Magnetic resonance method and apparatus using atlas-based masking for quantitative susceptibility mapping
10732235 ยท 2020-08-04
Assignee
Inventors
- Kieran O'Brien (Highgate Hill, AU)
- Markus Barth (St Lucia, AU)
- Steffen Bollmann (Dutton Park, AU)
- Benedicte Marechal (Lausanne, CH)
Cpc classification
G01R33/5608
PHYSICS
G01R33/56
PHYSICS
International classification
G01V3/00
PHYSICS
G01R33/561
PHYSICS
G01R33/24
PHYSICS
Abstract
In a magnetic resonance method and apparatus, deficiencies in conventional masking in quantitative susceptibility mapping (QSM) are addressed by the inclusion of an additional step in the conventional QSM post-processing pipeline. In this additional step, atlas-based segmentation techniques, which have been developed for morphological applications such as T1w MPRAGE are used in order to provide the mask. This mask is then fed to the remainder of the QSM post-processing pipeline.
Claims
1. A method for quantitative susceptibility mapping (QSM), comprising: operating a magnetic resonance (MR) data acquisition scanner in order to execute a gradient echo MR data acquisition sequence, so as to acquire MR data from a current subject during an echo of said gradient echo MR data acquisition sequence using a radio-frequency (RF) reception coil combination of said MR data acquisition scanner; providing the acquired MR data to a processor and, in said processor, executing an MR image reconstruction algorithm in order to reconstruct an MR image of the current subject; in said processor, executing a QSM mapping algorithm in which said at least one MR image of the current subject is used as an input, said QSM mapping algorithm including a masking step in order to generate a mask that is used in subsequent steps of said QSM mapping algorithm in order to generate a magnetic susceptibility map that represents a spatial distribution of magnetic susceptibility values of tissue in the current subject; in said processor, executing said masking step by accessing an atlas comprising a plurality of MR images respectively for a plurality of subjects, wherein each of said MR images in said atlas was generated from a same echo in a gradient echo MR data acquisition sequence as said echo during which said MR data were acquired from said current subject, and combining said MR images from said atlas to produce said mask; and in said processor, generating a data file comprising said magnetic susceptibility map and making the data file available from the processor in electronic form.
2. A method as claimed in claim 1 comprising operating said MR data acquisition scanner in order to acquire said MR data from the brain of the current subject and, in said processor, executing said masking step to generate skull stripped data as said mask.
3. A method as claimed in claim 1 comprising: in said processor, executing said masking step by accessing, as said atlas, an atlas comprising a plurality of MR image sets respectively for a plurality of subjects, wherein each of said MR image sets comprises respective MR images of a respective subject reconstructed from MR data acquired from the respective subject during a plurality of echoes in a gradient echo MR data acquisition sequence; in said processor, executing said masking step by combining the respective MR image for each respective echo in the plurality of MR image data sets in order to form a normative representation of all of said subjects for each echo; and in said processor, executing said masking step by combining the respective normative representations of all subjects for each echo dependent on the echo time during which said MR data were acquired from said current subject.
4. A method as claimed in claim 3 comprising combining the respective normative representations of all subjects for each echo dependent on said echo during which said MR data were acquired from said current subject by identifying two echo times among said normative representations that are adjacent to the echo time for said current subject, and forming a weighted average of said two adjacent echo times.
5. A method as claimed in claim 3 comprising combining said normative representations of all subjects for each echo dependent on said echo during which said MR data were acquired from the current subject by identifying a closest echo, among said normative representations, to said echo during which said MR data were acquired from the current examination subject, and using said normative representation of all subjects for said closest echo as said combination.
6. A magnetic resonance (MR) apparatus for quantitative susceptibility mapping (QSM), comprising: a MR data acquisition scanner comprising a radio-frequency (RF) reception coil combination; a computer configured to operate MR data acquisition scanner in order to execute a gradient echo MR data acquisition sequence, so as to acquire MR data from a current subject during an echo of said gradient echo MR data acquisition sequence using said RF reception coil combination of said MR data acquisition scanner; a processor provided with the acquired MR data, said processor being configured to execute an MR image reconstruction algorithm in order to reconstruct an MR image of the current subject; said processor being further configured to execute a QSM mapping algorithm in which said at least one MR image of the current subject is used as an input, said QSM mapping algorithm including a masking step in order to generate a mask that is used in subsequent steps of said QSM mapping algorithm in order to generate a magnetic susceptibility map that represents a spatial distribution of magnetic susceptibility values of tissue in the current subject; said processor being configured to execute said masking step by accessing an atlas comprising a plurality of MR images respectively for a plurality of subjects, wherein each of said MR images in said atlas was generated from a same echo in a gradient echo MR data acquisition sequence as said echo during which said MR data were acquired from said current subject, and combining said MR images from said atlas to produce said mask; and said processor being configured to generate a data file comprising said magnetic susceptibility map and make the data file available from the processor in electronic form.
7. An apparatus as claimed in claim 6 wherein said computer is configured to operate said MR data acquisition scanner in order to acquire said MR data from the brain of the current subject, and said processor is configured to execute said masking step to generate skull stripped data as said mask.
8. An apparatus as claimed in claim 6 wherein: said processor is configured to execute said masking step by accessing, as said atlas, an atlas comprising a plurality of MR image sets respectively for a plurality of subjects, wherein each of said MR image sets comprises respective MR images of a respective subject reconstructed from MR data acquired from the respective subject during a plurality of echoes in a gradient echo MR data acquisition sequence; said processor is configured to execute said masking step by combining the respective MR image for each respective echo in the plurality of MR image data sets in order to form a normative representation of all of said subjects for each echo; and said processor is configured to execute said masking step by combining the respective normative representations of all subjects for each echo dependent on the echo time during which said MR data were acquired from said current subject.
9. An apparatus as claimed in claim 8 wherein said processor is configured to combine the respective normative representations of all subjects for each echo dependent on said echo during which said MR data were acquired from said current subject by identifying two echo times among said normative representations that are adjacent to the echo time for said current subject, and forming a weighted average of said two adjacent echo times.
10. An apparatus as claimed in claim 8 said processor is configured to combine said normative representations of all subjects for each echo dependent on said echo during which said MR data were acquired from the current subject by identifying a closest echo, among said normative representations, to said echo during which said MR data were acquired from the current examination subject, and using said normative representation of all subjects for said closest echo as said combination.
11. A non-transitory, computer-readable data storage medium encoded with programming instructions, said storage medium being loaded into a computer of a magnetic resonance (MR) apparatus comprising an MR data acquisition scanner having a radio-frequency (RF) reception coil combination, said programming instructions causing said computer to: operate said MR data acquisition scanner in order to execute a gradient echo MR data acquisition sequence, so as to acquire MR data from a current subject during an echo of said gradient echo MR data acquisition sequence using said RF reception coil combination of said MR data acquisition scanner; execute an MR image reconstruction algorithm in order to reconstruct an MR image of the current subject; execute a QSM mapping algorithm in which said at least one MR image of the current subject is used as an input, said QSM mapping algorithm including a masking step in order to generate a mask that is used in subsequent steps of said QSM mapping algorithm in order to generate a magnetic susceptibility map that represents a spatial distribution of magnetic susceptibility values of tissue in the current subject; execute said masking step by accessing an atlas comprising a plurality of MR images respectively for a plurality of subjects, wherein each of said MR images in said atlas was generated from a same echo in a gradient echo MR data acquisition sequence as said echo during which said MR data were acquired from said current subject, and combining said MR images from said atlas to produce said mask; and generate a data file comprising said magnetic susceptibility map and make the data file available from the processor in electronic form.
12. A data storage medium as claimed in claim 11 wherein said programming instructions cause said computer to operate said MR data acquisition scanner in order to acquire said MR data from the brain of the current subject, and execute said masking step to generate skull stripped data as said mask.
13. A data storage medium as claimed in claim 11 wherein said programming instructions cause said computer to: execute said masking step by accessing, as said atlas, an atlas comprising a plurality of MR image sets respectively for a plurality of subjects, wherein each of said MR image sets comprises respective MR images of a respective subject reconstructed from MR data acquired from the respective subject during a plurality of echoes in a gradient echo MR data acquisition sequence; execute said masking step by combining the respective MR image for each respective echo in the plurality of MR image data sets in order to form a normative representation of all of said subjects for each echo; and execute said masking step by combining the respective normative representations of all subjects for each echo dependent on the echo time during which said MR data were acquired from said current subject.
14. A data storage medium as claimed in claim 13 wherein said programming instructions cause said computer to combine the respective normative representations of all subjects for each echo dependent on said echo during which said MR data were acquired from said current subject by identifying two echo times among said normative representations that are adjacent to the echo time for said current subject, and form a weighted average of said two adjacent echo times.
15. A data storage medium as claimed in claim 13 wherein said programming instructions cause said computer to combine said normative representations of all subjects for each echo dependent on said echo during which said MR data were acquired from the current subject by identifying a closest echo, among said normative representations, to said echo during which said MR data were acquired from the current examination subject, and use said normative representation of all subjects for said closest echo as said combination.
16. A method as claimed in claim 1, wherein the processor is configured to execute the QSM mapping algorithm using, as the input of the at least one MR image of the current subject, a magnitude image.
17. A method as claimed in claim 16, wherein the processor is configured to step use the magnitude image as an input to a volumetric morphometric package that processes T1-weighted magnetization-prepared 180 degrees radio-frequency pulses and rapid gradient-echo (MPRAGE) images.
18. A method as claimed in claim 17, wherein the processor is configured to an generate the atlas as a QSM magnitude atlas template in registration with incoming data.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
(1)
(2)
(3)
(4)
(5)
DESCRIPTION OF THE PREFERRED EMBODIMENTS
(6) A magnetic resonance system 1 according to the invention is schematically shown in
(7) Basic components of the magnetic resonance scanner 2 are a basic field magnet 3, a gradient system 4 with magnetic field gradient coils to generate magnetic field gradients in the x-, y- and z-directions, and a whole-body radio-frequency (RF) antenna 5. The magnetic field gradient coils can be controlled independently of one another in the x-, y- and z-directions so that gradients can be applied in arbitrary logical spatial directions (for example in the slice-selection direction, in the phase coding direction or in the readout direction) via a predetermined combination, wherein these directions normally depend on the selected slice orientation. The transmission (radiation) of RF signals in order to induce of magnetic resonance signals in the examination subject O can take place via the whole-body antenna 5. The MR signals are received with a local coil 6, which can be composed of one or more individual reception coils. The local coil 6 can also be used to radiate the RF signals. All of these components are known in principle to those skilled in the art and therefore are only schematically shown in
(8) The components of the magnetic resonance scanner 2 are controlled by a control computer, which can be formed by a number of individual computers (which may be spatially separated and connected among one another via suitable cables or the like). This control computer 10 is connected via a terminal interface 17 with a terminal 20 via which an operator can control the entire system 1. In the present case, this terminal 20 (as a computer) is equipped with keyboard, one or more monitors and additional input devices (for example mouse or the like) so that a graphical user interface is provided to the operator.
(9) Among other things, the control computer 10 has a gradient controller 11 that can in turn have multiple sub-components. Via this gradient controller 11, the individual gradient coils are provided with control signals according to a gradient pulse sequence GS. These gradient pulses are radiated (activated) at precisely provided time positions and with a precisely predetermined time curve during a measurement.
(10) The control computer 10 also has a radio-frequency transmission unit 12 in order to feed electrical signals respectively representing radio-frequency pulses into the whole-body radio-frequency coil 5 (or the local coil 6) according to a predetermined radio-frequency pulse sequence RFS of the pulse sequence MS. The radio-frequency pulse sequence RFS includes excitation and/or refocusing pulses. The reception of the magnetic resonance signals then occurs with the use of the reception coils of the local coil 6, and the raw data RF received in this manner are read out and processed by an RF reception unit 13. The magnetic resonance signals are passed in digital form as raw data RF to a reconstruction computer 14, which reconstructs the image data BD from the raw data using the reconstruction algorithm described above, and stores the image data BD in a memory 16 and/or passes the image data BD via the interface 17 to the terminal 20 so that the operator can view the image. The image data BD can also be stored at other locations via a network NW and/or be displayed and evaluated.
(11) Control commands are transmitted via an interface 18 to other components of the magnetic resonance scanner 2 (such as the bed 7 or the basic field magnet 3, for example), and measurement values or other information are received.
(12) The gradient controller 11, the RF transmission unit 12 and the RF reception unit 13 are controlled, in a coordinated manner, by a measurement control processor 15. Via corresponding commands, this ensures that the desired gradient pulse sequences GS and radio-frequency pulse sequences RFS are emitted. Moreover, for this purpose it must be ensured that the magnetic resonance signals are read out by the reception coils of the local coil array 6 by the RF reception unit 13 at the appropriate point in time and are processed further. The measurement control processor 15 likewise controls the interface 18.
(13) As shown in
(14) In accordance with the invention, the magnitude image is fed to a volumetric morphometric package that is conventionally applied to T1w MPRAGE images. The customization of this pipeline in accordance with the invention requires the creation of QSM magnitude atlas template in registration with the incoming data, in order to provide the skull stripped data to the masking step.
(15) Two embodiments are disclosed below for creating the optimized GRE/QSM atlas. The primary parameter that influences signal and contrast in a GRE acquisition, such as signal loss in a T2*w image, occurs around the sinuses and ears, and is subject-dependent. In view of the fact that image acquisition occurs at a specified TE (echo time), in accordance with the invention a representative atlas is created from the scanning of normal volunteers from images acquired for a specific TE, as shown in
(16) As noted above, signal loss in a T2*w acquisition is dependent on the echo time during which the MR data are acquired. The manner of generating an atlas as illustrated in
(17)
(18) As shown in
(19) For Subject 1 shown in
(20) A further alternative is to generate a priori information from a subsequent segmentation of the T2*w data, similar to what is done for T1w MPRAGE images. This can then be used as an input to define significant edges to assist in the regularization of the dipole inversion, which means there would also be an output from the label map in
(21) The present invention also encompasses a non-transitory, computer-readable data storage medium encoded with programming instructions that, when loaded into the control computer 10 of the magnetic resonance system in accordance with the invention shown in FIG. 1, cause the system to operate in order to implement any or all embodiments of the method according to the invention, as described above.
(22) Although modifications and changes may be suggested by those skilled in the art, it is the intention of the Applicant to embody within the patent warranted hereon all changes and modifications as reasonably and properly come within the scope of the Applicant's contribution to the art.