the echo planar imaging method in Figure 8 The pulse sequence of the scanning mode is illustrated in FIG.

[0088] Figure 10 A mode for acquiring reference lines for combined GRAPPA+SMS calibration in a method for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination object according to one exemplary embodiment of the present invention is shown, wherein k-space data are detected by means of underscanning during accelerated echo planar imaging.

[0089] Figure 11 The following diagram is shown, which shows a method for implementing the echo planar imaging method in Figure 10 The pulse sequence of the scanning mode is illustrated in FIG.

[0090] Figure 12A mode for acquiring reference lines of a first core for bipolar GRAPPA calibration in a method for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination object according to one embodiment of the present invention is shown, wherein k-space data are detected by means of underscanning during accelerated echo planar imaging.

[0091] Figure 13 The following diagram is shown, which shows the method for implementing Figure 12 The pulse sequence of the scanning mode is illustrated in FIG.

[0092] Figure 14 A mode for acquiring a reference line of a second core for bipolar GRAPPA calibration in a method for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination object according to one embodiment of the present invention is shown, wherein k-space data are detected by means of underscanning during accelerated echo planar imaging.

[0093] Figure 15 The following diagram is shown, which shows the method for implementing Figure 14 The pulse sequence of the scanning mode illustrated in FIG, and

[0094] Figure 16 A magnetic resonance imaging system according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0095] FIG1 shows a flow chart of the Figure 10 The diagram illustrates a conventional procedure for performing Nyquist ghosting correction in magnetic resonance imaging. In step 1.I, during phase correction acquisition, phase encoding lines are acquired in the center of k-space during the first three repetitions REP1, REP2, and REP3. The unencoded phase encoding lines acquired in the center of k-space during the first three repetitions REP1, REP2, and REP3 are then averaged in step 1.II to obtain k-space data for the navigation acquisition NA. Based on the averaged k-space data, the Nyquist effect is corrected or eliminated from the image data in step 1.III. Step 1.III is also symbolically represented below the flowchart by a diagrammatic view, with an artifact-laden image of the brain shown on the left and the corrected image on the right. In this conventional variant, the Nyquist correction line is scanned only during the first three repetitions, and the acquired correction coefficients are then applied to all other repetitions of the actual imaging method. The method for correcting the Nyquist effect illustrated in FIG. 1 is described in detail in US Pat. No. 6,043,651 B2.

[0096] At this point, it should be noted that, before the k-space data is scanned for imaging, three uncoded k-space lines with alternating polarity are typically also scanned as navigator data in each repetition. The variant described in FIG1 corresponds to the acquisition of phase-coded k-space lines for the B0 field image described in FIG2 . The repetitions REP4, ..., REPM illustrated in FIG2 follow, in this particular case, the first three repetitions REP1, REP2, REP3 illustrated in FIG1 .

[0097] FIG2 shows a diagram 20 illustrating a conventional method for obtaining a field map based on an image. The method illustrated in FIG2 corresponds to the method described in Chi et al., “Field-Mapping-Embedded EPI for Geometric Distortion Correction,” Proc. Intl. Soc. Mag. Reson. Med. 30 (2022) 1099.

[0098] In this method, additional k-space segments to be acquired are encoded using phase encoding gradients in each repetition, which are then combined to form a suitable k-space (see steps 2.I, 2.II). In step 2.I, a series of repetitions REP4, REP5, ..., REPM are performed, wherein k-space data or raw data are acquired layer by layer using different phase encodings during echo-planar imaging of the examination object. Unlike the method illustrated in FIG1 , in each repetition, in addition to the undersampled k-space data used for imaging, phase encoding lines are also acquired for the subsequent field map generation, with different phase encodings applied for each repetition. In step 2.II, the undersampled k-space lines acquired in step 2.I are rearranged. Furthermore, in step 2.II, reference data REF are additionally acquired, which are subsequently used for a calibration core to determine the k-space data missing in step 2.I during the acquisition of the k-space data. It should be noted that the acquisition of the reference data is performed in a separate acquisition. That is to say, two separate acquisitions are performed, namely one acquisition for acquiring the under-sampled k-space data and the phase encoding lines and a separate acquisition for acquiring the reference data.

[0099] Subsequently, in step 2.III, the missing k-space data are ascertained for the individual slices based on the core and underscanned k-space data ascertained in step 2.II.

[0100] In step 2.IV, image reconstruction is performed slice by slice based on the completed k-space data using a two-dimensional fast Fourier transform (2D-FFT).

[0101] In step 2.V, the phase difference (eg, the phase development between the first and third echo) is calculated based on the multiple reconstructed slice images. The first and third echoes are recorded with the same polarity) and the total phase offset

[0102] Subsequently, the image data from each channel is combined in step 2.VI and the phase is unwrapped ("unwrapped") in step 2.VII to create a field map. This eliminates abrupt phase changes in the image. Because the phase is cyclical through 360 degrees, non-uniformity may occur between pixels (i.e., one phase difference becomes 0, while the next phase difference becomes 360 degrees).

[0103] FIG3 shows a view 30 illustrating a pulse sequence for implementing the method for image-based field map determination illustrated in FIG2 . The view comprises four subviews arranged one above the other, wherein the first subview shows the excitation radio frequency pulses RF, the second subview shows a series of readout gradients Gx, the third subview shows a series of phase encoding gradients Gy, and the fourth subview shows a series of slice selection gradients Gz.

[0104] At this point it should be noted that received radio frequency pulses, which represent received measurement signals, are not illustrated in this diagram.

[0105] The left side of the diagram shows an excitation pattern A-RF with an RF excitation pulse RF and a slice selection gradient Gz displayed simultaneously therewith.

[0106] The scanning pattern PAS, shown between two vertically extending dashed lines in the center of the view, is used to scan phase-shifted phase encoding lines for the navigation recording NA to acquire k-space data for Nyquist correction. To locate the desired output position in k-space, the frequency encoding gradient Gx, slice selection gradient Gz, and phase encoding gradient Gy are initially applied simultaneously. Subsequently, three frequency encoding gradients Gx, alternating in polarity, are applied. During the generation of the phase correction data, the polarity of the readout gradients is reversed to reflect the polarity of the readout gradients in the kx direction on the axis along the direction of the readout gradients Gx. This is necessary to acquire three unipolar k-space images for the subsequent image-based field map calculation.

[0107] At the end of the intermediate scanning mode PAS, the phase coding gradient Gy and the slice coding gradient Gz are switched.

[0108] Shown on the right side of the diagram is an EPI pulse sequence EPI with Slice-GRAPPA scan mode.

[0109] exist Figure 44 shows a mode 40 for acquiring reference lines for GRAPPA calibration in a method for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination object according to one exemplary embodiment of the present invention, wherein k-space data are detected by means of GRAPPA underscanning during accelerated echo planar imaging.

[0110] As already explained in detail in the general section, within the scope of the method according to the invention, three phase encoding lines for correcting the Nyquist effect are recorded without encoding using a first scanning mode (not shown), and the corresponding pulse sequence is displayed in the first repetition within the scope of the imaging process. For the subsequent repetitions, the scanning mode is now modified, as in Figure 4 As shown in .

[0111] For this purpose, the first acquisition mode for acquiring the reference line of the core (which is required to determine the data missing due to undersampling in the GRAPPA imaging method) is modified as follows: additional so-called phase spikes or phase encoding gradients are added. The subsegments REP1, REP2, REP3 of k-space to be acquired are selected by means of the phase encoding gradients or Gy gradients already shown in FIG. 3 and are obtained by means of the phase encoding gradients or Gy gradients in FIG. Figure 5 The additional phase peaks illustrated in FIG continue to move between the individual phase encoding lines of a repetition. With the described method, even with eight repetitions, a k-space with 24 phase encoding lines that is very suitable for GRAPPA calibration can be acquired.

[0112] exist Figure 5 FIG. 50 shows an example of an echo planar imaging method for performing an Figure 4 The pulse sequence of the scanning mode is illustrated in FIG. First, Figure 5 The excitation view A-RF already illustrated in FIG3 is shown in the view 50 illustrated in FIG3. The subsequent modified first scanning view AS1 arranged between the vertical dashed lines is reflected in FIG3. Figure 4 The second repetition REP2 shown in FIG, with the aid of which the reference data should be detected. Figure 4Positioned at the starting point of the trajectory of the second repetition REP2 shown in FIG, frequency encoding gradients Gx, slice selection gradients Gz, and phase encoding gradients Gy are initially displayed simultaneously. Phase encoding gradients Gy, shown one above the other with different amplitudes (a thick line and a dashed line) at the same location, are associated with different repetitions. Depending on the amplitude strength, different phases in k-space are manipulated in different repetitions. Subsequently, similar to the pulse sequence shown in FIG3 , three frequency encoding gradients Gx with alternating polarities follow, each followed by a smaller second or third phase encoding gradient Gy, each representing a phase spike. Finally, the modified first excitation mode AS1 is terminated by an additional frequency encoding gradient Gx and, simultaneously, a fourth phase encoding gradient Gy, to reach the starting point of the trajectory forming the echo train of the actual EPI pulse sequence EPI for echo-planar imaging A-EPI. Here, the phase encoding gradient Gy that terminates the modified first scanning mode is again represented by multiple phase encoding gradients Gy, shown one above the other, each having different amplitudes and associated with different repetitions. The EPI pulse sequence EPI represents the second scanning pattern AS2 and, similar to the first scanning pattern AS1 modified by an increased phase encoding gradient Gy, has a consistent, but negatively polarized, phase encoding gradient Gy.

[0113] It should be noted that in Figure 5 In the embodiment illustrated in FIG, Figure 4 The repetitions REP1, ..., REP3 shown in FIG are preceded by an unmodified repetition (not shown) by which three uncoded phase space lines with alternating polarities are represented as navigation data for Nyquist correction. Figure 5 The phase encoding gradients shown in the first scanning mode, in particular the second and third phase encoding gradients (also called phase spikes), are omitted because the navigator data is recorded uncoded. In other words, in contrast to the conventional method illustrated in FIG2 , the navigator data, reference data, and k-space data are acquired for echo-planar imaging using a single pulse sequence and a common recording. As already mentioned above, the overall recording time can be reduced by integrating the acquisition of reference data into the echo-planar imaging recording process.

[0114] exist Figure 6 6 shows an alternative mode 60 for acquiring reference lines for GRAPPA calibration in a method for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination object according to one exemplary embodiment of the present invention, wherein k-space data are detected by means of undersampling during accelerated echo planar imaging.

[0115] With Figure 4In contrast to the pattern 40 shown in FIG, the successive phase encoding lines of the repetitions REP1, REP2, REP3 are arranged offset by three phase encoding lines in each case.

[0116] exist Figure 7 FIG. 70 is a diagram illustrating a method for implementing an echo planar imaging method in FIG. Figure 6 The pulse sequence of the scanning mode is illustrated in FIG.

[0117] exist Figure 7 First shown in Figures 3 and Figure 5 The excitation pattern A-RF illustrated in FIG. The modified first scanning pattern AS1 following it is reflected in Figure 6 The first repetition REP1 is shown in FIG. Figure 6 The starting point of the trajectory of the first repetition REP1 shown in FIG is positioned, and the frequency encoding gradient Gx, the slice selection gradient Gz and the phase encoding gradient Gy are first displayed simultaneously. Figure 5 Similarly to the pulse sequence shown in FIG, three frequency encoding gradients Gx with alternating polarity are followed, wherein the first two frequency encoding gradients Gx are each followed by a phase encoding gradient Gy, which is referred to as a phase spike. The phase encoding gradients have the same polarity as the corresponding Figure 5 The phase encoding gradient Gy shown in FIG is three times larger than the surface, since it is used to jump over two adjacent phase encoding lines. Finally, the modified first excitation pattern AS1 is terminated by an additional phase encoding gradient to reach the starting point of the trajectory of the echo train that forms the actual EPI pulse sequence EPI. The EPI pulse sequence EPI represents the second scanning pattern AS2 and is similar to the first scanning pattern AS1 modified by the amplified phase encoding gradient Gy, but with a consistent, but negatively polarized, phase encoding gradient.

[0118] exist Figure 8 8 shows a mode 80 for acquiring reference lines for SMS calibration in a method for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination object according to one exemplary embodiment of the present invention, wherein k-space data are detected by means of SMS underscanning during accelerated echo planar imaging.

[0119] The reference data for the Slice-GRAPPA core calculation are shown in the SMS imaging. The data points to be recorded are shown here so that the slice peak Gz can be processed as a shift of the data in the kz direction (see US10557903 B2). Therefore, in the first repetition REP1, the data points kz=1 / ky=1, kz=2 / ky=2, kz=1 / ky=3 are recorded. The scanning order of the different k-space lines within a repetition is symbolically represented by the numbers 1, 2, 3 associated with these k-space lines. This is shown in the figure. Figure 9The kinetic energy is generated by alternating spikes Gz along the kz direction and monopolar spikes Gy along the ky direction.

[0120] In the second repetition REP2, additional k-space segments kz=2 / ky=1, kz=1 / ky=2, and kz=2 / ky=3 are acquired. In this case, the polarity of the kz peak Gz is rotated accordingly. Subsequently, further subsegments of k-space are acquired using the third and fourth repetitions REP3 and REP4.

[0121] Alternatively, the acquisition can also be performed "in lines" along the kz direction, i.e., a single movement is performed before the first echo, and then ky=1, 2, 3 with a fixed kz are acquired. As described in US Pat. No. 10,557,903 B2, slice-specific calibration data can then be generated along kz by Fourier transformation. To obtain a core with 24 k-space lines, 16 repetitions are required.

[0122] exist Figure 9 FIG. 1 shows a method for implementing an echo planar imaging method in FIG. Figure 8 FIG90 illustrates a pulse sequence for a scan mode.

[0123] exist Figure 9 First, it is shown in FIG3, Figure 5 and Figure 7 The excitation pattern A-RF illustrated in FIG. The modified first scanning pattern AS1 that follows is reflected in Figure 8 The first repetition REP1 is shown in FIG. Figure 8 The starting point of the trajectory of the first repetition REP1 shown in FIG is positioned, and the frequency encoding gradient Gx, the slice selection gradient Gz and the phase encoding gradient Gy are first displayed simultaneously. Figure 7 The pulse sequence shown in FIG is similarly followed by three frequency encoding gradients Gx with alternating polarity, wherein the first two frequency encoding gradients Gx are each followed by a positively polarized phase encoding gradient Gy, referred to as a phase spike. The phase encoding gradient Gy has the same polarity as the corresponding Figure 7 The phase encoding gradient Gy shown in is only one-third the size of the surface because it is used to control adjacent phase encoding lines. Simultaneously with the phase encoding gradient Gy, referred to as the phase peak, the alternately polarized slice selection gradient Gz is switched.

[0124] Finally, the modified first excitation pattern AS1 is terminated by an additional phase encoding gradient Gy, with which the final slice selection gradient Gz is simultaneously introduced in order to reach the starting point of the echo train that forms the trajectory of the actual EPI pulse sequence EPI. The EPI pulse sequence EPI represents the second scanning pattern AS2 and, similar to the modified first scanning pattern AS1, has a modified scanning pattern characterized by a uniform, but negatively polarized, phase encoding gradient Gy and a slice selection gradient Gz of alternating polarity.

[0125] exist Figure 10 1 shows a mode 100 for acquiring reference lines for combined GRAPPA+SMS calibration in a method for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination object according to one exemplary embodiment of the present invention, wherein k-space data are detected by means of underscanning during accelerated echo planar imaging.

[0126] By Figure 10 The method illustrated in can obtain reference data for GRAPPA as well as for SMS. Figure 8 The method is modified as follows for the scan pattern 80 illustrated in FIG: : The spacing along the ky direction corresponds to the underscan factor (here 2). The pattern of the bracketed lines (dotted, thin, medium, thick) associated with each repetition REP1, ..., REP4 corresponds to the pattern of the circumferential lines of the k-space line associated with the respective repetition. For calibration of the Slice GRAPPA core, the recorded data are rearranged according to DE 10 2017 209 988 B3 so that the underscanned recording pattern remains unchanged. For calibration of the conventional GRAPPA core, the lines are received as blocks. As in Figure 8 As in the method illustrated in , other arrangements of the k-space points acquired in repetitions REP1, REP2, REP3, REP4 are also conceivable, as long as a completely scanned region is achieved in the end. In this case, 32 repetitions are already required for two cores each with 24 k-space lines.

[0127] exist Figure 11 The diagram illustrates a diagram 110 showing a method for implementing Figure 10 The pulse sequence of the shaping method is illustrated in FIG.

[0128] exist Figure 11 First, it is shown in FIG3, Figure 5 、 Figure 7 and Figure 9 The excitation mode A-RF already illustrated in FIG. The modified first scanning mode AS1 following it is reflected in Figure 10 The first repetition REP1 is shown in FIG. Figure 10At the starting point of the trajectory of the first repetition REP1, the frequency encoding gradient Gx, the slice selection gradient Gz and the phase encoding gradient Gy are first displayed simultaneously. Figure 9 The pulse sequence shown in FIG is similarly followed by three frequency encoding gradients Gx with alternating polarity, wherein the first two frequency encoding gradients Gx are each followed by a positively polarized phase encoding gradient Gy, referred to as a phase spike. The phase encoding gradients have the same polarity as the corresponding Figure 9 The phase encoding gradient Gy shown in FIG has a double area compared to FIG, because the phase encoding gradient is used to cross the phase encoding line in the ky direction and to control the next phase encoding line. Simultaneously with the phase encoding gradient Gy, which is called the phase peak, the phase encoding gradient Gy is connected as in FIG. Figure 9 The slice selection gradient Gz is alternately polarized as in .

[0129] Finally, the modified first excitation pattern AS1 is terminated by an additional phase encoding gradient Gy, with which the final slice selection gradient Gz is simultaneously introduced in order to reach the starting point of the echo train that forms the trajectory of the actual EPI pulse sequence EPI. The EPI pulse sequence EPI represents the second scanning pattern AS2 and, similar to the modified first scanning pattern AS1, has a modified scanning pattern characterized by a uniform, but negatively polarized, phase encoding gradient Gy and a slice selection gradient Gz of alternating polarity.

[0130] exist Figure 12 FIG. 1 shows a mode 120 for acquiring reference lines for a first core for bipolar GRAPPA calibration in a method for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination subject according to one embodiment of the present invention, wherein k-space data are detected by means of underscanning during accelerated echo planar imaging. The scanning of k-space for the first core is very similar to that in FIG. Figure 4 The scanning mode 40 shown in FIG. 4 is performed and will not be described in detail.

[0131] exist Figure 13 FIG. 130 is a diagram illustrating a method for implementing a Figure 12 The pulse sequence of the scanning mode 120 is illustrated in FIG. Figure 13 The pulse frequency shown in Figure 5 The pulse frequencies 50 shown in FIG. 5 are structured similarly or identically and are therefore not explained in detail.

[0132] exist Figure 141 shows a pattern 140 for acquiring reference lines of a second core for bipolar GRAPPA calibration in a method for operating a magnetic resonance imaging system to generate magnetic resonance image data of an examination object according to one embodiment of the present invention, wherein k-space data are detected by means of underscanning during accelerated echo planar imaging. The scanning direction of the phase encoding line is Figure 14 In and in Figure 12 The scanning mode 120 shown in FIG. 1 is implemented in reverse.

[0133] exist Figure 15 FIG. 150 is a view showing the Figure 14 The pulse sequence of the scanning mode is illustrated in FIG. Figure 15 The pulse frequency shown in Figure 13 The pulse frequencies shown in FIG. 1 differ only in that the readout gradient Gx of the modified scanning mode AS1 has the same Figure 13 In contrast, the polarity is reversed and the fourth readout gradient Gx is omitted in the modified first scanning mode AS1 .

[0134] As described by Chi et al. in "Field-Mapping-Embedded EPI for Geometric Distortion Correction", Proc. 30 (2022) 1099, in the illustrated acquisition mode, it is preferred to first perform at least one uncoded navigation recording or a Nyquist ghost correction recording with the aid of a first unmodified scanning mode in order to proceed with the ghost correction. This is also suitable for this purpose in combination with the ghost correction methods already mentioned for SMS imaging (see US 10162037 B2 and US 11280870 B2). If only such recordings for ghost correction are performed, the correction is static, as in a conventional "external phase correction scan", i.e., dynamic changes cannot be incorporated. Since drift effects usually develop slowly in time, it is effective to insert these recordings, for example, eight times and to update the ghost correction coefficients, for example, by means of a sliding average, in order to compensate for the slow drift effects.

[0135] As already explained, it is also conceivable to acquire multiple GRAPPA and / or SMS kernels in very long image acquisitions, for example using a sliding average. This has the advantage that the image quality remains consistently high even in the presence of interfering influences such as phase shifts or patient motion. Alternatively, a stable kernel can be acquired by scanning a portion of the calibration k-space (e.g., the k-space center) or the entire calibration k-space multiple times during the measurement and performing a calibration on all averaged data. The acquisition mode schematically illustrated in the figure description is based on a navigator acquisition consisting of three lines. This has the advantage that the modification does not result in a change in the repetition time TR and the echo time TE. However, a minimum number of repetitions is required to acquire the entire data acquisition. Of course, it is also conceivable to increase the number of navigator lines by slightly increasing the echo time TE and, potentially, the repetition time TR, in order to acquire more reference lines per repetition.

[0136] exist Figure 16 1 shows a roughly schematic diagram of a magnetic resonance system 160 according to the present invention (hereinafter referred to as "MR system"). The magnetic resonance system comprises, on the one hand, an actual magnetic resonance scanner 102 having an examination room 103 or a patient corridor, into which a patient or examination subject O (e.g., in whose body a specific organ to be imaged is located) can be moved on a bed 108.

[0137] The magnetic resonance scanner 102 is generally equipped with a base field magnet system 104, a gradient system 106, an HF transmitting antenna system 105, and an HF receiving antenna system 107. In the embodiment shown, the HF transmitting antenna system 105 is a whole-body coil permanently installed in the magnetic resonance scanner 102, while the HF receiving antenna system 107 consists of local coils to be placed on the patient or subject (in the embodiment shown). Figure 16 In principle, however, it is also possible to use the entire coil of the HF transmit antenna system 105 as an HF receive antenna system and the local coils 107 as an HF transmit antenna system, provided that these coils can each be switched into different operating modes.

[0138] The MR facility 160 also has a central control device 113, which is used to control the MR facility 160. The central control device 113 includes a sequence control unit 114 for pulse sequence control. With the help of the sequence control unit, the time sequence of radio frequency pulses (HF pulses) and gradient pulses is controlled in relation to the imaging sequence PS selected according to the pulse sequence pattern PSS. Such an imaging sequence PS or the pulse sequence pattern PSS on which the imaging sequence PS is based can be preset in the measurement or control protocol P, for example. Usually, different control protocols P for different measurements are stored in the memory 119 and can be selected by the operator (and optionally changed if necessary) and then used to perform the measurement. The sequence control unit 114 also includes the control sequence determination device 91 according to the present invention. The control sequence determination device 91 generates control data SD, which are implemented in Figure 5 、 Figure 7 、 Figure 9 、 Figure 11 、 Figure 13 or Figure 15 The sequence control unit 114 outputs the excitation modulus and the reading modulus in the order shown in FIG. 1 to perform pulse sequence control.

[0139] To output the individual RF pulses, the central control unit 113 has a radio frequency transmitter 115 that generates and amplifies the radio frequency pulses and feeds them into the RF transmit antenna system 105 via a suitable interface (not shown in detail). To control the gradient coils of the gradient system 106, the control unit 113 has a gradient system interface 116. A sequence control unit 114 communicates with the radio frequency transmitter 115 and the gradient system interface 116 in a suitable manner, for example, by transmitting sequence control data SD, to transmit the pulse sequence PS in the order generated by the control sequence determination device 91. The control unit 113 also has a radio frequency receiver 117 (also communicating in a suitable manner with the sequence control unit 114) to coordinately acquire the magnetic resonance signals received by the RF transmit antenna system 107. A reconstruction unit 118 receives the acquired data as raw data or k-space data RD after demodulation and digitization and reconstructs MR image data from it. This image data BD can then be stored, for example, in a memory 119.

[0140] The central control unit 113 can be operated via a terminal with the input unit 111 and the display unit 109, so that the entire MR system 160 can also be operated by an operator via the terminal. MR images can also be displayed on the display unit 109, and measurements can be planned and started by means of the input unit 111, optionally in combination with the display unit 109, and, in particular, suitable control protocols with measurement sequences can be selected and optionally modified, as explained above.

[0141] The MR system 160 according to the invention and in particular the control device 113 can also have a plurality of further components, not shown in detail here, but usually present at these devices, such as a network interface, in order to connect the entire system to a network and to be able to replace raw data RD and / or image data BD or parameter maps, but also other data, such as patient-related data or control protocols.

[0142] How suitable raw data are obtained by irradiating RF pulses and generating gradient fields and how MR images BD can be reconstructed therefrom is known in principle to a person skilled in the art and will not be explained in detail here.

[0143] It is clear from the above description that the present invention effectively offers the possibility of improving the method for operating a magnetic resonance imaging system for generating magnetic resonance image data with respect to the required duration.

[0144] It should also be pointed out at this point that the features of all exemplary embodiments or developments disclosed in the figures can be used in any desired combination.

[0145] Finally, it should be pointed out again that the detailed methods and configurations described above are embodiments, and those skilled in the art will also be able to change the basic principles in a wide range of fields without departing from the scope of the present invention. For the sake of completeness, it should also be noted that the use of the indefinite article "a" or "an" does not exclude that the feature in question can exist multiple times. Similarly, the term "unit" does not exclude that it is composed of multiple parts, which can optionally also be distributed in space. Regardless of the grammatical part of speech of a particular term, people with male, female or other gender identities are included.

Claims

1. A method for operating a magnetic resonance imaging system to generate magnetic resonance image data (BD) of an examination subject (O), wherein k-space data (RD) are detected during accelerated echo-planar imaging (A-EPI) by undersampling according to a pulse sequence pattern (PSS), the pulse sequence pattern (PSS) having a plurality of repetitions, the repetitions each comprising - a first scanning mode (AS1) for recording (NA)k-space data (RD) for Nyquist ghost correction or for generating magnetic field maps, - a subsequent second scanning mode (AS2) for accelerated echo-planar imaging (A-EPI), and - a common excitation pattern (A-RF) for both recordings (NA, A-EPI), wherein the first scanning pattern (AS1) of the proper subset of the multiple repetitions (REP1, REP2, REP3, REP4) is modified such that: based on the k-space data (RD) detected using the modified first scanning pattern (AS1), missing k-space data (RD) due to the underscanning are supplemented and / or artifacts occurring in image space due to the underscanning are corrected, No separate recording process is performed to detect the reference data.

2. The method according to claim 1, wherein when modifying the first scanning pattern (AS1), the duration of the repetitions (REP1, REP2, REP3, REP4) associated with the modification remains unchanged.

3. A method according to claim 1 or 2, wherein the modified first scanning mode (AS1) and the number of repetitions (REP1, REP2, REP3, REP4) comprised by the proper subset are selected so that a complete scan of the k-space sub-region is performed, which is required at least for the complete supplementation of the missing k-space data (RD).

4. The method according to claim 1 or 2, wherein the modified first scanning pattern (AS1) comprises a first scanning pattern with a different phase encoding (k y ) in the reading direction (k x ) scans the k-space lines.

5. The method according to claim 1 or 2, wherein the modified first scanning pattern (AS1) comprises a plurality of layers which are encoded by at least partially different layers (k z ) in the reading direction (k x ) scans the k-space lines.

6. The method according to claim 1 or 2, wherein the same type of scanning pattern (AS) is used for the modified first scanning pattern (AS1) and the second scanning pattern (AS2).

7. The method according to claim 1 or 2, wherein the updating of the k-space data (RD) detected by the modified first scanning mode (AS1) is performed at different time points of the accelerated echo planar imaging (A-EPI).

8. The method according to claim 1 or 2, wherein the modified first scanning pattern (AS1) comprises: - at least four consecutive phase encoding gradients (Gy) and At least three consecutive readout gradients (Gx) with alternating polarity, which alternate temporally with the phase encoding gradient (Gy). 9 . The method according to claim 8 , wherein the slice selection gradient (Gz) is switched in time synchronously with the second to fourth phase encoding gradients (Gy).

10. The method according to claim 8, wherein the amplitude of the phase encoding gradient (Gy) of the modified first scanning pattern (AS1) is selected as a function of the underscanning factor of the second scanning pattern (AS2).

11. The method according to claim 1 or 2, wherein one of the following specific imaging techniques is used for the accelerated echo planar imaging: -GRAPPA, -SMS / Slice-GRAPPA, -SENSE, -Combination of GRAPPA and SMS, - Using bipolar imaging technology.

12. A control device (113) for a magnetic resonance imaging system, The control device is designed to control the magnetic resonance imaging system using the method according to any one of claims 1 to 11.

13. A magnetic resonance imaging system comprising a control device (113) according to claim 12.

14. A computer program product comprising a computer program that can be directly loaded into a storage unit in a memory of a control device (113) of a magnetic resonance imaging system, the computer program comprising program code sections for executing all steps of the method according to any one of claims 1 to 11 when the program is executed in the control device (113). 15 . A computer-readable medium on which a program section executable by a computing unit is stored, so that when the program section is executed by the computing unit, the steps of the method according to claim 1 are performed.

Images