WO2009112987A1 - Coil selection for parallel magnetic resonance imaging - Google Patents

Coil selection for parallel magnetic resonance imaging Download PDF

Info

Publication number
WO2009112987A1
WO2009112987A1 PCT/IB2009/050925 IB2009050925W WO2009112987A1 WO 2009112987 A1 WO2009112987 A1 WO 2009112987A1 IB 2009050925 W IB2009050925 W IB 2009050925W WO 2009112987 A1 WO2009112987 A1 WO 2009112987A1
Authority
WO
WIPO (PCT)
Prior art keywords
coil
physical
coil elements
matrix
elements
Prior art date
Application number
PCT/IB2009/050925
Other languages
French (fr)
Other versions
WO2009112987A8 (en
Inventor
Michel P. J. Jurrissen
Peter Boernert
Mariya I. Doneva
Miha Fuderer
Original Assignee
Koninklijke Philips Electronics N.V.
Philips Intellectual Property & Standards Gmbh
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Koninklijke Philips Electronics N.V., Philips Intellectual Property & Standards Gmbh filed Critical Koninklijke Philips Electronics N.V.
Priority to JP2010550301A priority Critical patent/JP5602030B2/en
Priority to CN200980109060.4A priority patent/CN101971045B/en
Priority to RU2010142030/28A priority patent/RU2504795C2/en
Priority to US12/920,118 priority patent/US8502535B2/en
Priority to EP09720031.5A priority patent/EP2255214B1/en
Publication of WO2009112987A1 publication Critical patent/WO2009112987A1/en
Publication of WO2009112987A8 publication Critical patent/WO2009112987A8/en

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R33/00—Arrangements or instruments for measuring magnetic variables
    • G01R33/02—Measuring direction or magnitude of magnetic fields or magnetic flux
    • G01R33/022—Measuring gradient
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R33/00—Arrangements or instruments for measuring magnetic variables
    • G01R33/20—Arrangements or instruments for measuring magnetic variables involving magnetic resonance
    • G01R33/28—Details of apparatus provided for in groups G01R33/44 - G01R33/64
    • G01R33/32—Excitation or detection systems, e.g. using radio frequency signals
    • G01R33/34—Constructional details, e.g. resonators, specially adapted to MR
    • G01R33/341—Constructional details, e.g. resonators, specially adapted to MR comprising surface coils
    • G01R33/3415—Constructional details, e.g. resonators, specially adapted to MR comprising surface coils comprising arrays of sub-coils, i.e. phased-array coils with flexible receiver channels
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R33/00—Arrangements or instruments for measuring magnetic variables
    • G01R33/20—Arrangements or instruments for measuring magnetic variables involving magnetic resonance
    • G01R33/28—Details of apparatus provided for in groups G01R33/44 - G01R33/64
    • G01R33/32—Excitation or detection systems, e.g. using radio frequency signals
    • G01R33/36—Electrical details, e.g. matching or coupling of the coil to the receiver
    • G01R33/3664—Switching for purposes other than coil coupling or decoupling, e.g. switching between a phased array mode and a quadrature mode, switching between surface coil modes of different geometrical shapes, switching from a whole body reception coil to a local reception coil or switching for automatic coil selection in moving table MR or for changing the field-of-view
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R33/00—Arrangements or instruments for measuring magnetic variables
    • G01R33/20—Arrangements or instruments for measuring magnetic variables involving magnetic resonance
    • G01R33/44—Arrangements or instruments for measuring magnetic variables involving magnetic resonance using nuclear magnetic resonance [NMR]
    • G01R33/48—NMR imaging systems
    • G01R33/54—Signal processing systems, e.g. using pulse sequences ; Generation or control of pulse sequences; Operator console
    • G01R33/56—Image enhancement or correction, e.g. subtraction or averaging techniques, e.g. improvement of signal-to-noise ratio and resolution
    • G01R33/561—Image enhancement or correction, e.g. subtraction or averaging techniques, e.g. improvement of signal-to-noise ratio and resolution by reduction of the scanning time, i.e. fast acquiring systems, e.g. using echo-planar pulse sequences
    • G01R33/5611—Parallel magnetic resonance imaging, e.g. sensitivity encoding [SENSE], simultaneous acquisition of spatial harmonics [SMASH], unaliasing by Fourier encoding of the overlaps using the temporal dimension [UNFOLD], k-t-broad-use linear acquisition speed-up technique [k-t-BLAST], k-t-SENSE
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R33/00—Arrangements or instruments for measuring magnetic variables
    • G01R33/20—Arrangements or instruments for measuring magnetic variables involving magnetic resonance
    • G01R33/44—Arrangements or instruments for measuring magnetic variables involving magnetic resonance using nuclear magnetic resonance [NMR]
    • G01R33/48—NMR imaging systems
    • G01R33/54—Signal processing systems, e.g. using pulse sequences ; Generation or control of pulse sequences; Operator console
    • G01R33/56—Image enhancement or correction, e.g. subtraction or averaging techniques, e.g. improvement of signal-to-noise ratio and resolution
    • G01R33/565—Correction of image distortions, e.g. due to magnetic field inhomogeneities
    • G01R33/5659—Correction of image distortions, e.g. due to magnetic field inhomogeneities caused by a distortion of the RF magnetic field, e.g. spatial inhomogeneities of the RF magnetic field
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R33/00—Arrangements or instruments for measuring magnetic variables
    • G01R33/20—Arrangements or instruments for measuring magnetic variables involving magnetic resonance
    • G01R33/44—Arrangements or instruments for measuring magnetic variables involving magnetic resonance using nuclear magnetic resonance [NMR]
    • G01R33/48—NMR imaging systems
    • G01R33/54—Signal processing systems, e.g. using pulse sequences ; Generation or control of pulse sequences; Operator console
    • G01R33/543—Control of the operation of the MR system, e.g. setting of acquisition parameters prior to or during MR data acquisition, dynamic shimming, use of one or more scout images for scan plane prescription

Definitions

  • the invention relates to a method of selecting a set of coil elements from a multitude of physical coil elements comprised in a coil array for performing a magnetic resonance imaging scan, a computer program product and a magnetic resonance imaging system.
  • US 7,279,893 discloses an image reconstruction technique which determines linear combinations of receiver channel information that contribute the most to image SNR to reduce the computational burden placed on the reconstruction engine of an MRI system and which quantifies the SNR loss for the reduced receiver channel set for image reconstruction. It is a goal of the invention to provide an improved method of selecting a set of coil elements from a multitude of physical coil elements comprised in a coil array for performing a magnetic resonance imaging scan, an improved computer program product and an improved magnetic resonance imaging system
  • the present invention provides a method of selecting a set of coil elements from a multitude of physical coil elements comprised in a coil array for performing a magnetic resonance imaging scan optimized for a region of interest, the method comprising determining a coil sensitivity matrix of the multitude of physical coil elements for the region of interest and determining a projection matrix, the projection matrix projecting the sensitivity matrix of the multitude of physical coil elements onto a set of virtual coil elements. Subsequently, the contribution of each physical coil element of the multitude of physical coil elements to the set of virtual coil elements is determined from the projection matrix. Then, from this contribution the most significantly contributing physical coil elements are selected as the set of coil elements.
  • the method according to the invention has the advantage, that in principle all possible subsets of coil element combinations are considered in order to select the most suitable set of coil elements for performing a magnetic resonance imaging scan of the region of interest. An exhaustive search through all possible subsets is not required since a SENSE (sensitivity encoding) reconstruction approach with Cartesian k-space sampling is performed and the projection matrix approach yields a recommendation for the optimal coil set to be used, which does not require a search for the most appropriate set of coil elements on a coil element by coil element basis.
  • the method according to the invention allows for an automatic coil selection, i.e. selecting only the important coils for signal sampling, which reduces the load on memory and reconstruction.
  • the algorithm underlying the method allows selecting corresponding coils automatically without any user interaction facilitating optimal selection for SENSE type imaging.
  • the algorithm is important for optimal workflow and applicable already during the scan planning phase selecting the best coil element automatically. Furthermore, it can be used during real-time interactive scanning, coping with changed scan conditions, maintaining the best system performance to optimal real time coil selection.
  • the projection matrix is determined by a factorization of the coil sensitivity matrix.
  • the factorization is performed by singular value decomposition (SVD).
  • SVD has the advantage that in a fast manner a matrix can be factorized exactly in such a way that one of the factors corresponds to a matrix from which the projection matrix can be mathematically deduced with low computational effort.
  • the contribution of the physical coil element is determined by calculating the total of all projection matrix element magnitudes receiving a contribution from the physical coil element. This has the advantage that instead of considering the effect of an individual physical coil element onto a received signal the total effect is considered which results from a combined effect of all physical coil elements.
  • the method further comprises determining a noise correlation matrix of the multitude of physical coil elements, wherein the projection matrix is determined from the coil sensitivity matrix and the noise correlation matrix.
  • the method further comprises determining additional contributions of each physical coil element of the multitude of physical coil elements to the set of virtual coil elements using the noise correlation matrix, wherein the selection of the most significantly contributing physical coil elements as the set of coil elements further comprises weighting the contribution of each physical coil element according to the additional contributions derived from the noise correlation matrix.
  • the noise correlation matrix into the determination procedure which set of coils to select has the advantage, that couplings between coil elements are considered. This further increases the accuracy of the method according to the invention thus reflecting more precisely the physical conditions in a real MRI system.
  • the additional contributions derived from the noise correlation matrix are calculated by means of a further projection matrix, wherein the further projection matrix is determined by factorization of the noise correlation matrix.
  • the coil sensitivity matrix is a reception coil sensitivity matrix or a transmission coil sensitivity matrix. In other words it is possible to apply the method according to the invention for transmission as well as for reception purposes. In both cases, the most appropriate set of coils out of a multitude of transmission or reception coils can be used to excite and/or refocus magnetization or to acquire data from a region of interest.
  • the idea of using parallel imaging to shorten the acquisition time by the simultaneous use of multiple receive coils can be adapted for the parallel transmission of spatially selective multi-dimensional RF pulses by using only the relevant transmit coils in a given experiment.
  • optimal transmission with a limited set of transmitters and/or RF amplifies can be realized or the dimensionality of the RF pulse design problem can be reduced, by skipping the RF pulse calculation for transmit elements which do not significantly contribute to the final result. The latter may result in faster calculation times due to the reduced problem size.
  • the invention in another aspect, relates to a computer program product comprising computer executable instructions to perform the method according to the invention.
  • the invention in another aspect, relates to a magnetic resonance imaging system, the system comprising a coil array adapted for performing a magnetic resonance imaging scan of a region of interest, wherein the coil array comprises a multitude of physical coil elements, the system comprising means for determining a coil sensitivity matrix of the multitude of physical coil elements for the region of interest, means for determining a projection matrix, the projection matrix projecting the sensitivity matrix of the multitude of physical coil elements onto a set of virtual coil elements, means for determining from the projection matrix the contribution of each physical coil element of the multitude of physical coil elements to the set of virtual coil elements and means for selecting from the contribution the most significantly contributing physical coil element as a set of coil elements applicable for performing a magnetic resonance imaging scan of the region of interest.
  • the magnetic resonance imaging scan further comprises means for factorization of the coil sensitivity matrix, for example by singular value decomposition.
  • the magnetic resonance imaging system further comprises means for determining a noise correlation matrix of the multitude of physical coil elements, wherein the means for determining the projection matrix are adapted to determine the projection matrix from the coil sensitivity matrix and the noise correlation matrix.
  • the magnetic resonance imaging system further comprises means for determining additional contributions for each physical coil element of the multitude of physical coil elements for the set of virtual coil elements using the noise correlation matrix.
  • the invention in another aspect, relates to a method of selecting a set of coil elements from a multitude of physical coil elements comprised in a coil array for performing a magnetic resonance imaging scan of a region of interest, wherein each physical coil element of the set of coil elements exhibits a coil sensitivity, wherein the method comprises determining a first estimated information content for the image resulting from magnetic resonance imaging scan of the region of interest, wherein the determination of that image is performed by combining the coil sensitivities of a first set of the physical coil elements.
  • the method further comprises determining a second estimated information content for the image resulting from a magnetic resonance imaging scan of the region of interest, wherein the determination is performed by combining the coil sensitivities of a second set of the physical coil elements, wherein in the second set of the physical coil elements at least one coil is skipped compared to the first set of physical coil elements. Further, the loss of information resulting from skipping the at least one coil is calculated by comparing the first and second information content. Finally, the previous steps are repeated and the set of coil elements is determined which yields the minimum loss of information.
  • the "information content” can be understood as the number of bits required to represent the relevant information of the image.
  • the "information content” of an MR image is regarded here as the sum of the information-contents (or “mutual information”) of each pixel.
  • “mutual information” measures the amount of information that can be obtained about the tissue at a given location of the object, by observing a noisy reconstruction thereof. It can be expressed as
  • the invention relates to a computer program product comprising computer executable instructions to perform said method of selecting a set of coil elements based on the information content analysis disclosed above.
  • the invention in yet another aspect relates to a magnetic resonance imaging system, the system comprising a coil array adapted for performing a magnetic resonance imaging scan of a region of interest, wherein the coil array comprises a multitude of physical coil elements, the system comprising means for determining a first estimated information content for an image resulting from the magnetic resonance imaging scan of a region of interest, wherein the determination is performed by combining coil sensitivities of a first set of the physical coil elements.
  • the system further comprises means for determining a second estimated information content for an image resulting from the magnetic resonance imaging scan of the region of interest, wherein the determination is performed by combining the coil sensitivities of a second set of the physical coil elements, wherein in the second set of the physical coil elements at least one coil is skipped compared to the first set of physical coil elements.
  • the system further comprises means for calculating the loss of information resulting from skipping the at least one coil by comparing the first and second information content and means for repeating said steps and determining the set of coil elements yielding the minimum loss of information.
  • Fig. 1 is a schematic illustrating an MRI system according to the invention
  • Fig. 2 is a flowchart illustrating the method of selecting a set of coil elements from a multitude of physical coil elements comprised in a coil array for performing a magnetic resonance imaging scan of a region of interest
  • Fig. 3 is a further flowchart illustrating a method of selecting a set of coil elements from a multitude of physical coil elements
  • Fig. 4 shows a comparison between SVD-based and SNR (signal to noise ratio) based coil selection algorithms.
  • Fig. 1 is a schematic illustrating an MRI system according to the invention. Only major components of a preferred MRI system which incorporate the present invention are shown in fig. 1.
  • the magnetic resonance imaging apparatus comprises a data processing system 100, wherein the data processing system 100 typically comprises a computer screen 102 and an input device 104. Such an input device could be for example a keyboard or a mouse.
  • the MRI system in fig. 1 further comprises a memory 106 and an interface
  • the interface 108 is adapted for communication and data exchange with typical hardware MRI components.
  • Typical hardware MRI components are for example a magnet 122 which generates a main magnetic field for performing a magnetic resonance imaging scan of an object 120, like for example a person.
  • gradient coils 124 are controlled by the data processing system 100, wherein the gradient coils are necessary and used for performing a 3- dimensional data acquisition.
  • the gradient coils 124 are preferably self- shielded gradient coils for producing gradients along three mutual axis x, y and z.
  • the object or person 120 is located on a table 118 which is movable with respect to a set of RF coils 130.
  • the set of RF coils 130 is connected to a control unit 126 which is adapted for selecting a subset of coils 130 from the multitude of the coils shown in fig. 1 and to map them to the limited number of RF channels.
  • the MRI system of fig. 1 may exemplary have only a set of four MR receiver channels, located in the receiver/transmitter unit 132, wherein a total of eight reception coils 130 are available.
  • a coil element selection has to be performed which refers to dimensionality reduction, in which an optimal subset of physical coils is selected for image reconstruction.
  • the selection is performed by the data processing system 100 which further comprises a processor which is adapted to execute computer executable instructions of the computer program product 112.
  • the data processing system 100 comprises a computer program product 112 by means of a data acquisition module 114 which is adapted to control hardware units 122, 124, 126 and 132. Also, the movement of the table 118 can be controlled by the data acquisition module 114.
  • the computer program product 112 comprises a coil selection module 116. Based on the information received for example during a pre-scan procedure from the set of coils 130 the coil selection module 116 performs a selection which physical coils to use for performing a data acquisition of the region of interest 128.
  • a similar procedure can also be performed to select a limited number of channels for transmission.
  • the data processing system 100 and the corresponding software selects the corresponding transmit coils which are driven by the transmitter unit 132, which contains schematically the corresponding RF amplifiers.
  • step 200 a pre-scan is performed which is followed by step 202, the generation of a coil sensitivity matrix of the multitude of physical coil elements based on the pre-scan data.
  • step 204 a projection matrix is determined, the projection matrix projecting the sensitivity matrix of the multitude of physical coil elements onto a set of virtual coil elements.
  • step 206 in which from the projection matrix the contribution of each physical coil element of the multitude of physical coil elements to the set of virtual coil elements is determined.
  • step 208 from the contribution the most significantly contributing physical coil elements are selected as the set of coil elements which is used in step 210 in order to perform a main magnetic resonance scan.
  • a noise correlation matrix which also can be obtained by performing the pre-scan in step 200 can be considered.
  • a noise correlation matrix is generated after having performed the pre-scanning step 200.
  • the noise correlation matrix generated in step 212 is thereupon used in step 204 for performing the singular value decomposition on the coil sensitivity matrix and the noise correlation matrix in order to determine the projection matrix.
  • SENSE reconstruction of uniformly undersampled Cartesian data consists of solving a linear system of L equations with R variables, where L is the number of coil elements, and R is the reduction factor.
  • the R unknowns could be uniquely determined if they were measured with a basis of R orthogonal coils. Solving this system of equations could be presented as projecting the problem from the L to the R dimensional space and then solving it in the lower dimensional space.
  • the vector a consists of the complex pixel intensities in the reduced field of view (FOV) images for each coil, S is the L x R sensitivity matrix, b is a vector containing the original image pixels, which are superimposed in the aliased images, and n is a noise vector.
  • the reconstruction consists of solving the linear system of Eqs. (1) for each aliased pixel in the reduced FOV. This system is overdetermined, thus an estimation technique has to be used for finding the solution. Assuming Gaussian statistics for the noise, the best linear unbiased estimator of vector b is given by:
  • ⁇ is the L x L receiver noise covariance matrix
  • the noise covariance matrix may be replaced by an identity matrix, leading to the least squares estimator for b. In this case, the unfolding is still ensured, yet at an SNR penalty, which will be the more significant, the less equivalent the receivers are with respect to load, gain, and mutual coupling.
  • matrix factorization techniques might be used to obtain the linear least squares solution in a numerically stable fashion. The most common matrix factorization technique used for solving an overdetermined system of linear equations is the singular value decomposition (SVD).
  • ⁇ + is the transposed of ⁇ with every value, greater than a preset threshold, replaced with its reciprocal.
  • the rows of the new sensitivity matrix S' can be considered as sensitivities of a set of virtual coils, which sensitivity vectors are obtained as a superposition of the real physical coil sensitivities.
  • the coils can be ranked according to their relative contribution to the virtual coils.
  • the i-th row vector of S' can be written as a linear combination of the rows of S
  • the magnitude of p ⁇ represents the contribution of the k-th physical coil to the i- th virtual coil. In this way, the coil array elements can be ranked according to the weighting function:
  • the weighting function can be calculated for each reconstructed pixel in the full FOV (field of view) image or for a given region of interest, and the optimal coil set is chosen according to the total weight for this region.
  • ⁇ M r) w k reROI
  • Fig. 3 illustrates a further flowchart of a further method of selecting a set of coil elements from a multitude of physical coil elements comprised in the coil array for performing a magnetic resonance imaging scan.
  • the method starts with step 300, performing a pre-scan in order to obtain data for generating a noise sensitivity matrix in step 302. From said noise sensitivity matrix the 'information content' can be calculated in step 304 for N coils.
  • step 306 wherein the information content is calculated for N-I coils.
  • the difference between the calculated information content for N coils and the calculated information content for N-I coils is the information loss which is experienced when discarding said one coil.
  • This information loss is calculated in step 308.
  • Step 304-308 are repeated until in step 310 it is decided that information loss is calculated for all possible coil combinations and skipped coils. In case in step 310 all coils have been considered, in step 312 the coil with minimal information loss is discarded.
  • N acq acquired channels which need to be reduced to N rec elements, due to reconstruction performance and/or memory limitations.
  • This reduction from N acq to N rec elements can be established by performing fast channel combinations (pre-combination) of profiles in time domain or by skipping channels.
  • Pi(x) a vector in the 3- dimensional volume of the scan.
  • This P 1 can be estimated from a coil survey scan preparation scan. This is a scan with the same geometry as the actual scan, with large voxels (about 20mm) and a scan time of about 1 second. This coil survey scan can be used to make a coarse estimate of the spatial SNR distribution.
  • These coarse SNR distributions per coil i are normalized according to:
  • the summation over x is a summation over all voxels in the coil survey scan.
  • the 'average SNR' is different per scan protocol, FsNR(scan). This could be determined per scan, or a constant value of 50 may be sufficient.
  • the expected SNR P 1 (X) per coil- element i is
  • the SNRNrec could be maximized. But it is better to maximize the 'information content'.
  • This invention uses the 'information content' criterium, where Q NKC is maximized.
  • Fig. 4 shows a comparison between an SVD-based and SNR-based coil selection algorithm.
  • the SNR-based coil selection algorithm is an algorithm which considers signal to noise ratios obtained when combining coils on a coil-by-coil basis. This algorithm has the disadvantage of being slow since all kinds of coil element combinations need to be 'individually considered.
  • Fig. 4 shows the scatter plot of the mean SNR- factor for the pixels that were used for coil selection.
  • the mean SNR- factor is very well correlated between the two algorithms for 100 simulated data sets.
  • 100 different sets of computer generated coil sensitivity maps were used, while each set consisted of 32 coil elements.
  • the coil array elements were modeled as circular coils with 40 mm radius.
  • the coil element positions were randomly chosen within the area constrained by two coaxial cylinders, whereas their orientation was chosen tangentially to the cylinders. By using this approach simple wraparound coils were modeled.
  • Sensitivity maps were calculated using the Biot-Savart law for a 384 x 384 pixel matrix and FOV 384 x 384 mm 2 arranged as a plane, orthogonal to the cylinder axis.
  • the results of the simulation show that the SVD-based coil selection method gives comparable results to the SNR-based selection at greatly reduced computational complexity.
  • single coil set evaluation requires more calculations for the SNR- based coil selection than for the SVD-based selection, which leads to an even greater factor for the required computation times.
  • the computation time for a 384 x 384 matrix was evaluated on a Xion processor, CPU 2.4 GHz, 4 GByte memory.

Landscapes

  • Physics & Mathematics (AREA)
  • Condensed Matter Physics & Semiconductors (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • High Energy & Nuclear Physics (AREA)
  • Magnetic Resonance Imaging Apparatus (AREA)

Abstract

The invention relates to a method of selecting a set of coil elements from a multitude of physical coil elements comprised in a coil array for performing a magnetic resonance imaging scan of a region of interest.

Description

Coil selection for parallel magnetic resonance imaging
TECHNICAL FIELD
The invention relates to a method of selecting a set of coil elements from a multitude of physical coil elements comprised in a coil array for performing a magnetic resonance imaging scan, a computer program product and a magnetic resonance imaging system.
BACKGROUND AND RELATED ART
Parallel imaging using coil arrays with a large number of independent coil elements provides improved imaging performance and increased signal to noise ratio. Today, there is a fundamental trend to considerably increase the number of coil elements used for reception. Recently, 32-element coils have been introduced to boost SNR and to allow for higher reduction factors (Niendorf T., Hardy C. J., Giaquinto R.O., Gross P., Cline H. E., Zhu Y., Kenwood G., Cohen S., Grant A.K., Joshi S., Rofsky N.M., Sodickson D.K. (2006) Toward single breath-hold whole-heart coverage coronary MRA using highly accelerated parallel imaging with a 32-channel MR system. Magn Reson Med 56: 167-176).
Even more complex coil arrays have been proposed and realized, consisting of up to 128 individual elements. However, a number of these coil arrays have been used on systems supporting only 32 receive channels. The use of a large number of coil array elements can lead to memory storage problems and to increased reconstruction times. To overcome these problems, data reduction techniques could be applied.
These can be realized by a linear combination of the original coil data prior to reception, using an appropriate hardware combiner, data compression of the sample data before reconstruction, or by discarding particular data from coil elements with low signal content (either before or after reception). US 7,279,893 discloses an image reconstruction technique which determines linear combinations of receiver channel information that contribute the most to image SNR to reduce the computational burden placed on the reconstruction engine of an MRI system and which quantifies the SNR loss for the reduced receiver channel set for image reconstruction. It is a goal of the invention to provide an improved method of selecting a set of coil elements from a multitude of physical coil elements comprised in a coil array for performing a magnetic resonance imaging scan, an improved computer program product and an improved magnetic resonance imaging system
SUMMARY OF THE INVENTION
The present invention provides a method of selecting a set of coil elements from a multitude of physical coil elements comprised in a coil array for performing a magnetic resonance imaging scan optimized for a region of interest, the method comprising determining a coil sensitivity matrix of the multitude of physical coil elements for the region of interest and determining a projection matrix, the projection matrix projecting the sensitivity matrix of the multitude of physical coil elements onto a set of virtual coil elements. Subsequently, the contribution of each physical coil element of the multitude of physical coil elements to the set of virtual coil elements is determined from the projection matrix. Then, from this contribution the most significantly contributing physical coil elements are selected as the set of coil elements.
The method according to the invention has the advantage, that in principle all possible subsets of coil element combinations are considered in order to select the most suitable set of coil elements for performing a magnetic resonance imaging scan of the region of interest. An exhaustive search through all possible subsets is not required since a SENSE (sensitivity encoding) reconstruction approach with Cartesian k-space sampling is performed and the projection matrix approach yields a recommendation for the optimal coil set to be used, which does not require a search for the most appropriate set of coil elements on a coil element by coil element basis. The method according to the invention allows for an automatic coil selection, i.e. selecting only the important coils for signal sampling, which reduces the load on memory and reconstruction. The algorithm underlying the method allows selecting corresponding coils automatically without any user interaction facilitating optimal selection for SENSE type imaging. The algorithm is important for optimal workflow and applicable already during the scan planning phase selecting the best coil element automatically. Furthermore, it can be used during real-time interactive scanning, coping with changed scan conditions, maintaining the best system performance to optimal real time coil selection.
In accordance with an embodiment of the invention, the projection matrix is determined by a factorization of the coil sensitivity matrix. Preferably, the factorization is performed by singular value decomposition (SVD). SVD has the advantage that in a fast manner a matrix can be factorized exactly in such a way that one of the factors corresponds to a matrix from which the projection matrix can be mathematically deduced with low computational effort. In accordance with an embodiment of the invention, the contribution of the physical coil element is determined by calculating the total of all projection matrix element magnitudes receiving a contribution from the physical coil element. This has the advantage that instead of considering the effect of an individual physical coil element onto a received signal the total effect is considered which results from a combined effect of all physical coil elements. The combinatory effect of all physical coil elements contributing to a local signal corresponds to the real case when data acquisition is performed using multicoil techniques. In accordance with an embodiment of the invention, the contribution of the physical coil element is the averaged contribution of said physical coil element to all voxels in the region of interest. In accordance with a further embodiment of the invention, the method further comprises determining a noise correlation matrix of the multitude of physical coil elements, wherein the projection matrix is determined from the coil sensitivity matrix and the noise correlation matrix.
Alternatively, in a further embodiment of the invention the method further comprises determining additional contributions of each physical coil element of the multitude of physical coil elements to the set of virtual coil elements using the noise correlation matrix, wherein the selection of the most significantly contributing physical coil elements as the set of coil elements further comprises weighting the contribution of each physical coil element according to the additional contributions derived from the noise correlation matrix. Further including the noise correlation matrix into the determination procedure which set of coils to select has the advantage, that couplings between coil elements are considered. This further increases the accuracy of the method according to the invention thus reflecting more precisely the physical conditions in a real MRI system.
In accordance with an embodiment of the invention, the additional contributions derived from the noise correlation matrix are calculated by means of a further projection matrix, wherein the further projection matrix is determined by factorization of the noise correlation matrix. This has the same advantage as already discussed above with respect to the coil sensitivity matrix, namely a quick coil selection with low computational burden and high accuracy. In accordance with a further embodiment of the invention, the coil sensitivity matrix is a reception coil sensitivity matrix or a transmission coil sensitivity matrix. In other words it is possible to apply the method according to the invention for transmission as well as for reception purposes. In both cases, the most appropriate set of coils out of a multitude of transmission or reception coils can be used to excite and/or refocus magnetization or to acquire data from a region of interest. Thus, the idea of using parallel imaging to shorten the acquisition time by the simultaneous use of multiple receive coils can be adapted for the parallel transmission of spatially selective multi-dimensional RF pulses by using only the relevant transmit coils in a given experiment. In this way optimal transmission with a limited set of transmitters and/or RF amplifies can be realized or the dimensionality of the RF pulse design problem can be reduced, by skipping the RF pulse calculation for transmit elements which do not significantly contribute to the final result. The latter may result in faster calculation times due to the reduced problem size.
In another aspect, the invention relates to a computer program product comprising computer executable instructions to perform the method according to the invention.
In another aspect, the invention relates to a magnetic resonance imaging system, the system comprising a coil array adapted for performing a magnetic resonance imaging scan of a region of interest, wherein the coil array comprises a multitude of physical coil elements, the system comprising means for determining a coil sensitivity matrix of the multitude of physical coil elements for the region of interest, means for determining a projection matrix, the projection matrix projecting the sensitivity matrix of the multitude of physical coil elements onto a set of virtual coil elements, means for determining from the projection matrix the contribution of each physical coil element of the multitude of physical coil elements to the set of virtual coil elements and means for selecting from the contribution the most significantly contributing physical coil element as a set of coil elements applicable for performing a magnetic resonance imaging scan of the region of interest.
In accordance with an embodiment of the invention, the magnetic resonance imaging scan further comprises means for factorization of the coil sensitivity matrix, for example by singular value decomposition. Further, in a preferred embodiment of the invention the magnetic resonance imaging system further comprises means for determining a noise correlation matrix of the multitude of physical coil elements, wherein the means for determining the projection matrix are adapted to determine the projection matrix from the coil sensitivity matrix and the noise correlation matrix. In accordance with an embodiment of the invention, the magnetic resonance imaging system further comprises means for determining additional contributions for each physical coil element of the multitude of physical coil elements for the set of virtual coil elements using the noise correlation matrix. In another aspect, the invention relates to a method of selecting a set of coil elements from a multitude of physical coil elements comprised in a coil array for performing a magnetic resonance imaging scan of a region of interest, wherein each physical coil element of the set of coil elements exhibits a coil sensitivity, wherein the method comprises determining a first estimated information content for the image resulting from magnetic resonance imaging scan of the region of interest, wherein the determination of that image is performed by combining the coil sensitivities of a first set of the physical coil elements. The method further comprises determining a second estimated information content for the image resulting from a magnetic resonance imaging scan of the region of interest, wherein the determination is performed by combining the coil sensitivities of a second set of the physical coil elements, wherein in the second set of the physical coil elements at least one coil is skipped compared to the first set of physical coil elements. Further, the loss of information resulting from skipping the at least one coil is calculated by comparing the first and second information content. Finally, the previous steps are repeated and the set of coil elements is determined which yields the minimum loss of information. The "information content" can be understood as the number of bits required to represent the relevant information of the image. More precisely, the "information content" of an MR image is regarded here as the sum of the information-contents (or "mutual information") of each pixel. In this context, "mutual information" measures the amount of information that can be obtained about the tissue at a given location of the object, by observing a noisy reconstruction thereof. It can be expressed as
I(X; Y)=H(X)-H(XIY)
where I stands for mutual information, H for entropy, X for the true state of the tissue (in this context: the noiseless signal) and Y the reconstructed noisy signal. In essence, H(X) is the entropy of the signal and H(X|Y) is the entropy of the noise. In another aspect, the invention relates to a computer program product comprising computer executable instructions to perform said method of selecting a set of coil elements based on the information content analysis disclosed above.
In yet another aspect the invention relates to a magnetic resonance imaging system, the system comprising a coil array adapted for performing a magnetic resonance imaging scan of a region of interest, wherein the coil array comprises a multitude of physical coil elements, the system comprising means for determining a first estimated information content for an image resulting from the magnetic resonance imaging scan of a region of interest, wherein the determination is performed by combining coil sensitivities of a first set of the physical coil elements. The system further comprises means for determining a second estimated information content for an image resulting from the magnetic resonance imaging scan of the region of interest, wherein the determination is performed by combining the coil sensitivities of a second set of the physical coil elements, wherein in the second set of the physical coil elements at least one coil is skipped compared to the first set of physical coil elements. The system further comprises means for calculating the loss of information resulting from skipping the at least one coil by comparing the first and second information content and means for repeating said steps and determining the set of coil elements yielding the minimum loss of information.
BRIEF DESCRIPTION OF THE DRAWINGS
In the following preferred embodiments of the invention are described in greater detail by way of example only making reference to the drawings in which:
Fig. 1 is a schematic illustrating an MRI system according to the invention,
Fig. 2 is a flowchart illustrating the method of selecting a set of coil elements from a multitude of physical coil elements comprised in a coil array for performing a magnetic resonance imaging scan of a region of interest,
Fig. 3 is a further flowchart illustrating a method of selecting a set of coil elements from a multitude of physical coil elements,
Fig. 4 shows a comparison between SVD-based and SNR (signal to noise ratio) based coil selection algorithms.
DETAILED DESCRIPTION
Fig. 1 is a schematic illustrating an MRI system according to the invention. Only major components of a preferred MRI system which incorporate the present invention are shown in fig. 1. The magnetic resonance imaging apparatus comprises a data processing system 100, wherein the data processing system 100 typically comprises a computer screen 102 and an input device 104. Such an input device could be for example a keyboard or a mouse. The MRI system in fig. 1 further comprises a memory 106 and an interface
108. The interface 108 is adapted for communication and data exchange with typical hardware MRI components.
Typical hardware MRI components are for example a magnet 122 which generates a main magnetic field for performing a magnetic resonance imaging scan of an object 120, like for example a person. Further, gradient coils 124 are controlled by the data processing system 100, wherein the gradient coils are necessary and used for performing a 3- dimensional data acquisition. The gradient coils 124 are preferably self- shielded gradient coils for producing gradients along three mutual axis x, y and z.
The object or person 120 is located on a table 118 which is movable with respect to a set of RF coils 130. The set of RF coils 130 is connected to a control unit 126 which is adapted for selecting a subset of coils 130 from the multitude of the coils shown in fig. 1 and to map them to the limited number of RF channels. For example, the MRI system of fig. 1 may exemplary have only a set of four MR receiver channels, located in the receiver/transmitter unit 132, wherein a total of eight reception coils 130 are available. Thus, a coil element selection has to be performed which refers to dimensionality reduction, in which an optimal subset of physical coils is selected for image reconstruction.
The selection is performed by the data processing system 100 which further comprises a processor which is adapted to execute computer executable instructions of the computer program product 112. In the present embodiment, the data processing system 100 comprises a computer program product 112 by means of a data acquisition module 114 which is adapted to control hardware units 122, 124, 126 and 132. Also, the movement of the table 118 can be controlled by the data acquisition module 114. Further, the computer program product 112 comprises a coil selection module 116. Based on the information received for example during a pre-scan procedure from the set of coils 130 the coil selection module 116 performs a selection which physical coils to use for performing a data acquisition of the region of interest 128.
As can be clearly seen in the example of fig. 1, due to the relative position of the region of interest 128 and the set of coils 130 it would be appropriate to only select a subset of coils 130 for data acquisition which is located in close vicinity to said region of interest 128. Such a pre-selection can be already made without analyzing any reception signals from the coils 130 by just comparing the well known spatial positions of the coils 130 with respect to the actual position of the table 118. If for example a head examination of a person has to be performed the data processing system 100 always knows that in the example of fig. 1 only the very right arranged set of coils will be required for performing a respective MR scan. By pre-selecting a set of coils the computational burden for calculating an appropriate set of coils for performing an MR scan of the region of interest can be reduced since for the calculation procedure not all coils have to be taken into account.
A similar procedure can also be performed to select a limited number of channels for transmission. In analogy the data processing system 100 and the corresponding software selects the corresponding transmit coils which are driven by the transmitter unit 132, which contains schematically the corresponding RF amplifiers.
Referring to fig. 2, for selecting a set of coil elements for performing an MR scan of a region of interest, the steps depicted in fig. 2 are carried out. In step 200, a pre-scan is performed which is followed by step 202, the generation of a coil sensitivity matrix of the multitude of physical coil elements based on the pre-scan data. By performing a singular value decomposition in step 204, a projection matrix is determined, the projection matrix projecting the sensitivity matrix of the multitude of physical coil elements onto a set of virtual coil elements. This is followed by step 206, in which from the projection matrix the contribution of each physical coil element of the multitude of physical coil elements to the set of virtual coil elements is determined. In step 208, from the contribution the most significantly contributing physical coil elements are selected as the set of coil elements which is used in step 210 in order to perform a main magnetic resonance scan.
For more accurate results, a noise correlation matrix which also can be obtained by performing the pre-scan in step 200 can be considered. In one alternative embodiment, in step 212 a noise correlation matrix is generated after having performed the pre-scanning step 200. The noise correlation matrix generated in step 212 is thereupon used in step 204 for performing the singular value decomposition on the coil sensitivity matrix and the noise correlation matrix in order to determine the projection matrix. In the following, a detailed theoretical outline of the singular value decomposition based coil selection shall be given:
SENSE reconstruction of uniformly undersampled Cartesian data consists of solving a linear system of L equations with R variables, where L is the number of coil elements, and R is the reduction factor. The R unknowns could be uniquely determined if they were measured with a basis of R orthogonal coils. Solving this system of equations could be presented as projecting the problem from the L to the R dimensional space and then solving it in the lower dimensional space.
Consider a SENSE reconstruction problem with Cartesian k-space sampling at a reduction factor R in an arbitrary direction [3] using an L-element coil array. The reconstruction problem at each pixel is given by:
a = Sb + n (1)
The vector a consists of the complex pixel intensities in the reduced field of view (FOV) images for each coil, S is the L x R sensitivity matrix, b is a vector containing the original image pixels, which are superimposed in the aliased images, and n is a noise vector. The reconstruction consists of solving the linear system of Eqs. (1) for each aliased pixel in the reduced FOV. This system is overdetermined, thus an estimation technique has to be used for finding the solution. Assuming Gaussian statistics for the noise, the best linear unbiased estimator of vector b is given by:
Figure imgf000011_0001
where Ψ is the L x L receiver noise covariance matrix, which helps to optimize
SNR during the reconstruction. The noise covariance matrix may be replaced by an identity matrix, leading to the least squares estimator for b. In this case, the unfolding is still ensured, yet at an SNR penalty, which will be the more significant, the less equivalent the receivers are with respect to load, gain, and mutual coupling. Alternatively, matrix factorization techniques might be used to obtain the linear least squares solution in a numerically stable fashion. The most common matrix factorization technique used for solving an overdetermined system of linear equations is the singular value decomposition (SVD).
Using SVD, the sensitivity matrix S can be factorized as S = U∑VH, and the SVD solution of (1) is given by:
b = V∑+UHa, (3) where ∑+ is the transposed of ∑ with every value, greater than a preset threshold, replaced with its reciprocal.
Eq. (3) can be written in the form:
UHa = UHSb = ∑VHb (4)
considering only the rows of UH, ∑, and V corresponding to singular values greater than the threshold. The matrix UH projects both the data and the sensitivity matrix into lower dimensional space, corresponding to the rank of the matrix S. Let us denote this projection of the coil sensitivity matrix with S', the corresponding projection of the data vector with aO and the projection matrix, which transforms S to S' with P, where P = UH.
a' = UHa = Pa
S' = UHS = PS (5)
The rows of the new sensitivity matrix S' can be considered as sensitivities of a set of virtual coils, which sensitivity vectors are obtained as a superposition of the real physical coil sensitivities. The coils can be ranked according to their relative contribution to the virtual coils. The i-th row vector of S' can be written as a linear combination of the rows of S
s\ = ∑P*s* (6) k
where the weighting factors P(i, k) = p^ (k = 1, ..., nc) are the i-th row entries of matrix P. The magnitude of p^ represents the contribution of the k-th physical coil to the i- th virtual coil. In this way, the coil array elements can be ranked according to the weighting function:
^ =EkI (V)
The weighting function can be calculated for each reconstructed pixel in the full FOV (field of view) image or for a given region of interest, and the optimal coil set is chosen according to the total weight for this region. ∑Mr) wk = reROI
(8)
N ROI
In this way, a single step coil selection is performed using the singular value decomposition of the sensitivity matrix.
Fig. 3 illustrates a further flowchart of a further method of selecting a set of coil elements from a multitude of physical coil elements comprised in the coil array for performing a magnetic resonance imaging scan. The method starts with step 300, performing a pre-scan in order to obtain data for generating a noise sensitivity matrix in step 302. From said noise sensitivity matrix the 'information content' can be calculated in step 304 for N coils. This is followed by step 306 wherein the information content is calculated for N-I coils. The difference between the calculated information content for N coils and the calculated information content for N-I coils is the information loss which is experienced when discarding said one coil. This information loss is calculated in step 308. Step 304-308 are repeated until in step 310 it is decided that information loss is calculated for all possible coil combinations and skipped coils. In case in step 310 all coils have been considered, in step 312 the coil with minimal information loss is discarded.
In step 314 it is checked whether disregarding the coil will result in a number of physical coils which corresponds to a predetermined number of coils M which is desired to use to perform a final MR scan. If in step 314 there are still too many coils, i.e. N-1>M, in step 318 N=N-I is calculated, followed by a repetition of the steps 304-314. Each of said cycle reduces the number of coil elements by 1, until in step 314 the number of elements is reduced to M. Thereafter, in step 316 an MR scan is performed with said number of physical coil elements, the other coil elements being discarded due to the minimal information loss linked with discarding these coil elements.
In the following, this coil element selection method shall be described in more detail:
Assume we have Nacq acquired channels (coil elements), which need to be reduced to Nrec elements, due to reconstruction performance and/or memory limitations. This reduction from Nacq to Nrec elements can be established by performing fast channel combinations (pre-combination) of profiles in time domain or by skipping channels. To determine the optimal channel reduction, first an estimation must be made of the expected SNR from each channel, let this be Pi(x). Here is x a vector in the 3- dimensional volume of the scan. This P1 can be estimated from a coil survey scan preparation scan. This is a scan with the same geometry as the actual scan, with large voxels (about 20mm) and a scan time of about 1 second. This coil survey scan can be used to make a coarse estimate of the spatial SNR distribution. These coarse SNR distributions per coil i (PC1(X)) are normalized according to:
Figure imgf000014_0001
The summation over x is a summation over all voxels in the coil survey scan. The 'average SNR' is different per scan protocol, FsNR(scan). This could be determined per scan, or a constant value of 50 may be sufficient. Now the expected SNR P1(X) per coil- element i is
P1 (x) = FSNR (scan) ■ Pc, (x)
After combination of all channels an estimation of the SNR of the resulting image is given by
INacq <lNacq(x) = J ∑PΪ(x)
To determine which channels have to be selected for pre-combination and skipping, as criterium the SNRNrec could be maximized. But it is better to maximize the 'information content'. This invention uses the 'information content' criterium, where QNKC is maximized.
e«-ec =I>feLcM+i) Reducing to the number of elements to Nrec can be performed in many ways. The optimal method would calculate QNKC for every possible way to reduce the number of channels to Nrec. But this costs too much processing power.
A solution to this problem is given by reducing the number of channels by skipping a channel i. The quality is then reduced by
Figure imgf000015_0001
Now combine the channels or skip the channel that leads to a minimal information loss. It can be easily seen that this process can be repeated until the number of elements is reduced to Nrec.
Fig. 4 shows a comparison between an SVD-based and SNR-based coil selection algorithm. The SNR-based coil selection algorithm is an algorithm which considers signal to noise ratios obtained when combining coils on a coil-by-coil basis. This algorithm has the disadvantage of being slow since all kinds of coil element combinations need to be 'individually considered.
Fig. 4 shows the scatter plot of the mean SNR- factor for the pixels that were used for coil selection. The mean SNR- factor is very well correlated between the two algorithms for 100 simulated data sets. For simulation, 100 different sets of computer generated coil sensitivity maps were used, while each set consisted of 32 coil elements. The coil array elements were modeled as circular coils with 40 mm radius. The coil element positions were randomly chosen within the area constrained by two coaxial cylinders, whereas their orientation was chosen tangentially to the cylinders. By using this approach simple wraparound coils were modeled. Sensitivity maps were calculated using the Biot-Savart law for a 384 x 384 pixel matrix and FOV 384 x 384 mm2 arranged as a plane, orthogonal to the cylinder axis. The radii of the cylinders were chosen to be r = 271.5 and R = 300 mm such that all coils lie outside the FOV. In the z-direction, the coil positions vary between -200 and +200 mm from the considered FOV. The results of the simulation show that the SVD-based coil selection method gives comparable results to the SNR-based selection at greatly reduced computational complexity. Additionally, single coil set evaluation requires more calculations for the SNR- based coil selection than for the SVD-based selection, which leads to an even greater factor for the required computation times. The computation time for a 384 x 384 matrix was evaluated on a Xion processor, CPU 2.4 GHz, 4 GByte memory. The time for the SVD-based coil selection was t = 0.02s, and the time for the SNR-based coil selection was t = 16.3s, which is far too long for this clinical practice. These results clearly show that the SVD-based coil selection method according to the invention can be applied even in real time clinical MR studies.
List of reference numerals
100 data processing system
102 screen
104 input device
106 memory
108 interface
110 processor
112 computer program product
114 data acquisition module
116 coil selection module
118 movable table
120 patient
122 main coil
124 shim coils
126 selection module
128 region of interest
130 coils
132 receiver/transmitter unit

Claims

CLAIMS:
1. A method of selecting a set of coil elements from a multitude of physical coil elements (130) comprised in a coil array for performing a magnetic resonance imaging scan of a region of interest (128), the method comprising: determining a coil sensitivity matrix of the multitude of physical coil elements (130) for the region of interest (128), determining a projection matrix, the projection matrix projecting the sensitivity matrix of the multitude of physical coil elements onto a set of virtual coil elements, determining from the projection matrix the contribution of each physical coil element of the multitude of physical coil elements to the set of virtual coil elements, selecting from the contribution the most significantly contributing physical coil elements as the set of coil elements.
2. The method of claim 1 , wherein the projection matrix is determined by factorization of the coil sensitivity matrix.
3. The method of claim 2, wherein the factorization is performed by singular value decomposition.
4. The method of claim 1, wherein the contribution of a physical coil element
(130) is determined by calculating the total of all projection matrix element magnitudes receiving a contribution from the physical coil element.
5. The method of claim 4, wherein the contribution of the physical coil element (130) is the averaged contribution of said physical coil element to all voxels in the region of interest.
6. The method of claim 1, further comprising determining a noise correlation matrix of the multitude of physical coil elements, wherein the projection matrix is determined from the coil sensitivity matrix and the noise correlation matrix.
7. The method of claim 1 , further comprising determining additional contributions of each physical coil element of the multitude of physical coil elements to the set of virtual coil elements using a noise correlation matrix, wherein the selection of the most significantly contributing physical coil elements as the set of coil elements further comprises weighting the contribution of each physical coil element according to the additional contributions derived from the noise correlation matrix.
8. The method of claim 7, wherein the additional contributions are calculated by means of a further projection matrix, wherein the further projection matrix is determined by factorization of the noise correlation matrix.
9. The method of claim 1 , wherein the coil sensitivity matrix is a reception coil sensitivity matrix or a transmission coil sensitivity matrix.
10. A computer program product comprising computer executable instructions to perform the method steps as claimed in any of the previous claims 1 to 9.
11. A magnetic resonance imaging system, the system comprising a coil array (130) adapted for performing a magnetic resonance imaging scan of a region of interest (128), wherein the coil array comprises a multitude of physical coil elements (130), the system comprising: means (116) for determining a coil sensitivity matrix of the multitude of physical coil elements for the region of interest, means (116) for determining a projection matrix, the projection matrix projecting the sensitivity matrix of the multitude of physical coil elements onto a set of virtual coil elements, means (116) for determining from the projection matrix the contribution of each physical coil element of the multitude of physical coil elements to the set of virtual coil elements, means (116; 126) for selecting from the contribution the most significantly contributing physical coil elements as a set of coil elements applicable for performing a magnetic resonance imaging scan of the region of interest.
PCT/IB2009/050925 2008-03-14 2009-03-06 Coil selection for parallel magnetic resonance imaging WO2009112987A1 (en)

Priority Applications (5)

Application Number Priority Date Filing Date Title
JP2010550301A JP5602030B2 (en) 2008-03-14 2009-03-06 Coil selection for parallel magnetic resonance imaging
CN200980109060.4A CN101971045B (en) 2008-03-14 2009-03-06 Coil selection for parallel magnetic resonance imaging
RU2010142030/28A RU2504795C2 (en) 2008-03-14 2009-03-06 Selection of coils for parallel magnetic resonance imaging
US12/920,118 US8502535B2 (en) 2008-03-14 2009-03-06 Coil selection for parallel magnetic resonance imaging
EP09720031.5A EP2255214B1 (en) 2008-03-14 2009-03-06 Coil selection for parallel magnetic resonance imaging

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
EP08152772 2008-03-14
EP08152772.3 2008-03-14

Publications (2)

Publication Number Publication Date
WO2009112987A1 true WO2009112987A1 (en) 2009-09-17
WO2009112987A8 WO2009112987A8 (en) 2010-09-23

Family

ID=40613541

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/IB2009/050925 WO2009112987A1 (en) 2008-03-14 2009-03-06 Coil selection for parallel magnetic resonance imaging

Country Status (6)

Country Link
US (1) US8502535B2 (en)
EP (1) EP2255214B1 (en)
JP (1) JP5602030B2 (en)
CN (1) CN101971045B (en)
RU (1) RU2504795C2 (en)
WO (1) WO2009112987A1 (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102138793A (en) * 2010-01-29 2011-08-03 Ge医疗系统环球技术有限公司 Magnetic resonance imaging apparatus
CN103154761A (en) * 2010-08-20 2013-06-12 皇家飞利浦电子股份有限公司 Virtual coil emulation in parallel transmission mri
US10802101B2 (en) 2018-05-31 2020-10-13 General Electric Company Method and systems for coil selection in magnetic resonance imaging to reduce phase wrap artifact
US10859646B2 (en) 2018-05-31 2020-12-08 General Electric Company Method and systems for coil selection in magnetic resonance imaging to reduce annefact artifact
US10859645B2 (en) 2018-05-31 2020-12-08 General Electric Company Method and systems for coil selection in magnetic resonance imaging
US10866292B2 (en) 2018-05-31 2020-12-15 General Electric Company Methods and systems for coil selection in magnetic resonance imaging

Families Citing this family (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102654568A (en) * 2011-03-01 2012-09-05 西门子公司 Method and device for establishing excitation parameters for mr imaging
DE102011083406B4 (en) * 2011-09-26 2013-05-23 Siemens Ag Method for selecting a sub-scanning scheme for MR imaging, method for magnetic resonance imaging and magnetic resonance system
DE102012221465B4 (en) * 2012-11-23 2014-06-26 Siemens Aktiengesellschaft Dynamic adjustment of signal compression level for MRI image processing
CN103698732B (en) * 2013-12-12 2016-03-09 深圳先进技术研究院 Magnetic-resonance radio-frequency coil performance evaluation method and system
DE102014207236A1 (en) * 2014-04-15 2015-10-15 Siemens Aktiengesellschaft Method and control device for creating magnetic resonance recordings
WO2015164606A1 (en) * 2014-04-24 2015-10-29 The General Hospital Corporation Hierarchical mapping framework for coil compression in magnetic resonance image reconstruction
CN105467342B (en) * 2014-05-30 2017-12-22 上海联影医疗科技有限公司 Magnetic resonance multichannel collecting image rebuilding method and device
CN106574952B (en) * 2014-08-14 2020-12-22 皇家飞利浦有限公司 Method and apparatus for hardware RF receiver channel reduction
JP6439173B2 (en) * 2015-05-28 2018-12-19 日立オートモティブシステムズ株式会社 Balancer device for internal combustion engine
US10408906B2 (en) * 2016-06-22 2019-09-10 Comsats Institute Of Information Technology GPU based implementation of SENSE (a parallel MRI algorithm) using QR decomposition
CN106872921B (en) * 2017-01-22 2019-08-16 上海东软医疗科技有限公司 MR imaging method and device
CN108627783B (en) 2017-03-23 2022-01-14 通用电气公司 Radio frequency coil array and magnetic resonance imaging transmitting array
CN108344960B (en) * 2018-02-09 2020-09-25 上海东软医疗科技有限公司 Coil unit gating method and device for magnetic resonance imaging system
CN108663640B (en) * 2018-05-09 2021-07-13 上海东软医疗科技有限公司 Method and device for determining position of magnetic resonance coil
CN109188326B (en) * 2018-09-29 2021-04-06 上海联影医疗科技股份有限公司 Magnetic resonance imaging method and magnetic resonance system
EP3644085A1 (en) * 2018-10-25 2020-04-29 Siemens Healthcare GmbH Selection of measuring coils in magnetic resonance imaging
EP3805772A1 (en) 2019-10-10 2021-04-14 Koninklijke Philips N.V. Magnetic resonance imaging receive antenna
CN110687490B (en) * 2019-10-10 2021-12-31 上海东软医疗科技有限公司 Parallel imaging method and device, storage medium and medical equipment
CN112834968A (en) * 2019-11-25 2021-05-25 通用电气精准医疗有限责任公司 Scanning control system and method for magnetic resonance imaging system
DE102021210162A1 (en) 2021-09-14 2023-03-16 Siemens Healthcare Gmbh Method for determining an optimized subset of coil elements from a plurality of coil elements for acquiring a magnetic resonance tomography image
EP4152029A1 (en) * 2021-09-15 2023-03-22 Siemens Healthcare GmbH Method and device for providing an antenna matrix for transmitting and / or receiving high frequency signals from a magnetic resonance recording
WO2024200017A1 (en) 2023-03-28 2024-10-03 Koninklijke Philips N.V. Method for acquiring a signal with overlapping receive coils for mr imaging

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070013375A1 (en) 2005-05-06 2007-01-18 James Akao Method and apparatus for adaptive channel reduction for parallel imaging
US7279893B1 (en) * 2006-04-20 2007-10-09 General Electric Company Receiver channel data combining in parallel mr imaging

Family Cites Families (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5138260A (en) 1990-11-21 1992-08-11 Picker International, Inc. Computer controlled switching of multiple rf coils
GB2360093A (en) * 2000-03-06 2001-09-12 Marconi Caswell Ltd NMR system with magnetic flux guide
US6717406B2 (en) * 2000-03-14 2004-04-06 Beth Israel Deaconess Medical Center, Inc. Parallel magnetic resonance imaging techniques using radiofrequency coil arrays
US6724923B2 (en) 2001-04-13 2004-04-20 Ge Medical Systems Global Technology Co., Llc Automatic coil selection of multi-receiver MR data using fast prescan data analysis
US6597173B1 (en) * 2002-07-10 2003-07-22 Ge Medical Systems Global Technology Co., Llc Method and apparatus for reconstructing zoom MR images
AU2003300665A1 (en) * 2003-01-20 2004-08-13 Koninklijke Philips Electronics N.V. Sub-sampled moving table mri
WO2005047914A1 (en) 2003-11-12 2005-05-26 Philips Intellectual Property & Standards Gmbh Parallel mr imaging method
JP2007511270A (en) * 2003-11-14 2007-05-10 コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ High-speed continuous moving bed magnetic resonance imaging with multiple stationary receiver coils for SENSE reconstruction
US7095229B2 (en) 2004-01-20 2006-08-22 Siemens Corporate Research Inc. Multicoil image analysis for magnetic resonance image segmentation, registration, and reconstruction
JP2006175058A (en) * 2004-12-22 2006-07-06 Ge Medical Systems Global Technology Co Llc Coil element selecting method and magnetic resonance imaging apparatus
CN100396240C (en) 2005-02-28 2008-06-25 西门子（中国）有限公司 MRI parallel imaging method utilizing sensitivity encoding in frequency domain
DE102006007769A1 (en) 2006-02-20 2007-08-30 Siemens Ag Patient`s magnetic resonance imaging producing method, involves selecting coil units based on result of pre-evaluation of magnetic resonance signals, and producing magnetic resonance imaging based on signals received from selected units
US8085046B2 (en) * 2008-08-28 2011-12-27 The General Hospital Corporation Coil array mode compression for parallel transmission magnetic resonance imaging
US8076938B2 (en) * 2009-03-31 2011-12-13 General Electric Company System and method of parallel imaging with calibration to a virtual coil

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070013375A1 (en) 2005-05-06 2007-01-18 James Akao Method and apparatus for adaptive channel reduction for parallel imaging
US7279893B1 (en) * 2006-04-20 2007-10-09 General Electric Company Receiver channel data combining in parallel mr imaging

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
BUEHRER MARTIN ET AL: "Array compression for MRI with large coil arrays", MAGNETIC RESONANCE IN MEDICINE, ACADEMIC PRESS, DULUTH, MN, US, vol. 57, no. 6, 1 June 2007 (2007-06-01), pages 1131 - 1139, XP007908560, ISSN: 0740-3194 *
MARIYA DONEVA ET AL: "Automatic coil selection for channel reduction in SENSE-based parallel imaging", MAGNETIC RESONANCE MATERIALS IN PHYSICS, BIOLOGY AND MEDICINE, CHAPMAN AND HALL, LONDON, GB, vol. 21, no. 3, 3 April 2008 (2008-04-03), pages 187 - 196, XP019596838, ISSN: 1352-8661 *
NIENDORF THORALF ET AL: "Toward single breath-hold whole-heart coverage coronary MRA using highly accelerated parallel imaging with a 32-channel MR system", MAGNETIC RESONANCE IN MEDICINE, ACADEMIC PRESS, DULUTH, MN, US, vol. 56, no. 1, 1 January 2006 (2006-01-01), pages 167 - 176, XP007908557, ISSN: 0740-3194, [retrieved on 20060605] *
PRUESSMANN K P ET AL: "SENSE: SENSITIVITY ENCODING FOR FAST MRI", MAGNETIC RESONANCE IN MEDICINE, ACADEMIC PRESS, DULUTH, MN, US, vol. 42, no. 5, 1 November 1999 (1999-11-01), pages 952 - 962, XP000866655, ISSN: 0740-3194 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102138793A (en) * 2010-01-29 2011-08-03 Ge医疗系统环球技术有限公司 Magnetic resonance imaging apparatus
CN103154761A (en) * 2010-08-20 2013-06-12 皇家飞利浦电子股份有限公司 Virtual coil emulation in parallel transmission mri
CN103154761B (en) * 2010-08-20 2015-11-25 皇家飞利浦电子股份有限公司 Virtual coil emulation in transmitted in parallel MRI
US9689945B2 (en) 2010-08-20 2017-06-27 Koninklijke Philips Electronics N.V. Virtual coil emulation in parallel transmission MRI
US10802101B2 (en) 2018-05-31 2020-10-13 General Electric Company Method and systems for coil selection in magnetic resonance imaging to reduce phase wrap artifact
US10859646B2 (en) 2018-05-31 2020-12-08 General Electric Company Method and systems for coil selection in magnetic resonance imaging to reduce annefact artifact
US10859645B2 (en) 2018-05-31 2020-12-08 General Electric Company Method and systems for coil selection in magnetic resonance imaging
US10866292B2 (en) 2018-05-31 2020-12-15 General Electric Company Methods and systems for coil selection in magnetic resonance imaging

Also Published As

Publication number Publication date
US8502535B2 (en) 2013-08-06
RU2504795C2 (en) 2014-01-20
JP2011513022A (en) 2011-04-28
EP2255214B1 (en) 2017-05-10
US20110006766A1 (en) 2011-01-13
CN101971045A (en) 2011-02-09
CN101971045B (en) 2014-04-02
WO2009112987A8 (en) 2010-09-23
EP2255214A1 (en) 2010-12-01
JP5602030B2 (en) 2014-10-08
RU2010142030A (en) 2012-04-20

Similar Documents

Publication Publication Date Title
US8502535B2 (en) Coil selection for parallel magnetic resonance imaging
US6160398A (en) Adaptive reconstruction of phased array NMR imagery
Huang et al. A software channel compression technique for faster reconstruction with many channels
US9396562B2 (en) MRI reconstruction with incoherent sampling and redundant haar wavelets
US9482732B2 (en) MRI reconstruction with motion-dependent regularization
RU2592039C2 (en) Formation of magnetic resonance images using multiple-point dixon method
US8427156B2 (en) Systems and methods for image reconstruction of sensitivity encoded MRI data
JP6640757B2 (en) Multi-shot magnetic resonance (MR) imaging system and method of operation
CN102736045B (en) Parallel transmission RF pulse design with local SAR constraints
US8581589B2 (en) System and method for variable mode-mixing in magnetic resonance imaging
US10429475B2 (en) Method for increasing signal-to-noise ratio in magnetic resonance imaging using per-voxel noise covariance regularization
US9229081B2 (en) Accelerated MRI with nonlinear spatial encoding gradients
WO2007023435A1 (en) Device and method for parallel magnetic resonance imaging
US20130207652A1 (en) System for Accelerated Magnetic Resonance Imaging Using Parallel Coils
US11125847B2 (en) System and method for out-of-view artifact suppression for magnetic resonance fingerprinting
CN102654568A (en) Method and device for establishing excitation parameters for mr imaging
CN107690587B (en) MRI with variable density sampling
US10386435B2 (en) Methods and systems for fast auto-calibrated reconstruction with random projection in parallel MRI
Doneva et al. Automatic coil selection for channel reduction in SENSE-based parallel imaging
Erdogmus et al. Image construction methods for phased array magnetic resonance imaging
Giovannetti et al. Coil sensitivity map-based filter for phased-array image reconstruction in Magnetic Resonance Imaging
US9360543B2 (en) System and method for parallel magnetic resonance image reconstruction using digital beamforming
Aggarwal et al. Optimal sampling in parallel magnetic resonance imaging
Yang et al. Virtual Coil Augmentation Technology for MR Coil Extrapolation via Deep Learning
Al Baqee ARelative COMPARISON OF NOISE PERFORMANCES BETWEEN DIFFERENT PMRI ALGORITHMS

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 200980109060.4

Country of ref document: CN

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 09720031

Country of ref document: EP

Kind code of ref document: A1

REEP Request for entry into the european phase

Ref document number: 2009720031

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 2009720031

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 2010550301

Country of ref document: JP

WWE Wipo information: entry into national phase

Ref document number: 12920118

Country of ref document: US

NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 6441/CHENP/2010

Country of ref document: IN

WWE Wipo information: entry into national phase

Ref document number: 2010142030

Country of ref document: RU