Time-resolved phase-contrast MR imaging with speed encoding
In a method and apparatus for time-resolved phase-contrast magnetic resonance (MR) imaging with speed encoding, MR signals are detected with multiple receivers in each of numerous time segments in order to acquire raw data in each of the time segments, in each case for numerous MR images with different speed encodings. Stationary image points and/or non-stationary image points are identified, dependent on the detected MR signals. A mask is defined, dependent on the identified stationary image points and/or the non-stationary image points, wherein the mask is locally variable. The numerous MR images for the numerous time segments are reconstructed from the acquired raw data, wherein the reconstruction occurs in an iterative process and with a temporal regularization, which is dependent on the mask.
1. A method for time-resolved phase-contrast magnetic resonance (MR) imaging with speed encoding, comprising:
operating an MR data acquisition unit comprising a plurality of radio-frequency (RF) receivers to acquire MR signals originating from stationary nuclear spins and non-stationary nuclear spins in an examination subject situated in the MR data acquisition unit, using multiple RF receivers among said plurality of RF receivers, in each of a plurality of time segments, with different speed encodings;
for each of said time segments, entering raw data, corresponding to the MR signals detected in the respective time segment, into an electronic memory as raw data points;
in a processor having access to said electronic memory, reconstructing intermediate images from said raw data, and automatically identifying in said intermediate images, stationary image points originating from said stationary nuclear spins and non-stationary image points originating from said non-stationary nuclear spins;
in said processor, defining a mask comprised of mask values that differ within said mask dependent on said stationary image points and said non-stationary image points; and
in said processor, reconstructing a plurality of final MR images respectively from the MR signals acquired in the respective segments by executing an iterative reconstruction algorithm with a temporal regularization that is dependent on said mask.
2. A method as claimed in claim 1 comprising defining and using said mask to suppress time-dependent changes in said stationary raw data points in the reconstruction of said MR images.
3. A method as claimed in claim 1 comprising, in said reconstruction algorithm, minimizing a target function that comprises a regularization term for said temporal regularization, and using said mask to provide a spatially-dependent weighting for said regularization term.
4. A method as claimed in claim 3 comprising calculating said regularization term as a difference between a first of said MR images and a second of said MR images respectively allocated to successive time segments among said plurality of time segments, and weighting said difference with said mask dependent on whether a raw data point is a stationary raw data point or a non-stationary raw data point.
5. A method as claimed in claim 4 wherein said target function is dependent on raw data acquired in respective, different time segments among said plurality of time segments.
6. A method as claimed in claim 3 , wherein the regularization term is a function of
∥{right arrow over (M)} p ·( {right arrow over (X)} s p −{right arrow over (X)} s p−1 )∥ L n
wherein p is an index for a time segment, s is an index for a speed encoding, {right arrow over (M)} p is the mask ( 40 , 45 ; 66 ) for a p th time segment, {right arrow over (X)} s p is a vector with image points of an MR image for the p th time segment, {right arrow over (X)} S p−1 is a vector with image points of an MR image for the (p−1) th time segment, and ∥·∥ L n indicates an L n norm.
7. A method as claimed in claim 1 comprising identifying said stationary raw data points and said non-stationary raw data points, and defining said mask, for each of multiple time segments among said plurality of time segments.
8. A method as claimed in claim 7 comprising, for each time segment, generating a phase-contrast image, as said intermediate image, for the respective time segment and identifying the stationary raw data points and the non-stationary raw data points as stationary image points and non-stationary image points respectively, in the respective phase-contrast image, and generating said mass from the respective phase-contrast image using a threshold value comparison.
9. A method as claimed in claim 1 comprising, in each of time segments, detecting said MR signals with a respectively different speed encodings that causes a raw data portion of said memory, reserved for the respective time segment, to be incompletely filled with raw data points.
10. A method as claimed in claim 9 comprising entering said MR signals into each region as raw data points according to a scanning pattern that is different for the respective different speed encodings.
11. A method as claimed in claim 10 comprising nesting said different scanning patterns in a central region of said memory, with each point in said central region being scanned according to at least one of said different scanning patterns.
12. A method as claimed in claim 10 comprising allocating the different scanning patterns to respective speed encodings in a time-dependent permutation.
13. A magnetic resonance (MR) apparatus for time-resolved phase-contrast MR imaging with speed encodings, comprising:
an MR data acquisition unit configured to receive an examination subject therein, and comprising a plurality of radio-frequency (RF) receivers;
a control unit configured to operate said MR data acquisition unit to acquire MR signals originating from stationary nuclear spins and non-stationary nuclear spins in an examination subject situated in the MR data acquisition unit, using multiple RF receivers among said plurality of RF receivers, in each of a plurality of time segments, with different speed encodings;
said control unit being configured, for each of said time segments, to enter raw data, corresponding to the MR signals detected in the respective time segment into an electronic memory as raw data points;
a processor having access to said electronic memory, said processor being configured to reconstruct intermediate images from said raw data and to automatically identify, in said intermediate images, stationary image points originating from said stationary nuclear spins and non-stationary image points originating from said non-stationary nuclear spins;
said processor being configured to define a mask comprised of mask values that differ within said mask dependent on said stationary image points and said non-stationary image points; and
said processor being configured to reconstruct a plurality of final MR images respectively from the MR signals acquired in the respective segments by executing an iterative reconstruction algorithm with a temporal regularization that is dependent on said mask.
14. A method for time-resolved magnetic resonance (MR) imaging with speed encoding comprising:
operating an MR data acquisition unit, comprising a plurality of radio-frequency (RF) receivers to acquire MR signals with multiple RF receivers among said plurality of RF receivers in each of a plurality of time segments by acquiring said MR signals, in each time segment, with a respectively different speed encoding;
operating said MR data acquisition unit to acquire said MR signals, in each time segment, with one of said different speed encodings that causes raw data, corresponding to the acquired MR signals, to be entered into a region of an electronic memory that incompletely fills said region, according to a scanning pattern, and using a different scanning pattern for each respective different speed encoding; and
in a processor, reconstructing a plurality of MR images from the raw data in said memory acquired during each of said time segments.
15. A method as claimed in claim 14 comprising nesting said different scanning patterns in a central region of said memory, with each point in said central region being scanned according to at least one of said different scanning patterns.
16. A method as claimed in claim 15 comprising selecting said different scanning patterns to cause a density of raw data points outside of said central region to be less than a density of raw data points within said central region.
17. A method as claimed in claim 15 comprising entering said raw data points into said central region in a coherent manner, and entering raw data points outside of said central region in an incoherent manner.
18. A method as claimed in claim 14 comprising allocating the different scanning patterns to respective speed encodings in a time-dependent permutation.
19. A magnetic resonance (MR) apparatus for time-resolved phase-contrast MR imaging with speed encodings, comprising:
an MR data acquisition unit configured to receive an examination subject therein, and comprising a plurality of radio-frequency (RF) receivers;
a control unit configured to operate to the MR data acquisition unit to acquire MR signals from the examination subject with multiple RF receivers among said plurality of RF receivers in each of a plurality of time segments by acquiring said MR signals, in each time segment, with a respectively different speed encoding;
said control unit being configured to operate said MR data acquisition unit to acquire said MR signals, in each time segment, with one of said different speed encodings that causes raw data, corresponding to the acquired MR signals, to be entered into a region of an electronic memory that incompletely fills said region, according to a scanning pattern, and using a different scanning pattern for each respective different speed encoding; and
a processor configured to reconstruct a plurality of MR images from the raw data in said memory acquired during each of said time segments.