MR IMAGING USING AN ACCELERATED MULTI-SLICE STEAM SEQUENCE

Abstract

The present disclosure is directed to controlling a magnetic resonance imaging system for generating magnetic resonance image data from an object under examination, in which magnetic resonance raw data is captured, and at least one multi-slice STEAM pulse sequence is generated. The multi-slice STEAM pulse sequence comprises one excitation module for each slice, in each of which are generated a first slice-selective RF excitation pulse and a second slice-selective RF pulse, and one readout module for each slice for acquiring magnetic resonance raw data, which readout module comprises a third slice-selective RF pulse and further sequence elements for spatial encoding and for receiving RF signals. Between the excitation module and the readout module of a first slice is implemented at least one excitation module or one readout module for another slice.

Claims

1. A method for controlling a magnetic resonance imaging system for generating magnetic resonance image data from an object under examination in which magnetic resonance raw data is captured, the method comprising: generating at least one multi-slice stimulated echo acquisition mode (STEAM) pulse sequence including: an excitation module for each slice of a plurality of slices, each excitation module comprising a first slice-selective RF excitation pulse and a second slice-selective RF pulse; a readout module for each slice of the plurality of slices for acquiring the magnetic resonance raw data, the readout module comprising a third slice-selective RF pulse and sequence elements for spatially encoding and receiving radio frequency (RF) signals, and wherein between the excitation module and the readout module of a respective slice of the plurality of slices, at least one excitation module or readout module for another respective slice is implemented; capturing the magnetic resonance raw data based upon the generated at least one multi-slice STEAM pulse sequence; and generating the magnetic resonance image data based upon the captured magnetic resonance raw data.

2. The method as claimed in claim 1, wherein for each repetition interval identified with the pulse sequence, sub-sequences of a constant sub-sequence duration are implemented successively in time, with each respective one of the sub-sequences comprising one excitation module.

3. The method as claimed in claim 1, wherein for each repetition interval identified with the pulse sequence, sub-sequences of a constant sub-sequence duration are implemented successively in time, with each respective one of the sub-sequences comprising one output module.

4. The method as claimed in claim 1, wherein for each repetition interval identified with the pulse sequence, sub-sequences of a constant sub-sequence duration are implemented successively in time, with each respective one of the sub-sequences comprising an excitation module and a readout module that are each assigned to different respective slices from among the plurality of slices.

5. The method as claimed in claim 1, wherein: the readout module corresponds to a sub-sequence of the same slice as the excitation module of a sub-sequence that was output (A−1) steps previously, and A represents an integer greater than one.

6. The method as claimed in claim 1, wherein: a number N of excitation modules and a number (N−A+1) of readout modules are implemented in an initial repetition interval TR, A−1 readout modules are implemented in a final interval, and a condition $T_{\mathrm{Block}}=\frac{TR}{N}$  is satisfied, with T.sub.Block denoting a duration of a sub-sequence.

7. The method as claimed in claim 1, wherein: a number of N excitation modules and a number (N−A+1) of readout modules are implemented in each one of a number of repetition intervals TR identified with the pulse sequence for each one of a number N of the plurality of slices, and a condition $T_{\mathrm{Block}}=\frac{TR}{N+A-1}$  is satisfied, with T.sub.Block denoting a duration of a sub-sequence.

8. The method as claimed in claim 6, further comprising: acquiring the N different ones of the plurality of slices a number of M times, wherein a number of successively-implemented sub-sequences has a value of M×N+A−1, and wherein M defines a number of readout modules per slice.

9. The method as claimed in claim 7, wherein the number of successively implemented sub-sequences has a value of M×(N+A−1), and wherein M defines a number of readout modules per slice.

10. The method as claimed in claim 1, wherein: a number of N slices of the plurality of slices are excited repeatedly in succession, a first excitation module of at least a second slice of the number of N slices is implemented earlier than a second excitation module for a first slice of the number of N slices, and a first readout module for the second slice is implemented later than a second excitation module for the first slice.

11. The method as claimed in claim 1, wherein the excitation module and the readout module of at least one sub-sequence comprise diffusion gradients that differ in terms of direction or amplitude.

12. The method as claimed in claim 1, wherein the first and the second slice-selective RF pulses each influence a plurality of different slices simultaneously.

13. A control-sequence determination device configured to determine a control sequence implemented on a magnetic resonance imaging system, the control-sequence determination device comprising: excitation-module generating circuitry configured to generate an excitation module for each slice of a plurality of slices, each excitation module comprising a first slice-selective RF excitation pulse and a second slice-selective RF pulse, readout-module generating circuitry configured to generate a readout module for each slice of the plurality of slices for acquiring magnetic resonance raw data, the readout module comprising a third slice-selective RF pulse and sequence elements for spatially encoding and receiving radio frequency (RF) signals, arranging circuitry configured to control the excitation-module generating circuitry and the readout-module generating circuitry such that, between the excitation module and the readout module of a respective slice from among the plurality of slices, at least one excitation module or readout module for another respective slice is implemented; and processing circuitry configured to capture the magnetic resonance raw data based upon the generated control sequence, and to generate magnetic resonance image data based upon the captured magnetic resonance raw data.

14. A non-transitory computer-readable medium having instructions stored thereon that, when executed by a processing circuitry of a magnetic resonance imaging system, cause the magnetic resonance imaging system to generate magnetic resonance image data from an object under examination in which magnetic resonance raw data is captured by: generating at least one multi-slice stimulated echo acquisition mode (STEAM) pulse sequence including: an excitation module for each slice of a plurality of slices, each excitation module comprising a first slice-selective RF excitation pulse and a second slice-selective RF pulse; a readout module for each slice of the plurality of slices for acquiring the magnetic resonance raw data, the readout module comprising a third slice-selective RF pulse and sequence elements for spatially encoding and receiving radio frequency (RF) signals, and wherein between the excitation module and the readout module of a respective slice from among the plurality of slices, at least one excitation module or readout module for another respective slice is implemented; capturing the magnetic resonance raw data based upon the generated at least one multi-slice STEAM pulse sequence; and generating the magnetic resonance image data based upon the captured magnetic resonance raw data.

Description

BRIEF DESCRIPTION OF THE DRAWINGS/FIGURES

[0060] The disclosure is described again below in greater detail using exemplary embodiments and with reference to the accompanying Figures, in which:

[0061] FIG. 1 is a schematic diagram of a conventional STEAM sequence for single-slice excitation;

[0062] FIG. 2 is a schematic diagram of a conventional STEAM sequence having spin-echo acquisition;

[0063] FIG. 3 shows a conventional multi-slice STEAM sequence having sequential output of sub-sequences;

[0064] FIG. 4 shows a multi-slice STEAM sequence according to an exemplary embodiment of the disclosure;

[0065] FIG. 5 shows a multi-slice STEAM sequence according to an alternative exemplary embodiment of the disclosure;

[0066] FIG. 6 shows a STEAM sequence having a GRE readout module according to an exemplary embodiment of the disclosure;

[0067] FIG. 7 is a comparative presentation of breast images that were acquired using a diffusion-weighted spin-echo EPI sequence and respectively a diffusion-weighted STEAM EPI sequence according to an exemplary embodiment of the disclosure;

[0068] FIG. 8 is a flow diagram 800 illustrating a method for controlling a magnetic resonance imaging system for producing magnetic resonance image data from an object under examination, in which magnetic resonance raw data is captured, according to an exemplary embodiment of the disclosure;

[0069] FIG. 9 is a schematic diagram of a control-sequence determination device according to an exemplary embodiment of the disclosure;

[0070] FIG. 10 is a schematic diagram of a magnetic resonance imaging system according to an exemplary embodiment of the disclosure;

[0071] FIG. 11 is a schematic diagram of sub-sequences having alternating excitation modules and readout modules for different slices.

DETAILED DESCRIPTION

[0072] FIG. 1 shows schematically a conventional STEAM sequence 20 having single-slice readout. The STEAM sequence 20 consists of three 90° RF pulses 1, 2, 3 and two crusher gradients 9, 10 having the same gradient moment, of which the first crusher gradient 9 is switched between the first RF pulse 1 and the second RF pulse 2, and the second crusher gradient 10 is switched after the third RF pulse 3. The stimulated echo 4 is formed a time interval T after the third RF pulse 3. There is the time that elapses between the first RF pulse 1 and second RF pulse 2. The stimulated echo 4 is generated via the following signal path:

[0073] A first RF excitation pulse 1 in FIG. 1 produces a transverse magnetization. In the time interval T between the first RF pulse 1 and the second RF pulse 2, the spins in the excited region of interest acquire a position-dependent phase PH as a consequence of position-dependent imperfections in the Bo field and in the first crusher gradients 9. The second RF pulse 2 negates the spatially dependent phase PH and flips the magnetization of the spins back into the longitudinal axis. In the time interval of duration TM between the second RF pulse 2 and the third RF pulse 3, gradients and Bo imperfections have no effect on the phase PH of the longitudinal magnetization. The phase dispersion PH-D acquired between the first RF pulse 1 and the second RF pulse 2 is so-to-speak “remembered” in the longitudinal axis. This longitudinal magnetization is excited again by the third RF pulse 3, and re-phased by the second crusher gradient 10 and as a consequence of the position-dependent B0 imperfections, with the result that what is known as a stimulated echo 4 is formed at the time T after the third RF pulse 3. The longitudinal magnetization that is excited by the third RF pulse 3 for the first time, i.e. was not located in the transverse plane in the time interval between the first and second RF pulses, is de-phased by the second crusher gradient 10 and therefore contributes no signal to the acquired signal 4. In addition, spoiler gradients 13 for de-phasing additional signal paths can be switched between the second RF pulse 2 and the third RF pulse 3.

[0074] During the time interval TM, the magnetization is subject only to the longitudinal relaxation process associated with the T1 relaxation, and not to the faster transverse relaxation process, also known as T2 relaxation. This makes the STEAM sequence interesting in particular for diffusion-weighted imaging, in which the two crusher gradients are replaced by diffusion gradients. By extending the mixing time TM, it is possible to increase a maximum diffusion weighting of the STEAM sequence without extending the echo time TE. In contrast, for a diffusion sequence based on the spin-echo technique, increasing the maximum diffusion weighting is always associated with an increase in the echo time. Therefore, STEAM is particularly interesting for diffusion imaging.

[0075] Since the magnetization is de-phased during the time interval TM, this time interval makes a diffusion contribution without extending the echo time TE relevant to the T2 relaxation. This situation differentiates the diffusion-weighted STEAM sequence from the classical spin-echo based diffusion sequences such as the Stejskal-Tanner sequence, in which an increase in the maximum diffusion weighting, the b-value, or in the diffusion time A, the time between the diffusion gradients, is always associated with a longer echo time TE and thus leads to an increase in the T2-induced signal loss.

[0076] The magnetization is subject to T2 relaxation only in the time intervals between the first and second RF pulses, and between the third RF pulse and the stimulated echo. The echo time TE of the STEAM sequence is accordingly TE=2×T.

[0077] Another advantage, in particular at higher field strengths, of a diffusion-weighted STEAM sequence over a spin-echo based diffusion sequence is that it manages without SAR-intensive refocusing pulses.

[0078] All the RF pulses 1, 2, 3 are slice-selective. This is achieved by switching a slice-selection gradient 5, 6, 7 during the application of the RF pulses 1, 2, 3 respectively, and by a suitable choice of the RF center frequency of the RF pulses 1, 2, 3. If the second RF pulse 2 and the third RF pulse 3 have the same bandwidth and duration, one slice-re-phasing gradient 8 after the first RF pulse 1 suffices.

[0079] A fast EPI readout module 11, for instance, can be used for the spatial encoding and readout of the stimulated echo 4 in the STEAM sequence. An EPI readout module comprises gradients for frequency-encoding and phase-encoding of the read-out signal. Details are known to a person skilled in the art and therefore not described further here. Using a fast technique such as EPI to read out an entire image has the advantage of relative insensitivity to movement. EPI does require complete fat suppression, however, to avoid ghosting artifacts in the phase-encoding direction.

[0080] The dominant spectral component of fatty tissue (CH.sub.2 group in the aliphatic chain) has a chemical shift of approximately 3.4 ppm compared with water, and therefore can be suppressed by spectrally selective pre-pulses 12 or a gradient-reversal technique (not shown). The gradient-reversal technique exploits the fact that, as a result of the chemical shift, the fat slice affected by a slice-selective RF pulse is shifted with respect to the water slice in the direction of the gradient. By suitable selection of the bandwidths of the RF pulse and the signs of the slice-selection gradients, it is possible to make this shift so large that fat spins of the dominant spectral component contribute no signal to the stimulated echo.

[0081] A diffusion sensitivity of the read-out signal can be achieved by suitable selection of the crusher gradients 9, 10. The crusher gradients 9, 10 are then also referred to as diffusion gradients.

[0082] The scheme shown produces a diffusion sensitivity (b-value contribution) of

[00003]$b=4{\pi }^2{\gamma }^2{G}^2\left[{\delta }^2\left(\Delta -\delta /3\right)+\frac{{\zeta }^3}{30}-\frac{\delta {\zeta }^2}{6}\right]$

along a gradient direction.

[0083] Here, G denotes the amplitude of a trapezoidal gradient, ζ a ramp time of a gradient, δ the duration of a gradient (ceiling time and ramp time), and Δ is the time difference between the two gradients.

[0084] Lengthening the time interval TM between the second RF pulse and the third RF pulse therefore also lengthens the time difference Δ between the two crusher gradients 9, 10 and hence the diffusion sensitivity of the sequence. Only the time intervals in which the spins that contribute to the stimulated echo signal are lying in the transverse plane make a contribution to the T2 weighting, however. These are the two time intervals of length T between the first RF pulse 1 and the second RF pulse 2, and between the third RF pulse 3 and the stimulated echo 4. The echo time TE of the sequence 20 is hence TE=2×T.

[0085] During the time interval TM, also called the mixing time, the magnetization is subject only to the slower longitudinal relaxation process (T1 relaxation).

[0086] The spoiler gradients 13 switched between the second RF pulse 2 and the third RF pulse are used for dephasing additional signal paths.

[0087] The STEAM sequence 20 shown in FIG. 1 allows a single excitation of one slice and readout of an image. A clinical dataset comprises about 30 to 50 slices. In addition, each slice is usually excited and read out repeatedly. This is necessary, for instance, to capture images with a different diffusion weighting (for instance to calculate an ADC value) and/or a different diffusion direction (for instance in order to compute an isotropic or trace-weighted diffusion image), and/or to increase the signal-to-noise ratio of the computed images by means of averaging.

[0088] Furthermore, data that is needed for instance for coil calibration in parallel imaging or for correcting image distortion is read out after a separate excitation. At the start of a measurement, the slices are also excited without any associated data-readout in order for the magnetization to approach dynamic equilibrium.

[0089] TR denotes below the time that elapses between the successive excitation of a particular slice. N denotes the number of slices that are excited during a TR interval. M denotes the number of excitations of a slice that are required to acquire all the data needed for image computation or image reconstruction. Thus the sequence 20 shown in FIG. 1 is implemented N×M times.

[0090] For describing the method according to the disclosure, it is helpful to divide the STEAM sequence of FIG. 1 into an excitation module EXC and a readout module ACQ. The excitation module EXC comprises e.g. the first RF pulse 1 and the second RF pulse 2, and the diffusion or crusher gradients 9 between the first RF pulse 1 and the second RF pulse 2. The excitation module EXC also comprises further RF pulses and gradients 12, which are applied or switched before the excitation pulse 1, for instance for fat suppression or for regional saturation of the magnetization.

[0091] The readout module ACQ comprises e.g. the third RF pulse 3 of the STEAM sequence 20, and the diffusion or crusher gradients 10, which are switched between the third RF pulse 3 and the “actual” readout module 11, for example an EPI readout module.

[0092] FIG. 2 describes a STEAM sequence 30 having spin-echo acquisition. The STEAM sequence shown in FIG. 2 has three RF pulses 1, 2, 3, which generate up to five echoes. In addition, unlike in FIG. 1, in the STEAM sequence shown in FIG. 2, additional gradients 15 are switched in a time interval after the second RF pulse. The transverse magnetization generated by the first RF pulse 1 is partially refocused by the second RF pulse 2, with the result that a spin echo 4 is formed at the time T after the second RF pulse, for example the first spin echo 14 in FIG. 2, provided the gradients 15, which are switched in the time interval T after the second RF pulse, each have the same moment as the gradients 9, which are switched between the first RF pulse 1 and the second RF pulse 2. This echo 14 can be read out using a second EPI module 16, for instance. Stimulated echo 4 and spin echo 14 have the same T2 weighting but different T1 weighting and diffusion weighting.

[0093] The method according to the disclosure is compatible with reading out the spin-echo. When applying the method according to the disclosure to a spin echo sequence, the gradients 15 and the additional EPI module 16 become part of the excitation module.

[0094] FIG. 3 shows the structure of the entirety of a conventional multi-slice STEAM sequence 31. In the example, the number of slices is N=5.

[0095] A sub-sequence S-SQ comprises an excitation module EXC.sub.n, a first fill time TM.sub.fill or a wait interval of the stated wait time, a readout module ACQ.sub.n and optionally an additional fill time TR.sub.fill or fill interval of the stated fill time. The excitation module EX.sub.n and the readout module ACQ.sub.n of a sub-sequence n belong to the same slice. The fill time TM.sub.fill between the excitation module EXC.sub.n and the readout module ACQn is chosen such that the time between the second RF pulse 2 and the third RF pulse 3 equals the desired TM time TM. Sub-sequences for different slices SL.sub.n are implemented sequentially. The excitation modules EXC.sub.n and readout modules ACQ.sub.n for all the slices SL.sub.1, . . . SL.sub.5 are shown in the bottom row of FIG. 3. Significant “idle times” spent waiting for the TM time to elapse or pass are the result.

[0096] T.sub.Exc denotes the duration of the excitation module EXC, T.sub.Acq denotes the duration of a readout module ACQ, T.sub.Block denotes the duration of a sub-sequence, and TA denotes the duration of an entire sequence. These notations are used to give the following equations:

[00004]$\begin{array}{cc}{T}_{\mathrm{Block}}=T_{Exc}+T{M}_{\mathrm{fill}}+T_{Acq}+T{R}_{\mathrm{fill}},& \left(1\right)\end{array}$$\begin{array}{cc}\mathrm{TR}=N\times T_{\mathrm{Block}},& \left(2\right)\end{array}$$\begin{array}{cc}\mathrm{TA}=M\times \mathrm{TR},& \left(3\right)\end{array}$$\begin{array}{cc}\stackrel{2+3}{.Math.}\mathrm{TA}=M\times N\times T_{\mathrm{Block}}.& \left(4\right)\end{array}$

[0097] FIG. 4 shows schematically the structure of a STEAM sequence 40 according to an exemplary embodiment of the disclosure. The multi-slice sampling is restricted to N=5 slices solely for the sake of a simpler diagram. The time interval between the excitation module EXC.sub.n and the readout module ACQ.sub.n of a particular slice n is used to switch excitation modules EXC.sub.k or readout modules ACQ.sub.k of other slices. Once again, it is possible to divide the sequence 40 into sub-sequences S-SQ, each of which comprises an excitation module EXC.sub.n, a first fill time .sub.fill, a readout module ACQ.sub.k and a second fill time .sub.fill, where k does not equal n and defines a position or number k, n of a slice SL.sub.k, SL.sub.n to be sampled. Therefore, the excitation module EXC.sub.n and the readout module ACQ.sub.k of a sub-sequence S-SQ belong to different slices SL.sub.n, SL.sub.k, and the first fill time .sub.fill is significantly shorter than the first fill time TM.sub.fill of the sequence shown in FIG. 3 for the same mixing time TM.

[0098] In FIG. 4, with the exception of the first and last TR intervals, between the excitation module EXC.sub.n and the readout module ACQ.sub.n of a particular slice SL.sub.n, i.e. of one and the same slice SL.sub.n, are output two excitation modules EXC.sub.k, EXC.sub.1 and are switched two readout modules ACQ.sub.m, ACQ.sub.j for other slices. This is equivalent to two sub-sequences S-SQ. Thus, in the time interval T.sub.Bloc, in which in the conventional single sub-sequence S-SQ is implemented, are implemented sequence modules EXC, ACQ equivalent to the 3 sub-sequences S-SQ. Since the total number of sub-sequences S-SQ used to acquire the data needed for image reconstruction does not change, an acceleration by a factor of approximately A=3 is achieved.

[0099] The actual acceleration of the acquisition is slightly less because in the two first sub-sequences S-SQ of the first TR interval TR, generally in the A−1 first sub-sequences S-SQ of the first TR interval TR, only excitation modules EXC are output, as can be seen in FIG. 4, and accordingly two further sub-sequences, in which just readout modules ACQ are switched are added to the end of the sequence 40.

[0100] The series of the slice indices of the readout modules follows the series of the slice indices of the excitation modules having the slice indices 1, 3, 5, 2, 4, 1, 3, 5, 2, 4, . . . but set back by A−1, so in the example where A=3, by 2 elements. This means that the excitation module EXC2 of the slice SL2 is followed by the readout module ACQ3 of the slice SL3; the excitation module EXC4 of slice SL4 is followed by the readout module ACQS of slice SLS, etc.

[0101] Complex relationships result between the variables N, TR, TM, etc. The maximum acceleration A is largely dependent on TM, the duration T.sub.EXC of an excitation module EXC, and the duration T.sub.ACQ of a readout module ACQ. Roughly speaking, this determines how many excitation modules and readout modules can be packed into the fill time TM.sub.fill.

[0102] For example, an operator specifies a required number N of slices, the b-value, the mixing time TM, and resolution parameters for diffusion imaging. These values then result in the minimum duration of an excitation module and of a readout module and the minimum echo time. Then, all possible acceleration factors A that can be achieved by adjusting the repetition time TR are calculated and offered to the user. If the user selects an acceleration factor A, the repetition time TR is adjusted accordingly. Normally, the user will select the maximum possible acceleration factor A. The value of the maximum acceleration factor A generally increases with the mixing time TM. For a fixed mixing time TM, the value of the repetition time TR decreases with the selected acceleration factor A. This is unusual in the sense that the repetition time TR is normally specified by the user.

[0103] In general, the following holds:

[0104] If {tilde over (T)}.sub.Block defines the duration of a sub-sequence S-SQ of the sequence 40 according to the disclosure, and if A−1 sub-sequences S-SQ of duration {tilde over (T)}.sub.Block fit into the fill time TM.sub.fill of the conventional sequence, then:

{tilde over (T)}.sub.Block=T.sub.Exc+.sub.fill+T.sub.Acq+.sub.fill   (5),

.sub.fill=TM.sub.fill−(A−1)×{tilde over (T)}.sub.Block   (6),

=N×{tilde over (T)}.sub.Block   (7),

=M×+(A−1)×{tilde over (T)}.sub.Block   (8),

[0105] Substituting equation 6 in equation 5 yields:

{tilde over (T)}.sub.Block=T.sub.Exc+TM.sub.fill−(A−1)×{tilde over (T)}.sub.Block+T.sub.Acq+.sub.fill

[0106] Solving for {tilde over (T)}.sub.Block:

A×{tilde over (T)}.sub.Block=T.sub.Exc+TM.sub.fill+T.sub.Acq+.sub.fill

[0107] Using equation 1 gives:

A×{tilde over (T)}.sub.Block=T.sub.Block−TR.sub.fill+.sub.fill   (9)

[0108] In addition, the acquisition time of the sequence according to the disclosure is obtained from equations 7 and 8:

=M×N×{tilde over (T)}.sub.Block+(A−1)×{tilde over (T)}.sub.Block

[0109] The minimum acquisition time is obtained when

[0110] TR.sub.fill=.sub.fill=0 to give:

[00005]$\begin{array}{cc}{\min }_{}=M\times N\times \frac{T_{Block}}{A}+\left(A-1\right)\times {\tilde{T}}_{B1ock}& \left(10\right)\end{array}$$\stackrel{2+3}{.Math.}{\min }_{}=\frac{T{A}_{\min }}{A}+\left(A-1\right)\times {\tilde{T}}_{\mathrm{Block}}$

[0111] Equation 10 reflects the result already obtained from the illustration: apart from the A−1 sub-sequences S-SQ added to the end of a sequence, the minimum acquisition time .sub.min of the sequence 40 according to the disclosure is shortened by a factor A compared with the minimum acquisition time TA.sub.min of the sequence 30 from the prior art. For typical parameters, the second term (A−1)×{tilde over (T)}.sub.Block in equation 10 is negligibly short compared with the first term

[00006]$\frac{T{A}_{\min }}{A}.$

Thus, the number A plays the role of an acceleration factor. The minimum repetition time .sub.min of the sequence according to the disclosure is correspondingly shorter by a factor A than the minimum repetition time TR from the prior art.

[0112] Taking equations 7, 9 and 2 gives initially:

[00007]$=N\times {\tilde{T}}_{\mathrm{Block}}\stackrel{9}{.Math.}=\times \frac{T_{Block}-T{R}_{\mathrm{fill}}+{\mathrm{fill}}_{}}{A}\stackrel{2}{.Math.}\frac{TR}{A}+\frac{{\mathrm{fill}}_{}-T{R}_{\mathrm{fill}}}{A}$

[0113] This yields, on setting TR.sub.fill=.sub.fill=0, the minimum repetition time:

[00008]$\begin{array}{cc}{\min }_{}=\frac{T{R}_{\min }}{A}.& \left(11\right)\end{array}$

[0114] FIG. 5 shows a schematic pulse sequence diagram containing a pulse sequence 50, which illustrates a second embodiment of the method according to the disclosure. Unlike the pulse sequence 40 shown in FIG. 4, for the pulse sequence shown in FIG. 5 the excitation and readout modules of all N slices are implemented in every TR interval TR. In the first A−1 sub-sequences S-SQ, in each TR interval TR is implemented only the excitation module EXC, and the readout module ACQ is replaced by a fill time (idle time) of duration T.sub.Acq. Correspondingly, in the last A−1 sub-sequences S-SQ in each TR interval TR, the excitation module EXC is replaced by a fill time of duration TE.sub.Exc. In this case, A−2 is again the number of sub-sequences S-SQ that are switched between the sub-sequence S-SQ containing the excitation module EXC.sub.n of a particular slice SL.sub.n and the sub-sequence S-SQ containing the readout module ACQ.sub.n of this slice SL.sub.n. The scheme shown in FIG. 5 is repeated in every TR interval. Equations 5 and 6 still hold. The repetition time TR and the acquisition time TA in this embodiment, however, are given by:

TR=(N+(A−1))×{tilde over (T)}.sub.Block   (12)

TA=M×TR  (13).

[0115] Thus, for a large number M of TR intervals, the embodiment shown in FIG. 5 is significantly less efficient than the embodiment of FIG. 4 in terms of time required. This embodiment can have advantages nonetheless. For example, it is possible to wait for a physiological trigger before each TR interval. Physiological triggering refers here to a method in magnetic resonance imaging that reduces motion artifacts caused by a quasi-periodic physiological movement by synchronizing the MR measurement with the physiological movement of the patient. In this process, the data acquisition takes place in portions known as acquisition phases, which are usually shorter than the period length of the physiological movement and each start in the same phase of the quasi-periodic physiological movement. This requires the use of a sensor to capture the physiological movement. A trigger algorithm analyzes the captured signal and generates a trigger signal as soon as the desired phase is detected within the physiological signal. A break point is inserted in the imaging sequence. When the sequence reaches the break point, the implementation of the sequence is interrupted until the trigger algorithm generates a trigger. Then, the sequence is continued until the next break point. In the embodiment shown in FIG. 5, break points can be inserted before each TR interval. After each trigger signal, each of the N slices is than excited and read out once. The physical TR time, i.e. the time between successive excitations of a particular slice, is then determined no longer by the time TR but by the period length of the physiological movement. Since the period length varies, the term “effective repetition time” is used, which equals the average length of a period of the physiological movement when triggering is in each period, or equals a multiple thereof if, for instance, a trigger signal is generated only in every n-th period. TR is the measurement duration per cycle, i.e. the duration of an acquisition phase. Examples of quasi-periodic physiological movements are respiration and cardiac movement. In the case of respiratory triggering, it is attempted, for instance, to generate a trigger signal in such a way, i.e. at a time instant, for which a measurement is made only during the relatively still phase at the end of expiration.

[0116] FIG. 6 shows a gradient echo STEAM sequence 60, referred to as a GRE sequence for short. In the gradient echo variant, the third RF pulse 3 is replaced by a series of RF pulses of smaller flip angle. As a result of the small flip angle, each of these RF pulses flips only some of the magnetization stored in the longitudinal axis into the transverse plane. This generates a plurality of stimulated echoes 4, which have different TM times and are separately phase-encoded. The right-hand readout module ACQ outlined by dashed lines is therefore repeated 1 times in FIG. 6, where 1 is the number of phase-encoding steps that are meant to be read out for each excitation. The totality of the 1 repetitions forms the readout module ACQ in the context of the disclosure. In the variant shown, a readout pre-phasing gradient 18 is switched in the excitation module EXC, the moment of which is half the size of the moment of a readout gradient 17. A diffusion-weighted gradient echo STEAM sequence is described, for example, in the article “Diffusion imaging Using Stimulated Echoes” by Klaus-Dietmar Merboldt, Wolfgang Hänicke, and Jens Frahm (MAGNETIC RESONANCE IN MEDICINE 19, 233-239 (1991)).

[0117] FIG. 7 compares breast images acquired using a diffusion-weighted spin echo EPI sequence, which are shown in the top row, with corresponding images acquired using a diffusion-weighted STEAM EPI sequence with a TM time of 500 ms, which are shown in the bottom row. The aliphatic fat components were suppressed in both image series on the basis of the SPAIR technique by means of a frequency-selective inversion pulse. In the top image row, remaining olefinic fat HA makes it harder to interpret the images. The hypointense regions above the chest muscle is one indicator that the bright signal HA in the top image row relates to olefinic fat. In EPI imaging, an olefinic fat signal is shifted with respect to the water signal in the phase-encoding direction because of the slightly different resonant frequencies. Unlike the adjoining tissue, the muscle does not contain any fat. Thus, the shift produces a hypointense region HA above the muscle. In the images acquired using the STEAM technique, the olefinic fat signal is largely suppressed as well. The hypointense region above the muscle is correspondingly less pronounced.

[0118] FIG. 8 shows a flow diagram 800 illustrating a method 800 for controlling a magnetic resonance imaging system for producing magnetic resonance image data BD from an object under examination O, in which magnetic resonance raw data RD is captured, according to an exemplary embodiment of the disclosure. In step 8.I, an excitation module EXC.sub.1 of a multi-slice STEAM pulse sequence 40 is output for a first slice of a slice stack to be imaged in a region of interest. The excitation module EXC.sub.1 comprises a first slice-selective RF excitation pulse 1 and a second slice-selective RF pulse 2 and also suitable gradients, which are used to define a first slice SL.sub.1 for readout.

[0119] In step 8.II, further excitation modules EXC.sub.3, EXC.sub.5 for a third slice SL.sub.3 and a fifth slice SL.sub.5 are output later than the output of the first excitation module EXC.sub.1.

[0120] Then in step 8.III, a first readout module ACQ.sub.1 for the first slice SL.sub.1 is output. Magnetic resonance raw data RD is acquired during the readout process. The first readout module ACQ.sub.1 comprises a third slice-selective RF pulse 3 and further sequence elements, for instance gradients and readout windows for receiving RF signals 4 from the first slice SL.sub.1.

[0121] In step 8.IV, excitation modules EXC.sub.n and readout modules ACQ.sub.n for the second to fifth slice SL.sub.2, SL.sub.3, SL.sub.4, SL.sub.5 of the slice stack to be imaged are then output alternately in the order EXC.sub.2, ACQ.sub.3, EXC.sub.4, ACQ.sub.5. The first repetition interval TR ends after step 4.

[0122] In step 8.V, the second, or a further repetition interval TR is then started. This further repetition interval now begins again with the output of the excitation module EXC.sub.1 for the first slice SL.sub.1. After this are then output alternately a readout module ACQ.sub.2 for the second slice SL.sub.2 and an excitation module EXC.sub.3 for the third slice SL.sub.3, and then a readout module ACQ.sub.4 for the fourth slice SL.sub.4 and an excitation module EXC.sub.5 for the fifth slice SL.sub.5. After excitation of the fifth slice SL.sub.5, the first slice SL.sub.1 is then read out by a corresponding readout module ACQ.sub.1. Then continues, in alternation, excitation and readout of the second to fifth slice in ascending order by suitable excitation modules EXC.sub.2, EXC.sub.4 and readout modules ACQ.sub.3, ACQ.sub.5, as is also illustrated in FIG. 4. This concludes the second repetition interval.

[0123] Step 8.V is repeated (M−1) times until every slice of the slice stack to be imaged has been excited M times, and also been read out in full except for a final readout of the second slice SL.sub.2 and the fourth slice SL.sub.4. This process is represented in FIG. 8 by the steps 8.Va, 8.Vb, in which it is determined whether the counter i for the excitations per slice SL.sub.n has not yet reached the predetermined total number M of excitations per slice SL.sub.n. If this is the case, which is represented in FIG. 8 by “y” (=yes), the flow moves to step 8.Vb in which the counter i is incremented, and then step 8.V is repeated. If in step 8.Va the predetermined total number M of excitations per slice is reached, which correspond to the M-th repetition interval, then the flow moves to step 8.VI, which is represented in FIG. 8 by “n” (=no).

[0124] Then, in step 8.VI is performed readout of the second slice SL.sub.2 and fourth slice SL.sub.4, which had still not been read out in the M-th repetition interval in step 8.V. This ends the acquisition process for the magnetic resonance raw data RD.

[0125] FIG. 9 shows a schematic diagram of a control-sequence determination device 90 according to an exemplary embodiment of the disclosure.

[0126] The control-sequence determination device 90 is designed to determine a control sequence 40, 50 according to the disclosure, which can be implemented on a magnetic resonance imaging system 100 (see FIG. 10). The control-sequence determination device 90 comprises an excitation-module generating unit (e.g. circuitry) 91 for generating an excitation module EXC.sub.n for each slice SL.sub.n of a slice stack, wherein a first slice-selective RF excitation pulse and a second slice-selective RF pulse are generated for each slice. Part of the control-sequence determination device 90 is also a readout-module generating unit (e.g. circuitry) 92, which is configured to generate one readout module ACQ.sub.n for each slice SL.sub.n for acquiring magnetic resonance raw data RD, wherein the readout module ACQ.sub.n comprises a third slice-selective RF pulse and further sequence elements, for instance gradients and readout windows for spatial encoding and for receiving RF signals.

[0127] In addition, the control-sequence determination device 90 comprises an arranging unit (e.g. circuitry) 93, which is configured to control the excitation-module generating unit 91 and the readout-module generating unit 92 in accordance with a pulse sequence scheme PS according to the disclosure, so that between the excitation module EXC.sub.n of one slice SL.sub.n and the associated readout module ACQ.sub.n, further excitation modules EXC.sub.k and/or readout modules ACQ.sub.1 for other slices SL.sub.k, SL.sub.1 are output.

[0128] FIG. 10 shows a sketch of a magnetic resonance system 100 according to the disclosure (referred to below as “MR system” for short). The magnetic resonance system 100 comprises the actual magnetic resonance scanner 102 containing an examination space 103 or patient tunnel, into which can be moved on a couch 108 a patient O, or rather in this case a patient or person under examination, in whose body is located, for instance, a specific organ to be imaged.

[0129] The magnetic resonance scanner 102 is equipped in the usual manner with a main magnetic field system 104, a gradient system 106, and also an RF transmit antenna system 105 and an RF receive antenna system 107. In the exemplary embodiment shown, the RF transmit antenna system 105 is a body coil that is fixed in the magnetic resonance scanner 102, whereas the RF receive antenna system 107 consists of local coils (represented in FIG. 10 by a single local coil) to be arranged on the patient or person under examination. In principle, however, the whole-body coil 105 can also be used as the RF receive antenna system, and the local coils 107 can be used as the RF transmit antenna system, provided these coils can each be switched into different operating modes.

[0130] The MR system 100 also comprises a central control device 113, which is used to control the MR system 100. This central control device 113 comprises a sequence control unit 114 for controlling the pulse sequence. This is used to control the timing of radiofrequency pulses (RF pulses) and gradient pulses according to a selected imaging sequence PS. Such an imaging sequence can be specified, for example, in a measurement or control protocol P. Different control protocols P for different measurements are typically stored in a memory 119, and can be selected by an operator (and possibly modified if required), and then used to perform the measurement. The sequence control unit 114 also comprises a control-sequence determination device 90 according to the disclosure that has the design shown in FIG. 9. The control-sequence determination device generates control data SD, which facilitates the series of excitation modules and readout modules shown in FIG. 9 that are output by the sequence control unit 114 for controlling the pulse sequence.

[0131] For the output of the individual RF pulses, the central control device 113 comprises a radiofrequency transmit device 115, which generates and amplifies the RF pulses and feeds said RF pulses into the RF transmit antenna system 105 via a suitable interface (not shown in detail). The control device 113 comprises a gradient system interface 116 for controlling the gradient coils of the gradient system 106. The sequence control unit 114 communicates in a suitable manner, e.g. by sending out sequence control data SD, with the radiofrequency transmit device 115 and the gradient system interface 116 for the emission of the pulse sequences PS in the order produced by the control-sequence determination device. The control device 113 also comprises a radiofrequency receive device 117 (likewise communicating with the sequence control unit 114 in a suitable manner) for the purpose of coordinated acquisition of magnetic resonance signals received by the RF transmit antenna system 107. A reconstruction unit 118 receives the acquired data after demodulation and digitization as raw data RD and reconstructs the MR image data therefrom. This image data BD can then be stored in a memory 119, for example.

[0132] The central control device 113 can be operated by means of a terminal having an input unit 110 (such as a user interface) and a display unit 109, via which an operator can hence also operate the entire MR system 100. MR images can also be displayed on the display unit 109, and the input unit 110, if applicable in combination with the display unit 109, can be used to plan and start measurements, and in particular to select and, if applicable, modify suitable control protocols containing suitable measurement sequences, as described above.

[0133] Furthermore, the MR system 100 according to the disclosure and e.g. the control device 113 can also comprise a multiplicity of further components, which are not shown here in detail but are typically present in such equipment, for instance components such as a network interface, to connect the entire system to a network and to be able to transfer raw data RD and/or image data BD or parameter maps, but also other data such as patient-related data or control protocols, for example.

[0134] The principles of how suitable raw data RD can be acquired by applying RF pulses and generating gradient fields, and how MR images BD can be reconstructed from said raw data, are known to a person skilled in the art and are not explained further here. Likewise, a variety of measurement sequences such as e.g. EPI sequences, GRE measurement sequences, or TSE measurement sequences (TSE=turbo spin echo) for generating dynamic or static images are known in principle to a person skilled in the art.

[0135] FIG. 11 shows a schematic diagram 120 of sub-sequences having alternating excitation modules EXC and readout modules ACQ for different slices.

[0136] The equation already used in the general section =TM(A−1)×T.sub.Block shall be explained briefly with reference to FIG. 11 for the case A=3.

[0137] In the method according to the disclosure, between the sub-sequence S-SQ1 containing the excitation module EXC1 for the first slice and the sub-sequence S-SQ3 containing the readout module ACQ1 for the first slice are implemented A−2 further sub-sequences, so in this case one further sub-sequence S-SQ2. The duration of a sub-sequence is T.sub.Block. This in turn means that the time difference between the RF pulse of a readout module ACQ2 of one sub-sequence S-SQ1 and the RF pulse of the readout module ACQ4 of the subsequent sub-sequence S-SQ2 is also equal to T.sub.Block, because the timing of the excitation module and the readout module within a sub-sequence is the same for all the sub-sequences.

[0138] The time difference between the second RF pulse of the excitation module EXC1 and the RF pulse of the readout module ACQ2 of a sub-sequence S-SQ1 therefore differs from the physical mixing time TM by precisely A−1 time intervals, so in the example by precisely two time intervals, T.sub.Block.

[0139] Finally, it shall be reiterated that the methods and devices described above are merely preferred exemplary embodiments, and that the disclosure can be modified by a person skilled in the art without departing from the scope of the disclosure insofar and/or as defined by the claims. So, although the method according to the disclosure has been described using the example of a diffusion-weighted single-shot EPI sequence, the disclosure is in no way restricted to diffusion-weighted sequences or single-shot sequences or EPI sequences, and any suitable type of sequences may be used

[0140] It is mentioned for the sake of completeness that the use of the indefinite article “a” or “an” does not rule out the possibility of there also being more than one of the features concerned. Likewise, the term “unit” does not exclude the possibility that said unit consists of a plurality of components, which may also be spatially distributed if applicable.