Method and magnetic resonance apparatus for generating at least one combination image dataset

Abstract

In a method and apparatus for generating a magnetic resonance (MR) image MR data are acquired from a subject as datasets in parallel with multiple RF coils, with first parallel dataset being acquired with a first parameter set and at least one further parallel dataset being acquired with a second parameter set. A first intermediate image dataset and at least one further intermediate image dataset are reconstructed with at least one of (a) the first intermediate image dataset being reconstructed from said first parallel dataset using a calibration data item derived from said at least one further parameter set, and (b) said at least one further intermediate image dataset is reconstructed from said at least one further parallel dataset using a calibration data item derived from said first parameter set. A combination image dataset is generated by combining said first intermediate image dataset and said at least one further intermediate dataset.

Claims

1. A method for generating a magnetic resonance (MR) image with an MR scanner comprising a radio-frequency (RF) coil array comprising a plurality of individual RF coils, said method comprising: in a control computer, generating control signals to the MR scanner in order to acquire the MR data from a subject as datasets respectively acquired in parallel, from an examination volume of the subject, with multiple RF coils among said plurality of RF coils, by acquiring a first parallel dataset with at least two of said RF coils with said MR scanner operated according to a first parameter set and by acquiring at least one further parallel dataset with said at least two of said RF coils with said MR scanner operated according to a second parameter set, each of said first and second parallel datasets comprising acquired data that are less than a total amount of data required to reconstruct an image of the examination volume, thereby requiring reconstruction using a calibration item to complete said total amount of data; providing said control signals from said control computer to said MR scanner and thereby operating said MR scanner while said subject is situated in the scanner, so as to acquire said first parallel dataset from the subject with said at least two of said RF coils operated according to said first parameter set and so as to acquire said at least one further parallel dataset from the subject with said at least two of said RF coils operated according to said second parameter set, and thereby obtaining said first and second parallel datasets comprising acquired data that are less than said total amount of data required to reconstruct said image of the examination volume, and thereby requiring said reconstruction using said calibration item to complete said total amount of data; providing said first and second parallel datasets to a processor and, in said processor, reconstructing a first intermediate image dataset and at least one further intermediate image dataset with at least one of (a) said first intermediate image dataset being reconstructed from said first parallel dataset using a calibration data item derived from said at least one further parameter set, and (b) said at least one further intermediate image dataset is reconstructed from said at least one further parallel dataset using a calibration data item derived from said first parameter set; and in said processor, generating a combination image dataset by combining said first intermediate image dataset and said at least one further intermediate dataset, and making said combination image dataset available at an output of said processor in electronic form as a data file.

2. A method as claimed in claim 1 comprising combining said first intermediate image dataset and said at least one further intermediate dataset to form said combination image dataset by addition.

3. A method as claimed in claim 1 comprising combining said first intermediate image dataset and said at least one further intermediate dataset to form said combination image dataset by subtraction.

4. A method as claimed in claim 1 comprising combining said first intermediate image dataset and said at least one further intermediate dataset to form said combination image dataset by an optimization method.

5. A method as claimed in claim 1 comprising operating said MR scanner to acquire at least three of said datasets of the examination volume of the subject in parallel.

6. A method as claimed in claim 5 comprising combining said first intermediate image dataset and said at least one further intermediate dataset to form said combination image dataset by a regression calculation.

7. A method as claimed in claim 6 comprising employing a linear fit or an exponential fit in said regression calculation.

8. A method as claimed in claim 1 comprising combining said first intermediate image dataset and said at least one further intermediate dataset to form said combination image dataset by a regression calculation.

9. A method as claimed in claim 1 wherein said first parameter set and said at least one further parameter set differ only with respect to echo time.

10. A method as claimed in claim 1 comprising selecting said first parameter set and said at least one further parameter set to have respective echo times to cause MR signals respectively originating from water protons and from fat protons to be acquired in-phase in one of said datasets and out-of-phase in another of said datasets.

11. A method as claimed in claim 1 comprising operating said MR scanner according to a turbo spin echo sequence to acquire said first dataset and said at least one further dataset in parallel.

12. A method as claimed in claim 1 comprising employing a GRAPPA-based method to reconstruct each of said first and second intermediate image datasets.

13. A method as claimed in claim 1 comprising employing a Auto-SMASH-based method to reconstruct each of said first and second intermediate image datasets.

14. A method as claimed in claim 1 comprising, when reconstructing said first and second intermediate image datasets, using a plurality of calibration data items that form a calibration line in each of said first and second intermediate image datasets.

15. A method as claimed in claim 1 comprising respectively weighting said calibration data items when reconstructing said first and second intermediate datasets.

16. A method as claimed in claim 15 comprising weighting said calibration data items as a function of a parameter in the respective first parameter set and said at least one further parameter set.

17. A method as claimed in claim 16 wherein said parameter is echo time.

18. A magnetic resonance (MR) apparatus comprising: an MR scanner comprising a radio-frequency (RF) coil array comprising a plurality of individual RF coils; a computer that operates the MR scanner so as to acquire MR data from a subject as datasets respectively acquired in parallel, from an examination volume of the subject, with multiple RF coils among said plurality of RF coils, by acquiring a first parallel dataset with at least two of said RF coils with said MR scanner operated according to a first parameter set and acquiring at least one further parallel dataset with said at least two of said RF coils with said MR scanner operated according to a second parameter set, each of said first and second parallel datasets comprising acquired data that are less than a total amount of data required to reconstruct an image of the examination volume, thereby requiring reconstruction using a calibration item to complete said total amount of data; a processor provided with said first and second parallel datasets, said processor reconstructing a first intermediate image dataset and at least one further intermediate image dataset with at least one of (a) said first intermediate image dataset being reconstructed from said first parallel dataset using a calibration data item derived from said at least one further parameter set, and (b) said at least one further intermediate image dataset is reconstructed from said at least one further parallel dataset using a calibration data item derived from said first parameter set; and said processor generating a combination image dataset by combining said first intermediate image dataset and said at least one further intermediate dataset, and making said combination image dataset available at an output of said processor in electronic form as a data file.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 shows a magnetic resonance system.

(2) FIG. 2 shows a flow diagram (prior art).

(3) FIG. 3 shows a flow diagram.

(4) FIG. 4 shows a first sequence diagram.

(5) FIG. 5 shows a second sequence diagram.

(6) FIG. 6 shows a third sequence diagram.

(7) FIG. 7 is a flowchart for generating a combination image dataset.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

(8) FIG. 1 shows a magnetic resonance scanner 1 with a radio-frequency (RF) coil 2 and an RF coil array 3 with RF coils 4, 5, 6 and 7 and a control computer 8.

(9) The RF coil 2 is generally of the type known as a body coil. This is used to stimulate magnetization. In contrast, the RF coil array 3 is provided to read out the measuring signal. The RF coils 4, 5, 6 and 7 of the coil array 3 read out the measuring signal simultaneously.

(10) FIG. 2 shows a known method for generating a combination image dataset with the Dixon technique using PPA methods. For clarity the use of just two coils, for example the coils 4 and 5, is assumed. Naturally any number of coils can be used.

(11) The coils 4 and 5 are used to record a first parallel dataset 9 and a further, second parallel dataset 10. In this process the parallel dataset 9 is acquired first and then the parallel dataset 10. The raw datasets 11 and 12 are therefore recorded using the same, e.g. first, parameter dataset and the raw datasets 13 and 14 are recorded using another, e.g. second, parameter dataset. The raw datasets 11 and 13 were acquired for example with the coil 4 and the raw datasets 12 and 14 were acquired with the coil 5. The number of raw datasets of a parallel dataset is based on the number of coils in the coil array 3 here.

(12) The raw datasets here are shown in the k-space diagram. The axis 15 represents the kx direction and the axis 16 the ky direction. Each raw dataset 11, 12, 13, and 14 has recorded k-space lines 17 and at least one calibration line 18, generally referred to as an ACS line. K space lines 19 that have been omitted during measurement and have to be reconstructed are shown broken.

(13) For the raw datasets 12 and 14 the reference characters have been omitted for clarity. However the structure of the raw datasets 12 and 14 corresponds to that of the raw datasets 11 and 13, which were acquired using the same parameters apart from the echo time.

(14) When two coils are used as the coil array, only one acceleration factor R=2 can be obtained and therefore only every second k-space line 19 can be omitted. If there are more coils correspondingly higher acceleration factors are possible. The calibration lines 18 are located at a point in the k-space, at which an omitted k-space line 19 should be located and as it were replace this according to the scan system.

(15) The measured k-space lines 17 here form the reduced dataset. The respective raw dataset is formed together with the calibration line 18.

(16) A GRAPPA method is used to obtain a reconstructed dataset 20 from the raw datasets 11 and 13, the reconstructed dataset 20 consisting of measured k-space lines 17, calibration lines 18 and reconstructed k-space lines 21. A Fourier transform is then used to obtain an intermediate image dataset 22 from the reconstructed dataset 20.

(17) Further processing steps can of course also be performed such as baseline correction or zero filling.

(18) A reconstructed dataset 23 is obtained from the raw datasets 12 and 14 and from this an intermediate image dataset 24 in the same manner. The intermediate image datasets 22 and 24 are then added or subtracted or otherwise combined to form a combination image dataset 25, depending on whether fat or water suppression is to take place.

(19) The problem occurs that the combination image dataset 25 can contain flow, spoiling, and other artifacts.

(20) FIG. 3 shows a flow which reduces the formation of such artifacts. Identical reference characters to those in FIG. 2 show identical objects. The recording of the measuring signals operates in the same manner as described in relation to FIG. 2.

(21) In contrast to FIG. 2 an intermediate step is inserted so that calibration data of another parallel dataset is taken into account, resulting in the raw datasets 11, 12, 13 and 14.

(22) For example the calibration lines 18 of the raw datasets 11 and 13, which were recorded with the same coil but using different parameter sets, can be averaged to form a calibration line 18. The calibration lines 18 here can be included with different weightings. However at the end the raw datasets 11 and 13 have the same calibration lines 18. This can of course also be the case when there are a number of calibration lines 18, with calibration lines being averaged at the same point of the k-space in each instance.

(23) The calibration lines 18 of another parallel dataset can be taken into account in an alternative manner in that they are inserted into another raw dataset. Thus the calibration line 18 of the raw dataset 13 can be inserted as the calibration line 18 into the raw dataset 11 and the calibration line 18 of the raw dataset 11 can be inserted as the calibration line 18 into the raw dataset 13. In fact this is only meaningful if more than one k-space line is omitted, as in the present example a measured k-space line 17 is then replaced. However this procedure is followed for the purpose of illustration.

(24) Reconstructed datasets 20 and 23, the calibration lines 18 and also, depending on the way in which external calibration lines are taken into account, calibration lines 18 are obtained from the raw datasets 11, 12, 13 and 14 thus modified.

(25) Either an averaging of calibration lines or an inclusion of the external calibration lines is always performed, in other words not a mixture as shown in FIG. 3. FIG. 3 is only intended to illustrate all the options.

(26) The intermediate image datasets 22 and 24 and the combination image dataset 25 then result from the reconstructed datasets 20 and 23.

(27) FIG. 4 shows a first sequence diagram for the recording of a parallel dataset, for example of the parallel dataset 9. High-frequency pulses and signals are plotted on the axis 26 and the read gradient is plotted on the axis 27. Phase encode and slice gradients are not shown.

(28) The sequence is configured as a TSE. In addition to an excitation pulse 28 there are therefore also two refocusing pulses 29. These produce two echo signals 30. While the echo signals 30 are being read out, the read gradients 31 are applied, while the gradient 32 is what is known as a preread gradient.

(29) The phase encode gradient is transferred between the refocusing pulses 29 so that the echoes 30 cover different k-space lines.

(30) The wait times are selected in such a manner between excitation pulse 28 and refocusing pulse 29 as well as refocusing pulse 29 and the echo maximum that the water and fat spins are in-phase at echo time TE.

(31) FIG. 5 shows a second sequence diagram for the recording of a parallel dataset, for example of the parallel dataset 10. The only difference compared with the sequence diagram shown in FIG. 4 is the selection of the echo time between excitation pulse 28 and the first echo 34, so the first echo signal 34 or at the time of the echo maxima the water and fat spins are out-of-phase. The echo time is therefore marked as TE2. TE2 is shorter than TE.

(32) FIG. 6 shows a third sequence diagram for the recording of two parallel datasets, e.g. of the parallel datasets 9 and 10. As an extension of FIG. 4 a bipolar read gradient 35 is applied so that after a first echo 30, in which the water and fat spins are out-of-phase, a second echo 34 is generated, in which the water and fat spins are in-phase. The in-phase echo here is located centrally between the refocusing pulses 29 and the out-of-phase echo is before or after. The recording of further k-space lines is suggested by the second refocusing pulse 29.

(33) A summary of FIGS. 4 to 6 shows that the parallel datasets can be recorded one after the other or interleaved. Even though a frequency encoding is shown in FIGS. 4 to 6, it is also possible to use spectroscopic imaging methods which do not use frequency encoding when the signal is read out.

(34) FIG. 7 is a flowchart for generating a combination image dataset.

(35) In step S1 a first parallel dataset is recorded (acquired), using a first parameter set. In step S2 a second or any number of further parallel datasets is/are recorded. Steps S1 and S2 can take place after one another or interleaved, for example as described in relation to FIGS. 4 to 6. In principle any known method can be used to generate parallel datasets, in other words any parallel imaging method in which calibration data ACS is generated. In a preferred example a Fast Dixon technique is used to acquire parallel datasets 9 and 10.

(36) In step S3 raw datasets 11, 12, 13 and 14 are generated, with calibration data 18 being used, which is determined with the aid of calibration data from an external parallel image dataset of the inherent coil. In a raw dataset 11 for example the measured calibration line 18 is replaced by a calibration line 18, which results from combining the calibration lines 18 of the raw datasets 11 and 12, with both the raw dataset 11 and the raw dataset 12 having been acquired with the coil 4. It is thus possible to improve the image quality of the subsequently generated image datasets.

(37) In the following step S4 a reconstructed dataset 20 and 23 is created for each parallel dataset 9 and 10 using a GRAPPA-based method. A Fourier transform produces intermediate image datasets 22 and 24 from these in step S5.

(38) In step S6 the intermediate image datasets 22 and 24 are combined to form the combination image dataset 25. This takes place as an addition so that the combination image dataset 25 only represents a water signal or as a subtraction so that the combination image dataset is a pure fat image.

(39) Although modifications and changes may be suggested by those skilled in the art, it is the intention of the inventors to embody within the patent warranted hereon all changes and modifications as reasonably and properly come within the scope of their contribution to the art.