Procedure for determining a B1 field map

Abstract

Method for determining a B1 field map in a magnetic resonance tomograph (17) which describes a local field strength of a high-frequency electromagnetic B1 alternating field irradiated by a transmitter (19) to excite nuclear spins for at least one measurement range, wherein, within a measurement sequence, several excitation pulses (1, 8) are irradiated by the transmitter (19), each of which changes the magnetization of an excitation range encompassing the measurement range according to an associated flip angle, and a first measurement value is recorded by a receiver (20) in a first measurement interval (12) and a second measurement value is recorded in a second measurement interval (15), relating to the magnetization in the measurement range, wherein, depending on the first and the second measurement value, a preliminary flip angle value is determined for the flip angle associated with a selected excitation pulse (1) in the measurement range.wherein the associated flip angle of the selected excitation pulse (1) is less than 90°, characterized in that a correction factor dependent on a pulse shape of the selected excitation pulse (1) is determined, wherein a corrected flip angle value is determined by multiplying the preliminary flip angle value with the correction factor, wherein the corrected flip angle value or a value derived therefrom is stored in the B1 field map to describe the local field strength in the measurement area, wherein in the measurement sequence the selected excitation pulse (1) is played back twice before the acquisition of the first and the second measured value, wherein a first gradient field (5) is activated between the plays back to dephase the transverse magnetization, wherein a common excitation pulse (8) is used to acquire the first and the second measured value, wherein in and / or before the first measurement interval (11) in which the first measured value is acquired,or in the second measurement interval (12), in which the second measurement value is acquired, a second gradient field (11) is activated to generate an echo of the previously dephased magnetization.

Description

[0001] The invention relates to a method for determining a B1 field map in a magnetic resonance tomograph, which describes a local field strength of a high-frequency electromagnetic B1 alternating field radiated by a transmitter device to excite nuclear spins for at least one measurement area, wherein, within a measurement sequence, several excitation pulses are radiated by the transmitter device, each of which changes the magnetization of an excitation area encompassing the measurement area according to an associated flip angle, and a first measurement value is acquired by a receiver device in a first measurement interval and a second measurement value in a second measurement interval, relating to the magnetization in the measurement area, wherein, depending on the first and the second measurement value, a preliminary flip angle value is determined for the flip angle associated with a selected excitation pulse in the measurement area.In addition, the invention relates to a magnetic resonance imaging scanner.

[0002] The field strength of the high-frequency field used in magnetic resonance imaging (MRI) scanners to excite the layers of a specimen, the so-called B1 field, is typically not homogeneous across the entire measurement range. Inhomogeneities can be caused by the MRI scanner's geometry or by the specimen itself. By taking this inhomogeneity into account during data analysis, the measurement quality of many imaging techniques, such as T1-weighted imaging, can be significantly improved. Determining B1 inhomogeneity can also be used for calibration and fault diagnosis in MRI scanners, or for pulse calculation in multi-channel operation. Therefore, it is known to measure B1 field maps that describe the local strength of the B1 field in different measurement ranges, particularly in the form of a two-dimensional pixel or three-dimensional voxel field.

[0003] The local B1 field strength can be determined by applying an excitation pulse with a specific, predetermined flip angle and measuring the flip angle achieved locally by this excitation pulse. A problem with this approach is that, even with a completely homogeneous B1 field, the actual flip angle achieved varies across the slice thickness when exciting individual slices. In magnetic resonance imaging (MRI) scanners, slice selection is achieved by choosing a suitable excitation pulse that excites the sample slice-specifically based on its frequency spectrum and a slice selection gradient. However, for homogeneous excitation of a defined slice, the excitation pulse would have to be a Fourier transform of a rectangular function in the frequency domain, i.e., a sinc function. A sinc excitation pulse, however, would have an infinite length.Sine pulses can be used as actual excitation pulses, which are time-limited by multiplication with a so-called window function. Depending on the specific window function chosen, however, this leads to ripple in the amplitude in the frequency domain and / or a roll-off of the amplitude at higher and lower frequencies, i.e., a "rounding" of the rectangular function in the frequency domain. Accordingly, the flip angle achieved with such an excitation pulse varies locally, oscillating in the direction of the slice selection gradient and / or decreasing from the center of an excitation slice towards the edges. However, in a magnetic resonance imaging (MRI) scanner, the magnetization is always measured as an average over an excited slice. A measured flip angle therefore deviates from a flip angle at the center of the excitation area, which is typically the target, with the error depending on the pulse shape used.

[0004] Several approaches to circumvent this problem are known in the prior art. On the one hand, it is possible to use preparation pulses whose flip angle is to be determined that are not layer-selective or that have a large layer width compared to the other measurement pulses used in the sequence. In this case, however, it is not possible to measure several adjacent layers in rapid succession, as these would be influenced by preceding non-layer-selective excitations.

[0005] More homogeneous excitation at the same slice width can be achieved by using preparation pulses with a high bandwidth-time product. However, such pulse shapes result in large amplitudes of the RF pulses. If high flip angles are also required, the bandwidth-time product may be technically limited.

[0006] The described inhomogeneities can have no effect on the measured flip angles if phase-based methods are used to determine them. However, these methods are susceptible to interference from inhomogeneities in the background magnetic field, can have long measurement times and / or limited dynamic ranges, and can lead to high SAR exposures. Therefore, it is preferable to use amplitude-based methods to determine flip angles and thus B1 maps.

[0007] The publication by M. Schär et al.: Simultaneous B0- and B1+-Map Acquisition for Fast Localized Shim, Frequency, and RF Power Determination in the Heart at 3 T. In: Magnetic Resonance in Medicine (2010), Vol. 63, pages 419-426, discloses a method for determining shimming correction values ​​using B0- and B1+ maps. The publication by C. Cunningham et al.: Saturated Double-Angle Method for Rapid B1+ Mapping. In: Magnetic Resonance in Medicine (2006), Vol. 55, pages 1326-1333, discloses a method that combines a double-angle method with a B1-insensitive sequence. The publication US 8,890,527 B1 discloses a method for measuring a B1 field. The publication by J. Wang et al.: Factors Influencing Flip Angle Mapping in MRI: RF Pulse Shape, Slice-Select Gradients, Off-Resonance Excitation, and B0 Inhomogeneities. In: Magnetic Resonance in Medicine (2006), Vol. 56, pages 463-468, discloses a method for calculating flip angle maps.

[0008] The invention is based on the objective of providing a method for determining a B1 field map that enables improved amplitude-based measurement of B1 field maps.

[0009] The object is solved according to the invention by a method of the type mentioned at the outset, wherein a correction factor dependent on a pulse shape of the selected excitation pulse is determined, wherein a corrected flip angle value is determined by multiplying the preliminary flip angle value with the correction factor, wherein the corrected flip angle value or a value derived therefrom is stored in the B1 field map to describe the local field strength in the measuring area.

[0010] According to the invention, it is proposed to consider the known pulse shape of the excitation pulse when determining a corrected flip angle value within the framework of determining a B1 field map. While only the pulse shape of the selected excitation pulse can be considered, preferably the pulse shapes of several excitation pulses, in particular all excitation pulses that influence the first and / or the second measured value, are taken into account. The correction factor can depend on the pulse shape of all considered excitation pulses. Excitation pulses are typically calculated or "synthesized" in discrete-time terms in magnetic resonance imaging (MRI) scanners and output after digital-to-analog conversion. Optionally, a frequency conversion can also be performed. The pulse shape is thus known, and the same excitation pulse can be emitted reproducibly.Based on this known pulse shape, reference measurements or calculations can be used to determine the extent to which a measured value for the flip angle, i.e., the preliminary flip angle value, deviates from a flip angle at the center of the excitation area. The flip angle at the center of the excitation area can then be determined as the corrected flip angle value. This allows inhomogeneous excitation in the measurement area, caused by the pulse shape of the selected excitation pulse, to be factored out, and any remaining inhomogeneities in the corrected flip angle value can be attributed to the inhomogeneities to be measured in the B1 field.

[0011] The first and second measured values ​​can describe the amplitude of an alternating magnetic field emitted by the excited object under investigation. A corresponding amplitude correlates with the transverse magnetization in the measurement area, which precesses in the constant B0 field of the magnetic resonance imaging scanner. To enable spatially resolved acquisition, frequency coding and / or phase coding can be used. Both approaches are known in the art and will therefore only be explained to the extent that they are relevant to the method according to the invention. In frequency coding, the total amplitude of an alternating field received by the object under investigation is not considered; instead, the amplitudes of the alternating field in different frequency ranges are taken into account for different slices in the direction of a readout gradient.In phase encoding, an additional gradient is briefly applied, which varies the phases of the preceding transverse magnetization depending on location. This results in a location-dependent mutual reinforcement or attenuation of the alternating fields radiated from different areas of the object under investigation due to the predetermined phase shift. In this case, the first and second measurements do not describe a local transverse magnetization, but rather amplitudes for specific spatial frequencies of the transverse magnetization in the direction of a phase-encoding gradient. By acquiring the first and second measurements for multiple phase encodings, a Fourier transform of the spatial frequencies into spatial space allows the determination of first and second local magnetization quantities that describe a spatially resolved transverse magnetization.In the case of phase encoding, the preliminary flip angle value for a specific measurement range is therefore determined from several first and second measured values, whereby a first and a second measured value are recorded for each phase encoding.

[0012] The preliminary flip angle value can be calculated as a quotient of the first and second measurements, or as a quotient of a value derived from the first and a value derived from the second measurement. The derived values ​​can, in particular, be the first and second local magnetization quantities. If the preliminary flip angle value is determined as a quotient of the measurements, the proportionality factors on which both measurements depend cancel each other out. This can simplify the calculation of a correction factor and reduce errors when transferring a correction factor determined by measurement. Alternatively, it would be possible to determine the preliminary flip angle value as a function of a difference between the measurements or the values ​​derived from them.

[0013] The first and second measurements can be taken after or partially before and partially after the application of the selected excitation pulse. The B1 field map or the flip angle values ​​can be acquired with spatial resolution, preferably for multiple layers or in a matrix-like manner in two or three dimensions.

[0014] In addition to the selected excitation pulse, further excitation or readout pulses can be used, in particular one per phase-encoding step. Additionally, gradients can be switched at various times during the measurement sequence, especially to perform layer selection or phase encoding, to generate gradient echoes, or to utilize so-called spoiler gradients that completely dephase any existing transverse magnetization.

[0015] The first and second measured values ​​can be recorded consecutively without intervening excitation pulses or separated from each other by at least one excitation pulse.

[0016] A predicted value for the preliminary flip angle can be determined based on the product of the flip angle assigned to the selected excitation pulse and a weighting function defined by the pulse shape of the selected excitation pulse. The correction factor is calculated as the quotient of the assigned flip angle of the selected excitation pulse and the predicted value. The weighting function can describe the variation in excitation intensity, i.e., the flip angle, along the layer selection gradient. As explained earlier, this variation depends on the pulse shape of the selected excitation pulse.

[0017] The predicted value can be calculated as a function of an integral over an integration range encompassing the measurement range, integrating the transverse magnetization dependent on the local flip angle, optionally taking a precession phase into account. Preferably, the spatial dependence of all excitation pulses prior to the respective measurement interval, which influence the local flip angle within the measurement range, is considered; that is, the associated flip angle of each excitation pulse is weighted with a pulse-shape-dependent weighting function.

[0018] The first and second measured values ​​each depend on the local transverse magnetization or the spatial frequencies of the local transverse magnetization at the respective measurement times. The transverse magnetization can be calculated from the flip angles of the preceding excitation pulses, taking into account switched gradient fields. The transverse magnetization and the measured values ​​may also depend on proportionality factors, which, for example, depend on the properties of the object under investigation or on inhomogeneities in the B1 field. These proportionality factors usually cancel out when calculating the correction factor. If this is not the case, estimates or factors from previous measurements can be used for any remaining proportionality factors.

[0019] The dependence of the actual achieved local flip angle on the pulse shape can be taken into account by the weighting function, at least for the selected excitation pulse, and preferably for all excitation pulses. A calculation of the expected measured values ​​can be performed, considering location-dependent flip angles for at least the selected excitation pulse. Thus, a locally varying transverse magnetization can be calculated. Within the context of predicting the first and second measured values, this transverse local magnetization can be integrated in the direction of the layer selection gradient or summed in a discrete calculation.

[0020] The correction factor can be calculated for several different assigned flip angles of the selected excitation pulse. In this case, it is possible to select the correction factor used to calculate the corrected flip angle value from the multiple calculated correction factors, depending on the first and / or second measurement, and in particular depending on the preliminary flip angle value.

[0021] The weighting function can be determined by a Fourier transform of the pulse shape and / or by a discrete-time simulation of the excitation of the nuclear spins by the selected excitation pulse. A Fourier transform of the pulse shape corresponds to the representation of the excitation pulse in the frequency domain. Since the resonance frequency of the nuclear spins is spatially dependent due to the layer selection gradient, this also corresponds to a location of the excitation, assuming that only resonant excitation of the nuclear spins occurs. A Fourier transform enables a simple and computationally efficient determination of the weighting function with typically sufficient accuracy.

[0022] By performing a discrete-time simulation of the excitation, for example by solving the Bloch equations with a discrete-time excitation field, it is possible to account for the fact that nuclear spins can also be excited at an excitation frequency adjacent to the resonance frequency. Such a simulation can be performed with a time resolution that corresponds to the time resolution of the discretely stored or calculated excitation pulses. However, secondary resonant excitations can also be accounted for by smearing the Fourier spectrum determined as described above, for example by convolution with a Gaussian function or by filtering.

[0023] As an alternative to the previously described calculation of a correction factor, it is also possible to determine it through supplementary measurements. For example, the associated flip angle can be known for at least one excitation pulse. In a preparatory measurement, the reference flip angle generated by this excitation pulse in a measurement area is determined, and the correction factor is then calculated as the quotient of the reference flip angle and the associated flip angle. This approach is based on the idea that, after prior calibration of the magnetic resonance imaging (MRI) scanner, the associated flip angle is identical to the set flip angle at the center of the excitation area, assuming a homogeneous B1 field. The measured reference flip angle corresponds to an average of the flip angles over the measurement area, with the flip angle typically decreasing with increasing distance from the center of the excited slice.The preliminary measurement establishes a relationship between a flip angle generated at the center of a measurement range and a flip angle averaged over the measurement range, and uses this relationship as a correction factor. Such reference measurements are preferably performed on a phantom that can be homogeneous and non-conductive and / or has the least possible influence on a B1 field distribution.

[0024] However, the correction factor can also be determined by using different methods with varying dependencies of a specific flip angle value on the pulse shape. Thus, a first and a second calibration flip angle can be determined using different first and second measurement methods, where the first calibration flip angle depends on the pulse shape of at least one excitation pulse used in the first method, and the second calibration flip angle is independent of the pulse shapes of the excitation pulses used in the second measurement method, with the correction factor being determined as a function of the first and second calibration flip angle values.The second calibration flip angle value is essentially independent of the pulse shapes of the excitation pulses, meaning that the effect of the pulse shape on the second calibration flip angle value is at least one, preferably at least two, orders of magnitude smaller than the effect of the pulse shape on the first calibration flip angle value.

[0025] To achieve a high degree of independence of the second calibration flip angle value from the pulse shape, block excitation can be used. In this method, the excitation pulse, whose flip angle is to be determined, excites a layer that is significantly wider than the layer excited during the measurement acquisition for flip angle determination. Phase-based methods for determining flip angles can also be used to determine the second calibration flip angle, since, as explained earlier, these are largely independent of the pulse shape.

[0026] The correction factor can be determined based on the preliminary flip angle value. In particular, multiple correction factors can be calculated and / or determined through additional measurements assigned to excitation pulses with the same pulse shape but different associated flip angles. The associated flip angles can be determined by the amplitude of the respective excitation pulse. Depending on the determined preliminary flip angle value, one of these correction factors can be selected, or interpolation can be performed between several of these correction factors. This approach can be advantageous when a layer profile, i.e., the spatial distribution of excitation by an excitation pulse, can change with the amplitude of the excitation pulse, i.e., with its associated flip angle.

[0027] The selected excitation pulse is played back twice in the measurement sequence before the acquisition of the first and second measured values, wherein the associated flip angle of the selected excitation pulse is less than 90°, wherein a first gradient field is activated between the plays to dephase the transverse magnetization, wherein a common excitation pulse is used to acquire the first and second measured values, wherein a second gradient field is activated in and / or before the first measurement interval or the second measurement interval to generate an echo of the previously dephased magnetization.

[0028] The combined excitation pulse acts as a readout pulse, meaning it rotates an available longitudinal magnetization to be measured in a transverse direction. A longitudinal magnetization corresponds to a magnetization parallel to the main magnetic field B0 of the magnetic resonance imaging (MRI) scanner. The phase of the precession of the transverse magnetization depends on the preceding preparation, that is, on the excitation by the two selected excitation pulses and the dephasing by the first gradient field. The first application of the selected excitation pulse tilts a portion of the existing longitudinal magnetization into the transverse plane, where it is dephased by the first gradient field. The second application of the selected excitation pulse tilts portions of this dephased transverse magnetization in the longitudinal direction.Optionally, any remaining transverse magnetization can be completely dephased by switching a spoiler gradient, so that it has no further influence on subsequent measurements.

[0029] The joint excitation pulse tilts a portion of the longitudinal magnetization into the transverse plane. The portion of the magnetization that was not previously tilted into the transverse plane by either of the two selected excitation pulses remains unprepared and is not dephased after the joint excitation pulse. The portion of the magnetization that was tilted into the transverse plane by the first excitation pulse and then back into the longitudinal direction of the B0 field by the second excitation pulse, after conversion to a transverse magnetization by the joint excitation pulse, exhibits a spatially dependent distribution determined by the dephasing effect of the first gradient field.

[0030] By measuring the first and second values ​​at different times within the first and second measurement intervals, and by switching at least the second gradient field, the portion of the magnetization dephased by the first gradient field and the portion not dephased by the first gradient field can be measured separately. The dephased and subsequently rephased magnetization is also referred to as a stimulated echo. The non-dephased magnetization can be measured as so-called "free induction decay." It can be advantageous to measure the non-dephased magnetization as a gradient echo by appropriately switching gradient fields after the second application of the selected excitation pulse; that is, first dephased by a gradient and then generating a gradient echo by applying a reverse gradient.A corresponding gradient echo for the non-dephased magnetization can be measured, for example, by applying the second gradient field before and / or during the first measurement interval. This provides an echo of the previously dephased magnetization in the first measurement interval, while simultaneously dephasing the previously non-dephased magnetization. After the first measurement interval and before and / or during the second measurement interval, a third gradient field can then be applied to cancel this subsequent dephasing and provide a gradient echo. In this case, the first measurement is proportional to that portion of the magnetization that was tilted into the transverse plane by the first application of the excitation pulse and back into the longitudinal direction by the second application.The second measurement is proportional to that part of the magnetization which was not tilted from the longitudinal direction by either of the two outputs of the excitation pulse.

[0031] The common excitation pulse can be played back multiple times, and after each playback, both the first and second measured values ​​can be read out, and at least the second gradient field can be switched. Additionally, phase coding can be performed after each common excitation pulse. In this case, the spatial frequencies of the corresponding magnetizations are determined by the first and second measured values, which can then be transformed by a subsequent Fourier transform into first and second local magnetization quantities that describe the respective local magnetization.

[0032] The associated flip angle of the common excitation pulse can be smaller than the associated flip angle of the selected excitation pulse. The associated flip angle of the common excitation pulse can be smaller by at least a factor of 2 or 4 than the associated flip angle of the common excitation pulse, and in particular, can be a maximum of 15°. This minimizes the effects of excitation inhomogeneities caused by the common excitation pulse. The common excitation pulse can excite a sub-region that does not encompass the entire excitation range of the selected excitation pulse. The common excitation pulse can therefore have a narrower spectral width than the selected excitation pulse. Consequently, the first and second measurements are taken for a sub-region of the layer excited by the selected excitation pulse. Thus, only the region in which the excitation pulse results in a relatively homogeneous excitation is investigated.

[0033] In an alternative embodiment of the method according to the invention, it is possible to acquire the first measured value before the selected excitation pulse is played and the second measured value after the selected excitation pulse has been played. For example, a two- or three-dimensional measurement can first be performed with a specific measurement sequence, then the selected excitation pulse is played, and the resulting transverse magnetization is dephased by a spoiler gradient, after which the preceding measurement sequence can be repeated. From the ratio of the first measured values ​​of the first execution of the measurement sequence before the selected excitation pulse is played and the second measured values ​​of the second execution of the measurement sequence after the selected excitation pulse has been played, a preliminary flip angle value for the selected excitation pulse can be determined, which can then be corrected as explained above.

[0034] In addition to the method according to the invention, the invention relates to a magnetic resonance imaging (MRI) scanner with a control unit, at least one transmitter, and at least one receiver. The control unit is configured to control the transmitter, to acquire the first and second measured values ​​via the receiver, and to process the first and second measured values ​​to determine the B1 field map according to the method according to the invention. The control unit can be configured to drive one or more gradient coils to provide slice selection gradients, phase coding gradients, gradients for dephasing or rephasing transverse magnetizations, and / or spoiler gradients. The receiver can be used to receive high-frequency alternating electromagnetic fields emitted by a probe excited by the excitation pulses.

[0035] Further advantages and details of the invention are shown in the following exemplary embodiments and the accompanying drawings. These show schematically: Fig. 1 a flowchart of an embodiment of the method according to the invention, Fig. 2 one in the in Fig. Measurement sequence used in the exemplary embodiment shown in 1, Fig. 3. A spatial distribution of the longitudinal and transverse magnetization in the object under investigation after the application of the Fig. 2 shown excitation pulses, Fig. 4. Calculating the correction factor in the Fig. 1 shown embodiment, Fig. 5 a flowchart of a further embodiment of the method according to the invention, and Fig. 6 an embodiment of a magnetic resonance tomograph according to the invention.

[0036] Fig. Figure 1 shows an embodiment of a method for determining a B1 field map in a magnetic resonance imaging scanner, which describes the local field strength of a high-frequency electromagnetic B1 alternating field, irradiated by a transmitter device to excite nuclear spins, for several matrix-like arranged measurement areas. The method is described with reference to the one in Fig. 2. Schematically represented measurement sequence explained.

[0037] In step S1, a layer of an object under investigation is first prepared, as described in section 2 of the sequence in Fig. Figure 2 illustrates this process. A layer selection gradient (not shown) is applied, and a layer of the test object is excited by the selected excitation pulse 1. The excitation pulse 1 rotates the magnetization at the center of the excited layer by an associated flip angle α. The amplitude of the excitation pulse 1 is chosen such that the flip angle α is < 90°.

[0038] Before the excitation pulse 1 is applied, the magnetization is essentially longitudinally oriented, meaning it points in the direction of the main magnetic field B0 of the magnetic resonance imaging scanner. The excitation pulse 1 provides a transverse magnetization proportional to sin(α). The remaining longitudinal magnetization is proportional to cos(α). The magnetization profile of the slice after excitation is shown in Fig. Figure 3 shows the transverse magnetization, and curve 4 shows the remaining longitudinal magnetization. If the excitation pulse 1 were an ideal sinc pulse, curve 3 would have a rectangular shape, meaning a sharply defined layer would have a constant flip angle and thus a constant magnetization. If a finite-length excitation pulse 1 is used, then, depending on the chosen window function, either oscillations of the flip angle and thus of the magnetization across the layer thickness or a rounding of the rectangular excitation profile occur. Oscillations should generally be avoided, which is why, as in Fig. 3, represented by curve 3, uses rounded layer profiles.

[0039] The flip angle α assigned to the excitation pulse 1 and the magnetization assigned to this flip angle are determined exclusively at the center of the layer, i.e. at the zero point in the X direction. Fig. 3, achieved. However, in magnetic resonance imaging (MRI) scanners, signals averaged over the entire excited area are always acquired. To determine a correct flip angle α for the excitation pulse 1, i.e., the flip angle at the center of the excited slice, the in Fig. The rounding of the layer profile shown in step 3 must therefore be taken into account during data evaluation. This will be explained later with reference to steps S8 and S9.

[0040] Between the first and second outputs of excitation pulse 1, a first gradient field 5 is activated to dephase the transverse magnetization. Transverse magnetizations are detected in magnetic resonance imaging (MRI) scanners by receiving the electromagnetic radiation emitted by the excited object under investigation. If several regions of the object emit radiation in phase, the received intensities add up. After dephasing, however, the oscillation phases in the individual regions of the object are shifted relative to each other in such a way that the received radiation intensity is reduced to essentially zero. The second output of excitation pulse 1 partially tilts the dephased transverse magnetization back in the longitudinal direction. This tilt can occur either towards or against the main field B0.The longitudinal magnetization that is available after the second excitation pulse 1 thus decays into two parts, a first part being proportional to cos. 2 (α) is, which has not been pivoted and dephased into the transverse plane by either of the two excitation pulses 1, and a second part, which is proportional to ½ sin 2 (α) has been dephased and is available for a stimulated echo during readout in region 6 of the sequence. Before such readout occurs, a spoiler gradient 7 is applied, which completely dephases the remaining transverse magnetization so that it is no longer included in subsequent measurements.

[0041] In steps S2 to S5, several first and second measurements are acquired. Multiple phase encodings are performed to enable spatially resolved measurement of the flip angle values. For this purpose, a common excitation pulse 8 is applied in step S2, which has a relatively small flip angle β and thus only flips a small portion of the longitudinal magnetization into the transverse plane. The layer profile, i.e., the spatial distribution of the magnetization caused by the excitation pulse 8, is shown in Fig. 3 is represented as curve 9. The corresponding flip angle for the excitation pulse 8 is also only reached in the center of the excited layer. The transverse magnetization in the layer center is proportional to sin(β). As in Fig. As shown in Figure 3, a common excitation pulse 8 is used, which excites a layer with a smaller width than the selected excitation pulse 1. By combining a flip angle β that is at least two to four times smaller than the flip angle α, and a relatively narrow layer profile for the common excitation pulse 8, the effects of the layer profile of the selected excitation pulse 1 on the preliminary flip angle value determined below can be reduced, with the correction in step S9, which will be explained later, enabling further improvements.

[0042] In step S3, a phase-encoding gradient 10 is applied, the gradient direction of which is essentially perpendicular to the layer selection gradient. The use of a phase-encoding gradient 10 enables phase encoding in one or two dimensions to allow spatially resolved determination of the flip angle of the selected excitation pulse 1. In this case, the first and second measurements do not directly describe the first and second local magnetization quantities that describe a local magnetization of the object under investigation, but rather spatial frequencies of these local magnetization quantities. As explained below, the first and second measurements are acquired for multiple phase encodings, after which the first and second local magnetization quantities can be determined by a Fourier transform, from which the preliminary flip angle value is calculated.The use of phase coding is well known in the field of magnetic resonance imaging and will not be explained in detail.

[0043] In step S4, the first and second measurements are acquired. For this purpose, a second gradient field 11 is first activated, which is aligned with the previously activated first gradient field 5. Simultaneously, the high-frequency radiation emitted by the object under investigation is acquired via an antenna and an analog-to-digital conversion in a first measurement interval 12. The gradient field 11 causes a portion of the magnetization dephased by the gradient field 5 to be rephased, resulting in the acquisition of a stimulated echo at time 13. The intensity of the stimulated echo, and thus the first measurement, is proportional to ½ sin (β)*sin 2(α). A gradient echo is detected during the second measurement interval 15 at time 16 by a third gradient field 14 directed opposite to the first and second gradient fields 5, 11. This echo is provided by a portion of the magnetization that was not tilted into the transverse plane by any of the selected excitation pulses 1. As a result, the corresponding transverse magnetization was initially in phase after the common excitation pulse 8, then dephased by gradient 11, and finally rephased by gradient 14. The amplitude of the gradient echo is proportional to sin (β) * cos 2 (α).

[0044] In step S5, it is checked whether the first and second measurements have been acquired for all predefined phase codings. If this is not the case, the procedure is repeated from step S2 for the remaining phase codings. After acquiring the first and second measurements for all phase codings, the first and second measurements are each Fourier-transformed in step S6 to obtain a first and a second local magnetization quantity. Dividing the first local magnetization quantity by the second local magnetization quantity cancels out the proportionality factors and the dependence on the flip angle β, so that the local flip angle α can be determined from the first and second local magnetization quantities. The quotient of the first and second local magnetization quantities is ½ tan β. 2 (α). A preliminary flip angle value can be calculated in step S7 using this relationship.

[0045] To account for the influence of the pulse shape and, in particular, the finite pulse length, a correction factor dependent on the pulse shape of the selected excitation pulse 1 is determined in step S8. An example of the calculation of such a correction factor will be given later with reference to Fig. 4 will be explained. Subsequently, in step S9, the preliminary flip angle value calculated in step S7 is multiplied by the correction factor calculated in step S8 to obtain a corrected flip angle value. The corrected flip angle value is proportional to the local field strength of the B1 field, which is used to output the excitation pulses 1 and 8. It can therefore be stored in the B1 field map to describe the field strength of the B1 field.

[0046] Fig. Figure 4 shows a flowchart for calculating the correction factor. For this purpose, in step S10, the pulse shape of the [unclear text] is [unclear text]. Fig. 2 shown selected excitation pulse 1 and in step S11 the pulse shape of the in Fig. The common excitation pulse 8 shown in Figure 2 is specified. The pulse shape can be specified as a discrete-time representation of the respective pulse, in particular as a digital representation of a signal to be output via a digital-to-analog converter. However, it would also be possible to specify the pulses in the form of analytical functions.

[0047] In steps S12 and S13, a weighting function is determined for each of the excitation pulses. For this purpose, a Fourier transform is performed to convert the pulse representation defined in steps S10 and S11 into a representation of the respective pulse in the frequency domain. As explained previously, a representation of the respective pulse in Fourier space corresponds to an excitation profile for that pulse. Since the object under investigation is located in a gradient field during excitation, there is a direct relationship between an applied excitation frequency and the location where resonant excitation of the nuclear spins occurs. A representation of the pulse in the frequency domain thus directly corresponds to a representation of the excitation pattern in spatial space, where the relationship between the frequency axis and the spatial axis is defined by the gradient field.The weighting function is specified in each case by normalizing the maximum of the frequency distribution to one.

[0048] The excitation pulses used are, for example, sinc pulses whose edge regions are smoothed by a window function. In this case, the layer profiles, i.e., the Fourier transform, essentially correspond to the shape of curves 3 and 9, respectively. Fig. 3, which represent the magnetization after the respective excitation pulse 1, 8 has been played.

[0049] Determining an excitation profile by Fourier transformation typically yields a sufficiently accurate excitation profile. To further improve the accuracy of the method, a discrete-time simulation of the excitation could be used instead of the Fourier transformation. This could be achieved, for example, by solving the Bloch equations for discrete-time excitation by the respective excitation pulse 1, 8 to obtain a spatial distribution of the magnetization. In this case as well, the resulting function should be normalized so that its maximum lies at one.

[0050] In steps S14 and S15, the respective flip angles for excitation pulses 1 and 8 are specified and, in steps S16 and S17, multiplied by the weighting function calculated in steps S12 and S13, respectively. Step S16 thus provides a function that describes the spatial dependence of the local flip angle generated by the selected excitation pulse 1 by multiplying the flip angle assigned to the selected excitation pulse 1 by the weighting function specified in step S12, which depends on the pulse shape of the selected excitation pulse 1. Similarly, step S17 provides a function that describes the spatial distribution of the local flip angle generated by the common excitation pulse 8.

[0051] In steps S18 and S19, predictive values ​​for the first and second measurements are calculated. The phase coding performed during the measurement can generally be neglected in this calculation. Phase coding is typically performed perpendicular to the direction of the layer selection gradient. However, the determination of the correction factor only considers a change in the flip angle in the direction of the layer selection gradient due to the pulse shape of the excitation pulse. Phase coding thus only results in a proportionality factor for the first and second measurements, which is at least approximately the same for both measurements. This proportionality factor therefore cancels out during the subsequent calculation, so it is sufficient to determine the correction factor or the predictive values ​​for the first and second measurements using only phase coding, especially for measurements without a phase coding gradient.Furthermore, the calculation neglects other factors that can influence the first and second measurements, but which are essentially the same for both measurements and are determined, for example, by relaxation effects. The predictive values ​​for the first and second measurements are thus calculated, neglecting any similar proportionality factors, solely based on the respective transverse and lateral magnetizations present after excitation pulses 1 and 8 of the measurement sequence.

[0052] The first measured value, i.e., the received radiation intensity in the first measurement interval 12, describes that part of the magnetization which is in the Fig. In the measurement sequence shown in Figure 2, the first excitation pulse 1 tilted the system into the transverse plane, where it was dephased by the gradient field 5, then tilted into the longitudinal direction by the second excitation pulse 1, and finally tilted back into the transverse direction by the common excitation pulse 8, where it was rephased by the gradient field 11. The first measured value S1 is thus calculated as follows: S1=K*1 / 2∫sin(β(z))sin2(α(z)) dz.

[0053] K represents a constant that describes the initially available magnetization as well as other quantities not considered in detail. Since K is assumed to be the same for the first and second measurements, and, as explained below, the predicted value for the preliminary flip angle depends on the quotient of the predicted values ​​for the first and second measurements, this constant can be set to 1. The quantities α(z) and β(z) correspond to the assigned flip angles of the excitation pulses 1 and 8, weighted by the weighting function, which were calculated in steps S16 and S17.

[0054] The second measurement is taken in the Fig. The measurement sequence shown in Figure 2 is measured in measurement interval 15 and relates to a gradient echo provided by the gradient fields 11, 14, the intensity of which depends on the part of the magnetization that was not tilted into the transverse plane by either the first or the second emission of the selected excitation pulse 1. The predicted value for the second measurement is therefore calculated as S2=K*∫sin(β(z))cos2(α(z)) dz.

[0055] In step S20, a prediction value α' for the preliminary flip angle value is calculated from the prediction values ​​S1 and S2 for the first and second measurements calculated in steps S18 and S19. The calculation is performed in the same way as the calculation of the preliminary flip angle value from the local magnetic quantities in step S6. Fig. 1. The forecast value is calculated as follows: α'=tan−1(2∗S1S2) In step S21, the correction factor is calculated by dividing the predicted value α' for the preliminary flip angle value by the flip angle α specified in step S14 and assigned to the selected excitation pulse 1.

[0056] Fig. Figure 5 shows a further embodiment of a method for determining a B1 field map. The method is based on the approach of first performing a phase-coded measurement, then applying the selected excitation pulse, dephasing the resulting transverse magnetization using a spoiler gradient, and then repeating the initial measurement. The readout pulses for excitation before the individual phase-coding steps have significantly smaller flip angles than the selected excitation pulse. The two separate measurements before and after the excited excitation pulse allow the determination of how much magnetization is "consumed" by the selected excitation pulse, and thus its flip angle.

[0057] The procedure assumes an equilibrium state of the object under investigation, in which the magnetization is essentially aligned in the direction of the magnetic resonance imaging (MRI) scanner's base magnetic field B0, depending on the strength of the MRI scanner's base magnetic field B0 and the temperature. In step S22, a readout pulse, i.e., an excitation pulse with a small associated flip angle, is first generated. Such a readout pulse has only a negligible effect on the remaining longitudinal magnetization but provides sufficient transverse magnetization to allow measurement. In the described procedure, the measurement is performed as a "free induction decay" measurement. Alternatively, it would of course be possible to detect the transverse magnetization using spin or gradient echo measurements, or similar methods.

[0058] In step S23, a phase gradient is switched, and in step S24, an analog-to-digital converter is activated to measure the received radiation intensity and thus the transverse magnetization within a measurement interval. In step S25, it is checked whether all phase codings have already been performed. If not, the procedure is repeated from step S22.

[0059] In step S26, the selected excitation pulse is applied to tilt a portion of the longitudinal magnetization into the transverse plane according to its associated flip angle. Subsequently, in step S27, a spoiler gradient is applied to dephase this magnetization, i.e., to change its phase so that it is no longer measurable in immediately following measurements.

[0060] Steps S28 to S31 correspond to steps S22 to S25, i.e., the previous measurement is repeated after the selected excitation pulse has been played.

[0061] In step S32, the results of both measurements are transformed into spatial space, so that for each measured point in spatial space, there is a first measurement value from the first measurement, which comprises steps S22 to S25, and a second measurement value from the second measurement, which comprises steps S28 to S31. From the first and second measurements, a preliminary flip angle value can be calculated, which describes the flip angle locally caused by the selected excitation pulse. If the influence on the longitudinal magnetization by the readout pulses played out in steps S22 and S28 is neglected, the quotient of the second and the first measurement value is approximately equal to the cosine of the local flip angle value. Using this relationship, a preliminary flip angle value can be calculated for each measured point in spatial space. In step S33, as described above, Fig. 4 explains, a correction factor is calculated and multiplied in step S34 by the preliminary flip angle value to calculate a corrected flip angle value and to store this or a derived value in a B1 field map.

[0062] The use of calculated correction factors in the detailed examples is purely illustrative. A correction factor could also be determined by direct measurement or by comparing measurements with a high or low dependence of the determined flip angle value on the pulse shapes used.

[0063] Fig.Figure 6 shows a magnetic resonance imaging (MRI) scanner 17 with a control unit 18, at least one transmitter 19, and at least one receiver 20. The control unit 18 is configured to control the transmitter 19 and to acquire the first and second measured values ​​via the receiver 20, as well as to process the first and second measured values ​​to determine the B1 field map. This can be done according to one of the methods described above. The control unit 18 also controls several gradient coils 21, 22 for generating gradient fields.

[0064] Although the invention has been illustrated and described in detail by the preferred embodiment, the invention is not limited by the disclosed examples and other variations can be derived by the person skilled in the art without leaving the scope of protection of the invention.

Claims

[1] Method for determining a B1 field map in a magnetic resonance tomograph (17) which describes a local field strength of a high-frequency electromagnetic B1 alternating field irradiated by a transmitter (19) to excite nuclear spins for at least one measurement range, wherein, within a measurement sequence, several excitation pulses (1, 8) are irradiated by the transmitter (19), each of which changes the magnetization of an excitation range encompassing the measurement range according to an associated flip angle, and a first measurement value is recorded by a receiver (20) in a first measurement interval (12) and a second measurement value is recorded in a second measurement interval (15), relating to the magnetization in the measurement range, wherein, depending on the first and the second measurement value, a preliminary flip angle value is determined for the flip angle associated with a selected excitation pulse (1) in the measurement range,where the associated flip angle of the selected excitation pulse (1) is less than 90°, , characterized by, that a correction factor dependent on a pulse shape of the selected excitation pulse (1) is determined, wherein a corrected flip angle value is determined by multiplying the preliminary flip angle value by the correction factor, wherein the corrected flip angle value or a value derived therefrom is stored in the B1 field map to describe the local field strength in the measurement area, wherein in the measurement sequence the selected excitation pulse (1) is played back twice before the acquisition of the first and the second measurement value, wherein a first gradient field (5) is activated between the plays back to dephase the transverse magnetization, wherein a common excitation pulse (8) is used to acquire the first and the second measurement value, wherein in and / or before the first measurement interval (11) in which the first measurement value is acquired, or the second measurement interval (12) in which the second measurement value is acquired,a second gradient field (11) is activated to generate an echo of the previously dephased magnetization. [2] Method according to claim 1, characterized by , that a prediction value for the preliminary flip angle value is determined as a function of a product of the flip angle assigned to the selected excitation pulse and a weighting function specified as a function of the pulse shape of the selected excitation pulse (1), wherein the correction factor is calculated as the quotient of the assigned flip angle of the selected excitation pulse (1) and the prediction value. [3] Method according to claim 2, characterized by , that the weighting function is determined by a Fourier transformation of the pulse shape and / or by time-discrete simulation of the excitation of the nuclear spins by the selected excitation pulse (1). [4] Method according to claim 1, characterized by, that for at least one excitation pulse (1, 8) the associated flip angle is known, wherein in a preparatory measurement the reference flip angle value generated by this excitation pulse (1, 8) in a measuring range is determined, after which the correction factor is determined as the quotient of the reference flip angle value and the associated flip angle. [5] Method according to claim 1, characterized by , that a first and a second calibration flip angle value are determined by different first and second measurement procedures, wherein the first calibration flip angle value depends on a pulse shape of at least one excitation pulse (1, 8) used in the first procedure and the second calibration flip angle is independent of the pulse shapes of the excitation pulses (1, 8) used in the second measurement procedure, wherein the correction factor is determined as a function of the first and the second calibration flip angle value. [6] Method according to any of the preceding claims, characterized by , that the correction factor is determined depending on the preliminary flip angle value. [7] Method according to any of the preceding claims, characterized by , that the associated flip angle of the common excitation pulse (8) is smaller than the associated flip angle of the selected excitation pulse (1). [8] Method according to any of the preceding claims, characterized by , that the common excitation pulse (8) excites a sub-area which does not encompass the entire excitation range of the selected excitation pulse (1). [9] Magnetic resonance imaging scanner with a control unit (18), at least one transmitting unit (19) and at least one receiving unit (20), characterized by, that the control device (18) is designed to control the transmitting device (19), to acquire the first and second measured values ​​via the receiving device (20) and to process the first and second measured values ​​to determine the B1 field map according to the method according to one of the preceding claims.

Images