System and method for magnetic resonance fingerprinting with non-locally sequential sampling of k-space
A system and method is provided for acquisition of magnetic resonance fingerprinting (“MRF”) data that includes determining a non-locally sequential sampling pattern for a Cartesian grid of k-space, performing a series of sequence blocks using acquisition parameters that vary between sequence blocks to acquire MRF data from a subject using the Cartesian grid of k-space and the determined non-locally sequential sampling pattern, assembling the MRF data into a series of signal evolutions, comparing the series of signal evolutions to a dictionary of known signal evolutions to determine tissue properties of the subject, and generating a report indicating the tissue properties of the subject.
1. A method for magnetic resonance fingerprinting (MRF) data acquisition using a computer system comprising:
a) determining a non-locally sequential sampling pattern for a Cartesian grid of k-space;
b) performing a series of sequence blocks using acquisition parameters that vary between sequence blocks to acquire MRF data from a subject using the Cartesian grid of k-space and the determined non-locally sequential sampling pattern;
c) assembling the MRF data into a series of signal evolutions;
d) comparing the series of signal evolutions to a dictionary of known signal evolutions to determine tissue properties of the subject; and
e) generating a report indicating the tissue properties of the subject.
2. The method of claim 1 , further comprising segmenting each point in the non-locally sequential sampling pattern into a plurality of echo times to control off-resonance effects.
3. The method of claim 1 , wherein determining the non-locally sequential sampling pattern includes selecting a Poisson-disc sampling pattern.
4. The method of claim 1 , wherein the non-locally sequential sampling pattern is a pseudorandom sampling pattern.
5. The method of claim 1 , further comprising using the computer system to reconstruct an undersampled image of the subject using a low-rank reconstruction.
6. The method of claim 5 , wherein the low-rank reconstruction is of the form:
min
α
1
2
y
-
E
U
k
α
2
2
+
λ
∑
r
R
r
(
α
)
where E represents the encoding matrix that contains sampling masks and coil sensitives, U k represents a subspace learned from the dictionary by using a singular-value decomposition (SVD), α=U k H x represents compressed low-rank images, and
∑
r
R
r
(
α
)
*
represents a local low-rank regularization on low-rank images with block size r.
7. The method of claim 1 , wherein comparing the series of signal evolutions to a dictionary of known signal evolutions includes reconstructing an image of the subject form the MRF data and determining a maximum of an inner product between the image of the subject and the dictionary.
8. A magnetic resonance fingerprinting (MRF) system comprising:
a magnet system configured to generate a polarizing magnetic field about at least a portion of a subject;
a magnetic gradient system including a plurality of magnetic gradient coils configured to apply at least one magnetic gradient field to the polarizing magnetic field;
a radio frequency (RF) system configured to apply an RF field to the subject and to receive magnetic resonance signals from the subject using a coil array;
a computer system programmed to:
determine a non-locally sequential sampling pattern for a Cartesian grid of k-space;
control the magnetic gradient system and the RF system to perform a series of sequence blocks using acquisition parameters that vary between sequence blocks to acquire MRF data from a subject using the Cartesian grid of k-space and the non-locally sequential sampling pattern;
assemble the MRF data into a series of signal evolutions;
compare the series of signal evolutions to a dictionary of known signal evolutions to determine tissue properties of the subject; and
generate a report indicating the tissue properties of the subject.
9. The system of claim 8 , wherein the computer system is further programmed to segment each point in the non-locally sequential sampling pattern into a plurality of echo times to control off-resonance effects.
10. The system of claim 8 , wherein the computer system is further configured to select a Poisson-disc sampling pattern to determine the non-locally sequential sampling pattern.
11. The system of claim 8 , wherein the non-locally sequential sampling pattern is a pseudorandom sampling pattern.
12. The system of claim 1 , wherein the computer system is further configured to reconstruct an undersampled image of the subject using a low-rank reconstruction.
13. The system of claim 12 , wherein the low-rank reconstruction is of the form:
min
α
1
2
y
-
E
U
k
α
2
2
+
λ
∑
r
R
r
(
α
)
where E represents the encoding matrix that contains sampling masks and coil sensitives, U k represents a subspace learned from the dictionary by using a singular-value decomposition (SVD), α=U k H x represents compressed low-rank images, and
∑
r
R
r
(
α
)
*
represents a local low-rank regularization on low-rank images with block size r.
14. The system of claim 1 , wherein comparing the series of signal evolutions to a dictionary of known signal evolutions includes reconstructing an image of the subject form the MRF data and wherein the computer system is further configured to determine a maximum of an inner product between the image of the subject and the dictionary.