DE102008051039A1 - Gradient-conditioned magnetic resonance signal shift correction method, involves determining image plane in object, and correcting signal shift before selection of signal for axes, where signal selection takes place along axes - Google Patents

Abstract

The method involves determining an image plane in an investigation object i.e. person (12), in which echo planar imaging is accomplished, relative to axes of coordinates of a gradient system (13) such that the coordinate axes run along a selection direction and a magnetic resonance (MR) signal is readout along the selection direction. A gradient-conditioned MR signal shift is determined for each coordinates axis of the gradient system. The signal shift is corrected before selection of the MR signal for the coordinate axes, where the signal selection takes place along the coordinate axes. An independent claim is also included for a device for correcting a gradient-conditioned magnetic resonance signal shift with magnetic resonance pictures.

Description

The The present invention relates to a method for correcting a gradient-induced MR signal shift in MR images obtained with an echoplanar imaging technique be included and a device for this.

The EPI (Echo Planar Imaging) imaging was based on a fast meandering Scanning of k-space by gradient refocusing. in this connection is after excitation of k or raw data space with a bipolar Go through gradient pulse scheme. Any dynamic or systematic Field errors in the gradient fields can vary depending on the time constant lead to a systematic echo shift in the readout time window. A systematic MR signal shift in the meandering scan of k-space causes an alternating right-left shift the individual echoes with a period 2. A possibly permanently existing Gradient offset of the same axis direction shifts the echo time depending on the sign relative to the readout gradient from the middle of the Read interval either forward or backward, since the echo maximum occurs at the time at which the time integral of the gradient Is zero. This systematic modulation of k-space leads to the reconstruction of the MR images by Fourier transformation too the so-called N / 2 ghost artifacts. Causes of such a signal shift can For example, be dynamic field errors due to eddy currents. Another reason for the MR signal shift can be a temporal change be the gradient fields due to bending vibrations of the gradient coils. Likewise the gradient pulses for creating the gradient fields, for example by signal propagation time differences relative to the RF pulse and the receive time window delayed be.

The systematic shift of the MR signal with the period two becomes in the prior art by a so-called phase correction scan recorded and thus the echo position of the individual echoes during reconstruction corrected after recording the measurement data. This phase correction scan usually takes place before the phase coding and directly after the excitation and consists of at least two gradient-refocused echoes with alternating gradient polarities. Alternative or supplementary can also be used a separately recorded complete EPI Echozug which does not phase encode.

To a one-dimensional Fourier transform of the phase correction data Then, using the linear phase response, the echo shift becomes positive and determines negative gradient amplitudes. An echo shift in the time domain corresponds to a linear phase response in the frequency domain. If more data available such as the full correction scan, can add extra Shifts for example for every single echo can be determined. This shift data will be used to correct the actual recorded image data and thus reduce the N / 2 ghost artifacts. The correction thus takes place after recording the measurement data.

In Magnetic resonance imaging (MRI), it is now possible, the image plane in the Examination object completely free to choose. In transversal, sagittal or coronal stratification, the readout gradient is while the MR signal is read out parallel to one of the three coordinate axes of the gradient system. When looking at rotated image planes the readout gradient is made up of several physical gradient axes super positioned. In general, however, field errors and thus resulting MR signal shifts for different coordinate axes of the gradient system is not identical. As a result, the Time for the zero crossing of the zeroth gradient moment, that of the gradient echo corresponds, is axisindividuell. Diverge these times for the am Reading off gradient pulses involved, this means that a linear phase and an echo position can be determined the one RMS value from both single axes in two-dimensional Case correspond. However, this does not necessarily mean that the both individual moments of the physical gradient axes at the same time be refocused. For example, the time focus centered on both echo components. Are the echo positions of the coordinate axes of the gradient system too far apart, before and after Fourier transformation is not enough signal available, to determine a linear phase. This means that the phase correction scan aslant is encoded by the k-space past the k-space center, where with growing Distance at the k-space center decreases the signal amplitude. This above described problems cause no or only insufficient Compensation of the MR signal shift, resulting in the already above mentioned N / 2 ghost artifacts on rotated image planes or layers leads.

It It is therefore an object of the present invention to provide such artifacts at image planes that are not parallel to one of the coordinate axes of the gradient system are to be avoided.

These The object is achieved with the features of the independent claims. In the dependent are preferred embodiments of the invention.

According to the invention is a Method for correcting a gradient-related MR signal shift provided in MR images obtained with echoplanar imaging be recorded. In a step according to the method, the image plane in the examination subject in which the echoplanar imaging carried out is to be, wherein the image plane relative to the coordinate axes of the Gradient system is determined such that the coordinate axes be determined, along which a readout sen for Ausle sen of the MR signal. In a further step, the MR signal shift for each coordinate axis of the gradient system. After determining the MR signal shift for every Coordinate axis can do this before reading out the MR signal for those Coordinate axes are corrected, along which the signal readout he follows. By determining the gradient axis-specific MR signal delay can the signal shift for each coordinate axis can be corrected axis specific, so that even with rotated image plane the appearance of the ghost artifacts can be minimized.

According to one preferred embodiment the MR signal shift with the help of at least two gradient-refocused Echoes are determined, where for the determination of the MR signal shift along a coordinate axis, the projection of the two gradient-refocused Echoes is determined on this coordinate axis. This means that in the case of the at least two gradient-refocused echoes, the projection is used on the involved physical gradient axes, by for the Determining the signal shift along an axis at least two gradient echoes due to magnetic field gradients along this axis be switched.

following is the generation of the gradient-refocused echoes by a For example, bipolar gradient scheme for determining the signal shift and referred to their correction as a correction measurement. Preferably finds this correction measurement after the signal excitation of the echoplanar imaging and before signal selection. In the two-dimensional case, for example both single axes directly in successive measurements of the same Be measured, the results when reading the EPI signal can be used directly after the signal excitation, according to the correction measurement is performed. Alternatively to this can the results of the correction measurement also for reading the following Application can be used, but contrary to the im State of the art known Phasenkorrek turscan the correction Circuit of the gradient for reading is done to achieve that the gradient echoes take place in the k-space center. When in the state the technique known Phasenkorrekturscan the correction takes place after reading the measured data by post-processing of the detected Signals.

In an embodiment will correct the signal shift before each echo planar imaging carried out, in which the readout direction components in more than one coordinate axis direction Has. Alternatively, the correction measurement may be unique at the beginning the examination of an investigator or in the context of the establishment of the MR system are determined and applied to the subsequent measurements become. Likewise, the corrected signal shift can also be applied to a subsequent correction measurement can be applied to check whether the gradient echoes on the individual coordinate axes really in the middle of the readout window at the same time for the different axes respectively. For a single application it is assumed that the gradient-related MR signal shifts over the Do not change time. When integrating the correction measurement into the imaging sequence after each signal excitation can time varying influences such as by heating components are better compensated by the MR system. For which imaging Measurement the correction measurement is used depends u. a. from the processing speed the involved signal processing devices. Is the processing speed very high high, the correction measurement can already be used for the directly following gradient switching the signal readout of the EPI imaging sequence can be used. As well The correction can be made only on the second after the correction measurement gradient switching of the EPI imaging sequence if this makes sense for signal processing reasons is.

The Aim of the correction of the MR signal shift along the individual Coordinate axes is the adaptation of the axis-specific delay such that the axis-specific MR signals for every Coordinate axis of the gradient system substantially simultaneously occur. The correction can be done by adding a axis - specific gradient moment to the gradient along the individual coordinate axes or for example by switching the control signals for switching the gradients such that the Gradientensignale for switching the gradient so controlled (delayed, for example and premature) be that in turn the MR signals on the individual axes at the same time in the k-space center respectively.

The invention further relates to a device for correcting this gradient-induced MR signal shift with a device for determining the image plane such that the coordinates tenachsen along which the readout direction runs. Furthermore, a device for determining the MR signal shift for each coordinate axis of the gradient system is provided, as well as a correction device which corrects the signal shift before reading the MR signal for those coordinate axes along which the signal readout takes place. The device for correction preferably operates as described in more detail above.

The Invention will be described below with reference to the accompanying Drawings closer explained.

1 shows an MR system in which a correction of a gradient-induced MR signal shift in rotated layers is possible,

2 shows how a gradient offset leads to a temporal signal shift in the MR signal,

3 shows a rotated image plane relative to the coordinate axes of a gradient system,

4 shows a section of an EPI imaging sequence at orthogonal image plane,

5 shows the imaging sequence of 4 at non-orthogonal image plane,

6 shows the imaging sequence of 5 with a correction measurement along the x-axis of the gradient system,

7 shows the sequence of 6 in which the signal shift is measured along the y-axis of the gradient system.

In 1 An MR system is shown, with which a gradient-induced MR signal shift at oblique or rotated image plane is possible in order to avoid artifacts in the subsequent MR image. The MR system has a magnet 10 to generate a polarization field B ₀ . One on a lounger 11 arranged examiner 12 is used to create MR images in the isocenter of the magnet to there MR images of the subject 12 to be able to record. Furthermore, a gradient system 13 shown schematically for the generation of linear magnetic field gradients, which are necessary for the localization of the recorded MR signals and thus for the subsequent generation of an MR image. It is known to those skilled in the art how fundamental MR images can be generated by irradiation of RF pulses and switching of gradient fields, so that this will not be explained in more detail below.

The MR system also has a central control unit 14 on. This central control unit 14 has a gradient control unit 15 which controls the switching of the magnetic field gradients in the three spatial directions. Furthermore, an image calculator 16 provided, which calculates MR images from the detected MR signals. A sequence control unit 17 is provided, which controls the process for the irradiation of RF pulses by an RF system (not shown) and the gradient. The sequence control unit 17 has, for example, a device for determining an image plane 18 which is one of an operator via an input unit 19 fixed image plane determined in the subject. The image plane may be on a display unit, for example 20 determined by the operator. Furthermore, a device for determining the MR signal shift for each coordinate axis of the gradient system 21 intended. The facilities 15 - 21 are symbolically represented as separate units and describe various functions of the MR system. Of course, the various devices or units can be combined in other units or can be integrated in other system components of the MR system.

In connection with 2 It is described how a gradient circuit which is not ideally adjusted leads to a temporal shift of the MR signal or to a phase offset. Will after irradiation of an RF pulse 22 a gradient along a coordinate axis of the gradient system, be it Gx, Gy or Gz is irradiated, and then a bipolar gradient is switched by a negative gradient component 23 and a positive gradient proportion 24 Thus, by reversing the polarity of the gradient, a gradient echo takes place, which ideally takes place at the echo time TE. The echo time is determined by the respective areas under the gradient dashed lines, since the echo maximum is recorded at the time at which the time integral of the gradient is zero. Is now like in 2 shown a positive gradient offset before, the echo is already at an earlier time, which is shifted by .DELTA.t over the ideal time TE forward. With a negative gradient offset, the MR signal is shifted backwards in time.

If, now, the image plane for generating the MR image is orthogonal to the coordinate system of the gradients, the signal readout in the plane lies parallel to one of the coordinate axes of the gradient system. Related to 3 this means that at orthogonal image plane, the readout direction along the axis x, y or z, wherein the in 3 shown coordinate system corresponds to the coordinate axes of the gradient system. If the picture plane 31 as in 3 shown, however is not orthogonal in the coordinate axes of the gradient system, so a simple compensation of the signal shift along an axis is no longer sufficient. The reading direction indicated by the arrow 32 is shown, is not parallel to one of the two axes. In the example shown, the read-out direction has a proportion along the coordinate axis x of the gradient system, with 33 is shown, and a share 34 along the coordinate axis y. In the case shown, the readout direction is orthogonal to the z-axis. However, it is also possible that the image plane 31 is stored such that even a z-components of the gradient readout is present.

In 4 FIG. 2 is a partial representation of a sequence diagram of an orthogonal layer guide echo planar imaging sequence. FIG. As is known, in echo planar imaging, after a single RF excitation 41 generates a series of echoes in the readout gradient (Gx), which fill different lines in the k-space plane by varying the phase-encoding gradient Gy. At the in 4 illustrated embodiment, a change of the Phasenkodiergradienten Gy is not shown for reasons of clarity. After selective signal excitation by the RF pulse 41 with simultaneous switching of the slice selection gradient 42 the repeated bipolar circuit of the read-out gradient, in each case gradient echoes are generated by this circuit, which schematically with 43 are designated. The distance between two such gradient echoes is TEE, while the total acquisition time is TA.

In 5 is now the sequence diagram of 4 shown with rotated layer, wherein the image plane is selected such that the read-out direction components in the x- and y-rich direction of the gradient system has, as in 3 shown. With such a rotated image plane, the readout gradient is superposed on a plurality of physical gradient axes, whereby, as in FIG 5 shown readout gradients 44 must be switched in the x direction and readout gradients 45 in the y direction. In order to minimize at such a position of the image plane, the ghost artifacts that can arise when the readout gradients, as in 2 shown are not ideally centered, so the gradient-related signal shifts must be corrected. In the 4 - 7 have the same elements the same reference numerals, so that these elements will not be further explained again.

In 6 It is now shown how the gradient-axis-specific pulse delay or signal shift is determined and corrected. The image plane in the imaging sequence of 6 corresponds to the image plane of the in 5 illustrated example with readout gradients in the x and y directions. In order to compensate for the MR signal shift along the x-coordinate axis of the gradient system, alternating gradients become respectively 46 Switched along the x-coordinate axis, leading to echoes 47 to lead. This signal projection along the x-axis can be used for the necessary correction of the readout gradient in the x-direction. When the gradient echoes 47 can not be induced in the middle of the read window, but temporally offset to the front or back, so for example by the gradient control unit 15 axis-specific for the x-axis additional corrective gradient moments before the actual readout gradient 44 be switched to reach for the x-axis that the signal echoes 43 each time TE in the middle of the read window (see. 2 ) respectively. In 7 is through the gradient circuit 48 determines the signal shift for the y-coordinate axis, so that subsequently in the readout gradient 45 either an additional gradient moment specific to the y-axis may be switched before the readout gradient pulse or by, for example, introducing a corrective delay in the gradient signals in the generation of the signals in the gradient control unit. Depending on the processing speed, the correction takes place directly before switching the associated readout gradients 44 . 45 , before the correction measurement took place, or the readout gradients are corrected in the next or the next readout train.

It may now be the case that the read-out direction is significantly smaller on one of the two axes than on the other axis, ie the section 34 would be much smaller than the section 33 , Under these conditions, the axial determination of the echo time shift for the smaller amplitude axis can be very difficult since there is only a small gradient amplitude and thus a very wide echo. At the same time, a relative temporal deviation from the axis having the larger amplitude does not manifest itself significantly in the form of artifacts, so that an optional assessment of the asymmetry of the division of the gradient amplitudes can be made and the above-described axes are individualized as a function of the result of this evaluation. For example, the axis-specific correction can only take place if the proportion along a gradient coordinate axis is greater than a predetermined threshold value.

Of course you can the method described above, which is related to two dimensions also be transferred to three axis, so that at a proportion of the read-out direction in the z-direction also the signal shift determined in the z direction and then corrected.

The above-described individual axis correction measurement can now be carried out before the first EPI measurement, the value determined for each further measurement being applied for each axis, wherein a normal correction measurement according to the prior art can be performed for each subsequent EPI measurement. In another embodiment, as in 5 - 7 represented the correction measurement between excitation and the actual correction measurement integrated and applied for all following readout gradients. The additional delay is applied in real time for all subsequent gradients or offset by one measurement.

Claims (10)

Method for correcting a gradient-related MR signal shift in MR images taken with echo planar imaging, with the following steps: - Determine an image plane in an examination subject in which echoplanar imaging is performed should, relative to coordinate axes of a gradient system such, that the coordinate axes are determined, along which a readout direction runs, along which the MR signal is read, - Determine the MR signal shift for each coordinate axis of the gradient system, and - Correct the signal shift before reading the MR signal for those Coordinate axes along which the signal is read. Method according to claim 1, characterized that the MR signal shift with the help of at least two gradient-refocused Echo is determined, where for the determination of the MR signal shift along the coordinate axes axis-individual an at least bipolar gradient scheme is switched accordingly axis-individually to generate the gradient echoes. Method according to claim 1 or 2, characterized in that correcting the MR signal shift after a signal excitation and before reading the signal. Method according to one of the preceding claims, characterized characterized in that correcting the signal shift before every echo planar imaging is performed at which the readout direction Having components in more than one coordinate axis direction. Method according to one of claims 1 to 3, characterized that determines the MR signal shift along the individual coordinate axes and subsequently to correct the MR signal shift of multiple echoplanar imaging. Method according to one of the preceding claims, characterized characterized in that it is determined what proportion of the readout direction in the direction of a coordinate axis, wherein the correction the MR signal shift along this coordinate axis only occurs if the proportion is greater than one is predetermined threshold. Method according to one of the preceding claims, characterized characterized in that the correction of the MR signal shift along the individual coordinate axis by fitting an axis-specific one delay such that the axis-specific MR signals substantially to occur at the same time. Method according to one of the preceding claims, characterized characterized in that the correction of the MR signal shift by Add an axis-specific gradient moment along the gradients the individual coordinate axes of the gradient system takes place in such a way that the axis-specific MR signals are substantially the same Time occur. Method according to one of the preceding claims, characterized characterized in that the correction of the MR signal shift takes place, by providing control signals for switching the gradients for each coordinate axis be controlled so that by the gradient circuit conditional MR signals for the individual coordinate axes at substantially the same time occur. Device for correcting a gradient-related MR signal shift in MR images taken with echo planar imaging, With: - one Device for determining an image plane in an examination subject, in which the echo planar imaging is Runaway leads, relative to coordinate axes a gradient system like this, that the coordinate axes be determined along which a readout direction, along the MR signal is read out, - a device for determining the MR signal shift for every coordinate axis of the gradient system, and A correction device, the signal shift before reading the MR signal for those Corrected coordinate axes along which the signal is read.