METHOD, CONTROLLER, AND MAGNETIC RESONANCE APPARATUS FOR MODEL-FREE DETERMINATION OF IMAGE REGIONS WITH ANOMALOUS DIFFUSION USING DIFFUSION-WEIGHTED MAGNETIC RESONANCE IMAGE DATA

Abstract

In a magnetic resonance apparatus and a method and controller for operating such an apparatus, first and second diffusion-weighted image data are reconstructed from first and second diffusion-encoded raw data that were respectively acquired using different diffusion-encoding gradient pulse sub-sequences. The different sub-sequences differ by respectively having a different parameter that characterizes the respective sub-sequence as a function of time. The first and second reconstructed image data are compared and a deviation of the image data from normal Gaussian diffusion behavior is determined model-free on the basis of the comparison result.

Claims

1. A method for determining a deviation of diffusion-weighted magnetic resonance (MR) image data of an examination object from Gaussian diffusion behavior, comprising: operating an MR data acquisition scanner to acquire first diffusion-encoded raw data by executing a first diffusion-encoding gradient pulse sub-sequence; operating said MR data acquisition scanner to acquire at least second diffusion-encoded raw data by executing at least one second diffusion-encoding gradient pulse sub-sequence, each of said first and second diffusion-encoding gradient pulse sub-sequence comprising at least one parameter that determines a sub-sequence characteristic as a function of time, and wherein said at least one parameter is changed during the acquisition of said at least second diffusion-encoding raw data relative to that parameter during acquisition of said first diffusion-encoded raw data; in a processor, reconstructing first diffusion-encoded image data from said first diffusion-encoded raw data and reconstructing at least second diffusion-encoded image data from said at least second diffusion-encoded raw data; in said processor, comparing said first diffusion-encoded image data and said at least second diffusion-encoded image data to obtain a comparison result; and in said processor, making a model-free determination, based on said comparison result, of a deviation, of at least one of said first diffusion-encoded image data or said at least second diffusion-encoded image data, from Gaussian diffusion behavior, and making an electronic designation of said deviation available as an output from said processor.

2. A method as claimed in claim 1 comprising comparing said first diffusion-encoded image data and said at least second diffusion-encoded image data to obtain said comparison result as a comparative value of respective image intensities of said first diffusion-encoded image data and said at least second diffusion-encoded image data, and determining said deviation dependent on a size of said comparative value.

3. A method as claimed in claim 1 comprising comparing said first diffusion-weighted image data and said at least second diffusion-weighted image data to obtain said comparison result as a comparative value selected from the group consisting of a difference of respective image intensities of said first diffusion-encoded image data and said at least second diffusion-encoded image data, and a quotient of respective image intensities of said first diffusion-encoded image data and said at least second diffusion-encoded image data.

4. A method as claimed in claim 1 comprising reconstructing each of said first diffusion-encoded image data and said at least second diffusion-encoded image data to depict diffusion in a single defined direction of a diffusion gradient used to acquire said first diffusion-encoded raw data and said at least second diffusion-encoded raw data.

5. A method as claimed in claim 1 comprising, in said processor, deriving a first tensor parameter from a diffusion tensor represented by said first diffusion-encoded image data, and deriving a second tensor parameter from a diffusion tensor represented by said at least second diffusion-encoded image data, and implementing said comparison as a comparison of said first and second tensor parameters to obtain a tensor parameter comparison result, and determining said deviation from Gaussian diffusion behavior from said tensor parameter comparison result.

6. A method as claimed in claim 5 comprising: operating said MR data acquisition scanner to acquire reference raw data by executing a reference scanning sequence; in said processor, determining said first tensor parameter from said first diffusion-encoded raw data and said reference raw data; and determining said at least second tensor parameter from said at least second diffusion-encoded raw data and said reference raw data.

7. A method as claimed in claim 5 comprising determining each of said first and at least second tensor parameters as a parameter selected from the group consisting of an apparent diffusion coefficient (ADC), fractional anisotropy, and relative anisotropy.

8. A method as claimed in claim 1 comprising reconstructing each of said first diffusion-encoded image data and said at least second diffusion-encoded image data as track-weighted image data comprising diffusion information depicted in three defined directions.

9. A method as claimed in claim 5 comprising: operating said MR data acquisition scanner to acquire reference raw data by executing a reference scanning sequence; in said processor, determining said first tensor parameter from said first diffusion-encoded raw data and said reference raw data; determining said at least second tensor parameter from said at least second diffusion-encoded raw data and said reference raw data; operating said MR data acquisition scanner to acquire at least third raw data with or without diffusion-weighting; and in said processor, determining each of said first and second tensor parameters using said first and said at least second raw data, said third raw data and said reference data using a repression method.

10. A method as claimed in claim 1 comprising: operating said MR data acquisition scanner to acquire third raw data using a third raw data acquisition parameter, which determines said characteristic, that is changed compared to the parameters respectively used for acquiring said first diffusion-weighted raw data and said at least second diffusion-weighted raw data, and generating said comparison result as a comparative value representing a model-free statistical parameter.

11. A method for depicting distribution of a deviation from Gaussian diffusion behavior in a field of view of an examination object, comprising: operating an MR data acquisition scanner to acquire first diffusion-encoded raw data by executing a first diffusion-encoding gradient pulse sub-sequence; operating said MR data acquisition scanner to acquire at least second diffusion-encoded raw data by executing at least one second diffusion-encoding gradient pulse sub-sequence, each of said first and second diffusion-encoding gradient pulse sub-sequence comprising at least one parameter that determines a sub-sequence characteristic as a function of time, and wherein said at least one parameter is changed during the acquisition of said at least second diffusion-encoding raw data relative to that parameter during acquisition of said first diffusion-encoded raw data; in a processor, reconstructing first diffusion-encoded image data from said first diffusion-encoded raw data and reconstructing at least second diffusion-encoded image data from said at least second diffusion-encoded raw data; in said processor, comparing said first diffusion-encoded image data and said at least second diffusion-encoded image data to obtain a comparison result; in said processor, generating a map of a spatial distribution of said deviation in the field of view of the examination object; and at a display in communication with said processor, displaying said map.

12. A controller for a magnetic resonance (MR) apparatus, said controller comprising: an input configured to receive first diffusion-encoded raw data acquired by executing a first diffusion-encoding gradient pulse sub-sequence with an MR data acquisition scanner; said input being configured to also receive at least second diffusion-encoded raw data acquired by executing at least one second diffusion-encoding gradient pulse sub-sequence with said MR data acquisition scanner, each of said first and second diffusion-encoding gradient pulse sub-sequence comprising at least one parameter that determines a sub-sequence characteristic as a function of time, and wherein said at least one parameter is changed during the acquisition of said at least second diffusion-encoding raw data relative to that parameter during acquisition of said first diffusion-encoded raw data; a processor configured to reconstruct first diffusion-encoded image data from said first diffusion-encoded raw data and to reconstruct at least second diffusion-encoded image data from said at least second diffusion-encoded raw data; said processor being configured to compare said first diffusion-encoded image data and said at least second diffusion-encoded image data to obtain a comparison result; and said processor being configured to make a model-free determination, based on said comparison result, of a deviation, of at least one of said first diffusion-encoded image data or said at least second diffusion-encoded image data, from Gaussian diffusion behavior, and to make an electronic designation of said deviation available as an output from said processor.

13. A magnetic resonance (MR) apparatus comprising: an MR data acquisition scanner; a control computer configured to operate said MR data acquisition scanner to acquire first diffusion-encoded raw data by executing a first diffusion-encoding gradient pulse sub-sequence; said control computer being configured to operate said MR data acquisition scanner to acquire at least second diffusion-encoded raw data by executing at least one second diffusion-encoding gradient pulse sub-sequence, each of said first and second diffusion-encoding gradient pulse sub-sequence comprising at least one parameter that determines a sub-sequence characteristic as a function of time, and wherein said at least one parameter is changed during the acquisition of said at least second diffusion-encoding raw data relative to that parameter during acquisition of said first diffusion-encoded raw data; a processor configured to reconstruct first diffusion-encoded image data from said first diffusion-encoded raw data and to reconstruct at least second diffusion-encoded image data from said at least second diffusion-encoded raw data; said processor being configured to compare said first diffusion-encoded image data and said at least second diffusion-encoded image data to obtain a comparison result; and said processor being configured to make a model-free determination, based on said comparison result, of a deviation, of at least one of said first diffusion-encoded image data or said at least second diffusion-encoded image data, from Gaussian diffusion behavior, and to make an electronic designation of said deviation available as an output from said processor.

14. A non-transitory, computer-readable data storage medium encoded with programming instructions, said storage medium being loaded into a computer system of a magnetic resonance (MR) apparatus that comprises an MR data acquisition scanner, said programming instructions causing said computer system to: receive first diffusion-encoded raw data acquired by executing a first diffusion-encoding gradient pulse sub-sequence with an MR data acquisition scanner; receive at least second diffusion-encoded raw data acquired by executing at least one second diffusion-encoding gradient pulse sub-sequence with said MR data acquisition scanner, each of said first and second diffusion-encoding gradient pulse sub-sequence comprising at least one parameter that determines a sub-sequence characteristic as a function of time, and wherein said at least one parameter is changed during the acquisition of said at least second diffusion-encoding raw data relative to that parameter during acquisition of said first diffusion-encoded raw data; reconstruct first diffusion-encoded image data from said first diffusion-encoded raw data and reconstruct at least second diffusion-encoded image data from said at least second diffusion-encoded raw data; compare said first diffusion-encoded image data and said at least second diffusion-encoded image data to obtain a comparison result; and make a model-free determination, based on said comparison result, of a deviation, of at least one of said first diffusion-encoded image data or said at least second diffusion-encoded image data, from Gaussian diffusion behavior, and make an electronic designation of said deviation available as an output from said computer system.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0078] FIG. 1 is a pulse graph for a sequence known as a Stejskal-Tanner sequence for diffusion-weighted imaging.

[0079] FIG. 2 is a flowchart of the basic steps of a method for determining a deviation of diffusion-weighted magnetic resonance image data of an examination object from normal diffusion behavior, according to an exemplary embodiment of the invention.

[0080] FIG. 3 shows anomaly maps in which regions are marked on an anatomy map in which an anomalous diffusion occurs.

[0081] FIG. 4 is a flowchart of the basic steps of a method for determining a deviation of diffusion-weighted magnetic resonance image data of an examination object from normal diffusion behavior, according to a second exemplary embodiment of the invention.

[0082] FIG. 5 is a flowchart of the basic steps of a method for determining a deviation of diffusion-weighted magnetic resonance image data of an examination object from normal diffusion behavior, according to a third exemplary embodiment of the invention.

[0083] FIG. 6 is a flowchart of the basic steps of a method for determining a deviation of diffusion-weighted magnetic resonance image data of an examination object from normal diffusion behavior, according to a fourth exemplary embodiment of the invention.

[0084] FIG. 7 is a block diagram of a magnetic resonance apparatus according to an exemplary embodiment of the invention.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0085] FIG. 1 shows a Stejskal-Tanner sequence 100. This is by far the most important diffusion-weighted pulse sequence. The first line of the graph, which is identified by RF/ADC, shows an RF excitation pulse 101 having a duration TRF.sub.1, which is radiated at the beginning of a pulse sequence at the same times as a slice selection gradient 106 (see second line GS), and an RF refocusing pulse 102 having a duration TRF.sub.2, which is radiated between two diffusion contrast gradient pulses 104, 105 (see third line GDW) and with which a slice selection gradient 107 (see second line GS) is likewise switched at the same time. The gradient pulses 104, 105 have the same polarity and usually the same amplitude and duration. The RF refocusing pulse 102 forms a spin echo 103 (see first line) which in the illustrated example is read out with an EPI echo readout train EPIR, composed of multiple readout windows. Furthermore, the graph in FIG. 1 shows in the second line from the bottom a gradient pattern GR in the readout direction (frequency-encoding direction) and in the lowest line a gradient pattern GP in the phase encoding direction.

[0086] The echo time TE is the time between the HF excitation pulse 101 and the echo 103. The formation of the echoes results from the diminishing moment of all activated gradients along the three axes. The position of the refocusing pulse is preferably chosen at TE/2 in order to also compensate the moments generated by static magnetic field gradients that cannot be influenced (e.g. due to BO inhomogeneities in the object) exactly at the echo instant.

[0087] The following is obtained for the b-value for the Stejskal-Tanner pattern shown in FIG. 1 having symmetrical, trapezoidal diffusion gradients 104, 105 with amplitude G and diminishing ramp time T.sub.r:

b=4π.sup.2γ.sup.2G.sup.2[δ.sup.2(Δ−δ/3)].  (1)

[0088] Here δ is the duration of a gradient and Δ is the time that elapses between activation on of the two gradient pulses 104, 105, i.e. the interval between those gradient pulses. The time intervals just defined are shown in FIG. 1. G is the gradient strength or gradient amplitude of the diffusion gradients 104, 105. The constant γ indicates the gyromagnetic ratio. Only the contribution of the diffusion contrast gradients 104 and 105 for the b-value of the sequence is taken into account in formula 1. The strong diffusion contrast gradients 104, 105 of the Stejskal-Tanner sequence make the sequence sensitive to macroscopic movement as well as undirected molecular Brownian motion, since the signal from moved spins is dephased. Anomalies, which are caused by movements overlaid on the Brownian motions, therefore easily occur during image recording using a sequence of this kind.

[0089] FIG. 2 is a flowchart 200 for one possible scanning process in which the inventive method is used according to a first exemplary embodiment.

[0090] With this method, first diffusion-weighted raw data DRD.sub.1 are acquired in step 2.Ia with a first pulse sequence having a first diffusion-encoding gradient pulse sub-sequence GD.sub.1, as is shown in FIG. 1. The first diffusion-encoding gradient pulse sub-sequence GD.sub.1 is characterized by time interval parameters δ, Δ, which were already described in connection with FIG. 1, having first parameter values δ.sub.1, Δ.sub.1. These first time interval parameter values δ.sub.1, Δ.sub.1 define the characteristic as a function of time of the first diffusion-encoding gradient pulse sub-sequence GD.sub.1. A first b-value b1 is calculated from the first time interval parameter values δ.sub.1, Δ.sub.1 according to equation 1. In a step 2.Ib second diffusion-weighted raw data DRD.sub.2 are additionally acquired with a second pulse sequence having a second diffusion-encoding gradient pulse sub-sequence GD.sub.2. The second diffusion-encoding gradient pulse GD.sub.2 is characterized by second time interval parameter values δ.sub.2, Δ.sub.2 which differ quantitatively from the first parameter values δ.sub.1, Δ.sub.1 at least with respect to one of the parameters δ, Δ. For example, δ.sub.2>δ.sub.1. A second b-value b2 can be calculated from the second parameter values δ.sub.2, Δ.sub.2. The two b-values b1, b2 match or should at least approximately match in the first exemplary embodiment illustrated in FIG. 2, and this may always be achieved by suitable choice of the second diffusion-encoding gradient pulse sub-sequence GD.sub.2.

[0091] In step 2.IIa first diffusion-encoded image data DBD.sub.1 are then reconstructed on the basis of the acquired first diffusion-weighted raw data DRD.sub.1. Conventional methods, such as a Fourier transform of the raw data in the image data space or the like, can be applied for this purpose. In step 2.IIb second image data DBD.sub.2 are additionally reconstructed on the basis of the acquired second diffusion-weighted raw data DRD.sub.2.

[0092] The image data are then compared image point-by-image point in step 2.111. In other words, the signal intensities S.sub.1 of the first image data DBD.sub.1 are compared with the signal intensities S.sub.2 of the second image data DBD.sub.2 for each image point. The comparison can include, for example, the formation of a quotient Q=S.sub.1/S.sub.2 or a difference D=S.sub.1 S.sub.2. The extent of the existing anomaly or deviation from a Gaussian diffusion can then be determined, for example, using the extent of the deviation of the quotient Q from the value 1 or the deviation of the difference D from the value 0.

[0093] Finally, the results of the comparison are shown on a graph in step 2.IV. For example, regions in which a deviation or an anomaly exceeds a minimum are marked in color or with contours against the background of an image recorded without diffusion weighting. A threshold value SW, for example, can be defined for this purpose. The overlaying of masked or unmasked anomaly maps and clinical routine or diffusion images makes anatomical orientation possible for the user.

[0094] FIG. 3 shows images of this kind, also called “anomaly maps”. An anomaly map is overlaid on an anatomic image in a left drawing AK1 of FIG. 3. Regions Δ.sub.1, Δ.sub.2, Δ.sub.3 with different hatching have varying degrees of anomalies, wherein denser hatching signifies a more pronounced anomaly. In the scenario shown in the left drawing AK1, regions Δ.sub.3 in the brain and in a portion of the neck have the most pronounced anomaly. A middle drawing AK2 of FIG. 3 shows the anomaly map by masking of sections Δ.sub.3 (shown in hatched lines) in which a threshold value SW of the anomaly is exceeded. Instead of planar marking, as is shown in the left drawing AK1 and the middle component drawing AK2 of FIG. 3, regions Δ.sub.3 with an anomaly extent above the threshold value can also be displayed as a contour in order to best illustrate the anatomical information of the clinical routine images in the relevant regions. The right drawing AK3 of FIG. 3 illustrates a procedure of this kind. An interactive adjustment of the threshold value allows the operator to match the display to his personal requirements. Pre-settings for tissue-specific, region-specific or pathology-specific threshold values can be taken from a database and used. Expedient threshold values can be determined from repeated scans with identical parameters: noise-induced (stochastic) variations can be separated from systematic diffusion anomalies thereby.

[0095] FIG. 4 is a flowchart 400 that illustrates a method for determining a deviation of diffusion-weighted magnetic resonance image data of an examination object from normal diffusion behavior according to a second exemplary embodiment of the invention. In the method illustrated in FIG. 4, which is based on the underlying structure of the method illustrated in FIG. 2 according to a first exemplary embodiment of the invention, track-weighted images each with three single images are generated instead of the simply diffusion-weighted images with a defined direction of the diffusion gradient. Analogously to step 2.Ia, first diffusion-weighted raw data DRD.sub.1 are first acquired for this purpose in step 4.Ia, now albeit with three-dimensional diffusion weighting. In other words, the diffusion gradients are now preferably switched sequential, not just in the slice direction or z direction, but also in the in x and y directions, so that the diffusion movement is acquired along each of the three spatial dimensions. The diffusion-encoding gradient pulse sub-sequence GD.sub.1 switched during recording of the first diffusion-weighted raw data DRD.sub.1 is played with first time interval parameter values δ.sub.1, Δ.sub.1. A first b-value b1 correlates with the parameter values. Furthermore, second diffusion-weighted raw data DRD.sub.2 is likewise acquired in step 4.Ib with a diffusion weighting in three different directions. The diffusion-encoding gradient pulse sub-sequence switched during recording of the second diffusion-weighted raw data DRD.sub.2 is played with second time interval parameter values δ.sub.2, Δ.sub.2. A second b-value b2 correlates with the second time interval parameter values δ.sub.2, Δ.sub.2, and this approximately matches the first b-value b1 in the exemplary embodiment illustrated in FIG. 4.

[0096] Single images, i.e. a total of three diffusion-weighted single images, are then reconstructed in each case in steps 4.IIa, 4.IIb, 4.IIc based on first to third diffusion-weighted image data DBD.sub.1a, DBD.sub.1b, DBD.sub.1e on the basis of the first diffusion-weighted DRD.sub.1. The first diffusion-weighted image data DBD.sub.1a represents the diffusion behavior in the z direction, the second diffusion-weighted image data DBD.sub.1b the diffusion behavior in the x direction and the third diffusion-weighted image data DBD.sub.1e the diffusion behavior in the y direction. Furthermore, single images, i.e. a total of three diffusion-weighted single images are then reconstructed in each case in steps 4.IId, 4.IIe, 4.IIf based on fourth to sixth diffusion-weighted image data DBD.sub.2a, DBD.sub.2b, DBD.sub.2c on the basis of the second diffusion-weighted raw data DRD.sub.2. The fourth diffusion-weighted image data DBD.sub.2a represents the diffusion behavior in the z direction, the fifth diffusion-weighted image data DBD.sub.2b the diffusion behavior in the x direction and the third diffusion-weighted image data DBD.sub.2e the diffusion behavior in the y direction.

[0097] First track-weighted, diffusion-weighted image data SDBD.sub.1 is then generated in step 4.IIIa on the basis of the first to third diffusion-weighted image data DBD.sub.1a, DBD.sub.1b, DBD.sub.1c. If the first to third diffusion-weighted image data DBD.sub.1a, DBD.sub.1b, DBD.sub.1c match first to third image signal intensities S.sub.n(b1), the track-weighted image signal intensities S.sub.Tr(b1) for the first track-weighted, diffusion-weighted image data SDBD.sub.1 result as

[00002]$\begin{array}{cc}{S}_{\mathrm{Tr}}\left(b.Math..Math.1\right)=S\left(b=0\right).Math.e^{\frac{1}{N}.Math..Math.S_n\left(b.Math..Math.1\right)}=S\left(b=0\right).Math.e^{\frac{1}{N}.Math..Math.-b.Math..Math.1.Math.\mathrm{Tr}\left(\mathrm{DT}\right)}.& \left(2\right)\end{array}$

[0098] The signal intensity S(b=0) matches the intensity for the case where no diffusion gradient is switched. The designation DT stands in equation 2 for the diffusion tensor DT which contains information about the diffusion behavior at a location r in the space. Analogously to step 4.IIIa, second track-weighted, diffusion-weighted image data SDBD.sub.2 is determined in step 4.IIIb on the basis of the fourth to sixth diffusion-weighted image data DBD.sub.2a, DBD.sub.2b, DBD.sub.2c. If the fourth to sixth diffusion-weighted image data DBD.sub.2a, DBD.sub.2b, DBD.sub.2c matches fourth to sixth image signal intensities S.sub.n(b2), the track-weighted image signal intensities S.sub.Tr(b2) for the second track-weighted, diffusion-weighted image data SDBD.sub.2 results as

[00003]$\begin{array}{cc}{S}_{\mathrm{Tr}}\left(b.Math..Math.2\right)=S\left(b=0\right).Math.e^{\frac{1}{N}.Math..Math.S_n\left(b.Math..Math.2\right)}=S\left(b=0\right).Math.e^{\frac{1}{N}.Math..Math.-b.Math..Math.2.Math.\mathrm{Tr}\left(\mathrm{DT}\right)}.& \left(3\right)\end{array}$

[0099] Finally, the image data is compared pixel-by-pixel in step 4.IV on the basis of the determined track-weighted image intensities S.sub.Tr(b1), S.sub.Tr(b2). In other words, the signal intensities S.sub.Tr(b1) of the first track-weighted, diffusion-weighted image data SDBD.sub.1 are compared with the signal intensities S.sub.Tr(b2) of the second track-weighted, diffusion-weighted image data SDBD.sub.2. The comparison can include, for example, the formation of a quotient Q=S.sub.Tr(b1)/S.sub.Tr(b1) or a difference D=S.sub.Tr(b1) S.sub.Tr(b2). The extent of the existing anomaly or the deviation from a Gaussian diffusion can then be determined, for example, with the use of how much the quotient Q differs from the value 1 or the difference D from the value 0.

[0100] Finally, the results of the comparison are shown in a graph in step 4.V. For example, regions in which a deviation or an anomaly exceeds a minimum are marked in color or with contours against the background of an image recorded without diffusion weighting. A threshold value SW, for example, can be defined for this purpose. The overlaying of masked or unmasked anomaly maps and clinical routine or diffusion images makes anatomical orientation possible for the user.

[0101] FIG. 5 is a flowchart 500 that illustrates a method for determining a deviation of diffusion-weighted magnetic resonance image data of an examination object from normal diffusion behavior according to a third exemplary embodiment of the invention. The exemplary embodiment shown in FIG. 5 differs from the exemplary embodiment shown in FIG. 2 to the extent that, in addition to the first and second raw data DRD.sub.1, DRD.sub.2, reference raw data RRD are also acquired, which are not diffusion-weighted. This is necessary if diffusion-encoding gradient pulse sub-sequences, whose b-values b1, b2 are significantly different, are activated for the acquisition of the first and second diffusion-weighted raw data DRD.sub.1, DRD.sub.2. The method can also be applied, however, with virtually identical b-values. The first diffusion-weighted raw data DRD.sub.1 are then acquired in step 5.Ia analogously to the first exemplary embodiment with a first pulse sequence having a first diffusion-encoding gradient pulse sub-sequence GD.sub.1 with first time interval parameter values δ.sub.1, Δ.sub.1 and a first b-value b1 again. Second diffusion-weighted raw data DRD.sub.2 are additionally acquired in step 5.Ib with a second pulse sequence having a second diffusion-encoding gradient pulse sub-sequence GD.sub.2 and second time interval parameter values δ.sub.2, Δ.sub.2, which differ quantitatively from the first time interval parameter values δ.sub.1, Δ.sub.1 at least in respect of one of said parameters δ, Δ. For example, δ.sub.2>δ.sub.1. A second b-value b2 can be calculated from the second parameter values δ.sub.2, Δ.sub.2. In contrast to the first and second exemplary embodiments, the two b-values b1, b2 can then differ, as was discussed in connection with FIG. 2 and FIG. 4.

[0102] Raw data RRD without diffusion weighting, i.e. for b=0 or G=0, is also acquired in step 5.Ic with a reference scanning sequence RMS.

[0103] First diffusion-encoded image data DBD.sub.1Bei are then reconstructed in step 5.IIa on the basis of the acquired first diffusion-weighted raw data DRD.sub.1. Second image data DBD.sub.2 are additionally reconstructed in a step 5.IIb on the basis of the acquired second diffusion-weighted raw data DRD.sub.2. Furthermore, reference image data RBD are determined in step 5.IIc without diffusion weighting on the basis of the raw data RRD.

[0104] First diffusion coefficients ADC.sub.1 are then determined in step 5.IIIa on the basis of the first diffusion-weighted raw data DRD.sub.1 and the reference image data RBD. The first diffusion coefficients ADC.sub.1 result as follows:

[00004]$\begin{array}{cc}{\mathrm{ADC}}_1=-\frac{1}{b.Math..Math.1}.Math.\ln \left(\frac{S\left(b.Math..Math.1\right)}{S\left(b=0\right)}\right),& \left(4\right)\end{array}$

[0105] where S(b1) are the signal intensities of the first diffusion-weighted image data DBD.sub.1 and S(b=0) are the signal intensities of the reference image data RBD.

[0106] Furthermore, second diffusion coefficients ADC.sub.2 are determined in step 5.IIIb on the basis of the second diffusion-weighted raw data DRD.sub.2 and the reference image data RBD. The second diffusion coefficients ADC.sub.2 result as follows:

[00005]$\begin{array}{cc}{\mathrm{ADC}}_2=-\frac{1}{b.Math..Math.2}.Math.\ln \left(\frac{S\left(b.Math..Math.2\right)}{S\left(b=0\right)}\right),& \left(5\right)\end{array}$

[0107] where S(b2) represents the signal intensities of the second diffusion-weighted image data DBD.sub.2.

[0108] The first and second diffusion coefficients ADC.sub.1 and ADC.sub.2 are then compared image point-by-image point in step 5.IV. The comparison can in turn comprise, for example, the formation of a quotient Q=ADC.sub.1/ADC.sub.2 or a difference D=ADC.sub.1 ADC.sub.2. The extent of the existing anomaly or deviation from a Gaussian diffusion can then be determined, for example, using the extent of the deviation of the quotient Q from the value 1 or the deviation of the difference D from the value 0.

[0109] Finally, the results of the comparison are shown on a graph in step 5.V. For example, regions in which a deviation or an anomaly exceeds a minimum are marked in color or with contours against the background of an image recorded without diffusion weighting. A threshold value SW, for example, can be defined for this purpose. The overlaying of masked or unmasked anomaly maps and clinical routine or diffusion images makes anatomical orientation possible for the user.

[0110] FIG. 6 shows a flow diagram 600 which illustrates a method for determining a deviation of diffusion-weighted magnetic resonance image data of an examination object from normal diffusion behavior according to a fourth exemplary embodiment of the invention. The fourth exemplary embodiment differs from the previous exemplary embodiments to the extent that now more than two different interval durations are taken into account. Raw data relating to three approximately identical b-values b1, b2 and b3 are recorded in the example shown in FIG. 6.

[0111] First diffusion-weighted raw data DRD.sub.1 are acquired first of all in step 6.Ia with a first pulse sequence having a first diffusion-encoding gradient pulse sub-sequence GD.sub.1 with first time interval parameter values δ.sub.1, Δ.sub.1.

[0112] Second diffusion-weighted raw data DRD.sub.2 are additionally acquired in step 6.Ib with a second pulse sequence having a second diffusion-encoding gradient pulse sub-sequence GD.sub.2 and second time interval parameter values δ.sub.2, Δ.sub.2, which differ quantitatively from the first time interval parameter values δ.sub.1, Δ.sub.1 at least in respect of one of the parameters δ, Δ.

[0113] Third diffusion-weighted raw data DRD.sub.3 are then additionally acquired in step 6.Ic with a third diffusion-encoding gradient pulse sub-sequence GD.sub.3 having third time interval parameter values δ.sub.3, Δ.sub.3.

[0114] First to third diffusion-weighted image data DBD.sub.1, DBD.sub.2, DBD.sub.3 is then reconstructed in steps 6.IIa, 6.IIb and 6.IIc. In contrast to the previous exemplary embodiments, model-free statistical parameters are then used, however, in step 6.111 instead of the division images or subtraction images to describe the anomaly of the diffusion behavior. For example, a maximum difference D.sub.max=max(S.sub.k−S.sub.n) can be determined for k, n=1, 3 or a maximum quotient Q.sub.max=max(S.sub.k/S.sub.n) for k, n=1, . . . , 3 from the first to third signal intensities S.sub.1, S.sub.2, S.sub.3 of the reconstructed first to third diffusion-weighted image data DBD.sub.1, DBD.sub.2, DBD.sub.3, and be evaluated as a measure of the anomaly of the diffusion behavior in the region to be examined. As already mentioned, the standard deviation of the determined comparative values or the breadth of the distribution of the determined comparative values can also be used as alternative statistical parameters.

[0115] Finally, the results of the comparison are shown in a graph in step 6.IV as in conjunction with the other exemplary embodiments, using the statistical parameter, such as, for example, the maximum difference D.sub.max or the maximum quotient Q.sub.max.

[0116] FIG. 7 illustrates highly schematically an inventive magnetic resonance system 1 (hereinafter called “MR system” for short). It has the actual magnetic resonance scanner 2 having an examination space 3 or patient tunnel into which an examination object O, or here a patient or test person, in whose body the examination object, for example a specific organ, is located, can be moved on a couch 8.

[0117] The magnetic resonance scanner 2 is conventionally fitted with a basic field magnetic system 4, gradient system 6 and an RF transmitting antenna system 5 and an RF receiving antenna system 7. In the illustrated exemplary embodiment the RF transmitting antenna system 5 is a body coil permanently installed in the magnetic resonance scanner 2, whereas the RF receiving antenna system 7 has local coils that are to be arranged on the patient or test person (symbolized by just a single local coil in FIG. 7). Basically, however, the body coil can also be used as an RF receiving antenna system and the local coils as an RF transmitting antenna system if these coils can each be switched into different operating modes.

[0118] The MR system 1 also has a central controller 13 used for controlling the MR system 1. This central controller 13 has a sequence controller 14 for pulse sequence control. The sequence of radio-frequency pulses (RF pulses) and gradient pulses is controlled by the sequence controller 14 as a function of a chosen imaging sequence. An imaging sequence of this kind can be specified, for example, within a scanning or control protocol. Different control protocols for different scans are conventionally stored in a storage device 19 and can be chosen by an operator (and optionally changed, if required) and then be used to carry out the scan.

[0119] For emitting the individual RF pulses, the central controller 13 has a radio-frequency transmitter 15 that generates and amplifies the RF pulses and feeds them via a suitable interface (not shown) into the RF transmitting antenna system 5. The controller 13 has a gradient system interface 16 for controlling the gradient coils of the gradient system 6. The sequence controller 14 communicates appropriately, e.g. by emitting sequence control data SD, with the radio-frequency transmitter 15 and gradient system interface 16 in order to emit the pulse sequences. The controller 13 also has a radio-frequency receiver 17 (likewise communicating appropriately with the sequence controller 14), in order to acquire magnetic resonance signals, i.e. raw data, received in a coordinated manner from the RF transmitting antenna system 7. A reconstruction processor 18 takes over the acquired raw data and reconstructs the MR image data therefrom. This image data can then be stored, for example, in a memory 19 and/or be processed further in an image data processor 20 in order, for example, to generate other image data and/or parameter maps from different image data, which can in turn likewise be stored in the memory 19. This image data processor 20 has an image data interface 11 for taking over first image data DBD.sub.1 which were reconstructed by the reconstruction processor 18 on the basis of first raw data DRD.sub.1 acquired with a first gradient pulse sub-sequence GD.sub.1, and for taking over second image data DBD.sub.2 which was reconstructed on the basis of second raw data DRD.sub.2 acquired with a second gradient pulse sub-sequence GD.sub.2. In a comparison unit 12 image regions in which there is an anomaly of diffusion behavior are then determined—as already described—on the basis of the first image data DBD.sub.1 and second image data DBD.sub.2 image.

[0120] The anomaly map data resulting from the comparison can be output again, for example stored in the storage device 19, by way of the image data interface 11. Alternatively the image data processing unit 20, in particular the comparator 12, can also be integrated in the reconstruction processor 18 here or be connected externally over a network or the like to the central controller 13.

[0121] The central controller 13 can be controlled via a terminal having an input unit 10 and a display unit 9, at which terminal the entire MR system 1 can therefore also be operated by one operator. MR images can also be displayed on the display unit 9, and the input unit 10, optionally in combination with the display unit 9, scans can be planned and started and, in particular, suitable control protocols with suitable scanning sequences can be chosen, as described above, and optionally be modified.

[0122] The inventive MR system 1, and in particular the controller 13, can have a number of further components, not illustrated in detail herein but conventionally present in devices of this kind, such as, for example a network interface to connect the entire system to a network and to be able to exchange raw data and/or image data or parameter maps, but also further data such as, for example, patient-relevant data or control protocols.

[0123] Those skilled in the art knows how suitable raw data can be acquired and MR images reconstructed therefrom by radiating RF pulses and generating gradient fields, so this need not be described in more detail herein. Similarly, a wide variety of scanning sequences, such as, e.g. EPI scanning sequences or scanning sequences for generating diffusion-weighted images, are known in principle to those skilled in the art.

[0124] As mentioned, the method for determining a deviation of diffusion-weighted magnetic resonance image data of an examination object from normal diffusion behavior is not limited to diffusion-weighted imaging with the use of a Stejskal-Tanner sequence.

[0125] Furthermore, the described method is not limited to medical applications. Use of the term “unit” herein does not preclude the described item from being composed of components which can optionally also be spatially distributed.

[0126] Although modifications and changes may be suggested by those skilled in the art, it is the intention of the Applicant to embody within the patent warranted hereon all changes and modifications as reasonably and properly come within the scope of the Applicant's contribution to the art.