US 8,798,349 B2Granted

Systems and methods for detecting arterial input function (AIF)

AssigneeThe Chinese University of Hong Kong

InventorsLin Shi, Defeng Wang, Pheng Ann Heng

Life of this patent

Estimating the term…

Abstract

Disclosed are methods, systems and apparatuses for detection of arterial input function (AIF) in MRI, specially DCE MR images, comprising automatic selection of AIF based on affinity propagation (AP) clustering method.

Claims14

Dependent claims

A method of detecting arterial input function (AIF) in dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) scans of a tissue or an organ in a subject, comprising:

(a) designing a mask to a major region of a scan imaging;
(b) calculating a concentration of intravascular contrast agent Cp (t) by
Cp
⁡
(
t
)
=
k
⁢
⁢
Δ
⁢
⁢
R
2
*
=
-
1
TE
⁢
ln
⁢
S
⁡
(
t
)
S
⁡
(
0
)
,
wherein t represents time;
ΔR2 * represents transverse relaxation at time t, S (t) is a signal intensity at time t, S (0) is a pre-injection baseline signal intensity, k is a scaling factor, wherein S (0) is the mean intensity of the time slot that the contrast agent arrives at the tissue;
(c) calculating the area under the concentration curve;
(d) discarding Paus percent of the curves with the smallest area, and Paub percent of the remaining curves with the biggest area;
(e) determining peak values of every pixel in every scan imaging and excluding Ppks percent of a pixel which peak value is smaller than a big peak value of all pixels in the scan;
(f) calculating a dissimilarity value between an AIF and a processed concentration time curve by
disij=√{square root over ((xi−xj)(xi−xj)′)}{square root over ((xi−xj)(xi−xj)′)},
and selecting Psim percent of small dissimilarity values, wherein xi represents the processed concentration time curve and xj represents the AIF curve;
(g) clustering concentration time curves using a fast-AP clustering algorithm, in which concentration time curves of remaining pixels are simultaneously considered as input data in the cluster analysis;
(h) selecting additional scans to be processed by above steps (a) to (g);
(i) processing the additional scans by above steps (a) to (g);
and
(j) calculating a final estimate of AIF, which is a mean curve with the lowest first moment based on the selection of the scans.

45 more elements

A system of detecting arterial input function (AIF) in a dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) of a tissue or an organ in a subject, comprising:

(a) a first module configured for designing a mask to a major region of a scan imaging;
(b) a second module configured for calculating a concentration of intravascular contrast agent Cp (t) by
Cp
⁡
(
t
)
=
k
⁢
⁢
Δ
⁢
⁢
R
2
*
=
-
1
TE
⁢
ln
⁢
S
⁡
(
t
)
S
⁡
(
0
)
,
wherein t represents time;
ΔR2 * represents transverse relaxation at time t, S (t) is a signal intensity at time t, S (0) is a pre-injection baseline signal intensity, k is a scaling factor, wherein S (0) is the mean intensity of the time slot that the contrast agent arrives at the tissue;
(c) a third module configured for calculating an area under the concentration curve so as to discard Paus percent of the curves with the smallest area, and Paub percent of the remaining curves with the biggest area;
(d) a fourth module configured for determining peak values of every pixel in every slice image and excluding Ppks percent of a pixel which peak value is smaller than a big peak value of all pixels in the slice;
(e) a fifth module configured for calculating a dissimilarity value between an AIF and a processed concentration time curve by
disij=√{square root over ((xi−xj)(xi−xj)′)}{square root over ((xi−xj)(xi−xj)′)},
and selecting Psim percent of small dissimilarity values, wherein xi represents the processed concentration time curve and xj represents the AIF curve.
(f) a sixth module configured for clustering concentration time curves using a fast-AP clustering algorithm, in which concentration time curves of remaining pixels are simultaneously considered as input data in the cluster analysis.

40 more elements

One or more non-transitory computer readable media storing computer executable code for instructing one or more processors for detecting arterial input function (AIF) in dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) scans of a tissue or an organ in a subject, comprising:

(a) a first processor for designing a mask to a major region of a scan imaging;
(b) a second processor for calculating a concentration of intravascular contrast agent Cp (t) by
Cp
⁡
(
t
)
=
k
⁢
⁢
Δ
⁢
⁢
R
2
*
=
-
1
TE
⁢
ln
⁢
S
⁡
(
t
)
S
⁡
(
0
)
,
wherein t represents time;
ΔR2 * represents transverse relaxation at time t, S (t) is a signal intensity at time t, S (0) is a pre-injection baseline signal intensity, k is a scaling factor, wherein S (0) is the mean intensity of the time slot that the contrast agent arrives at the tissue;
(c) a third processor for calculating the area under the concentration curve;
(d) a fourth processor for discarding Paus percent of the curves with the smallest area, and Paub percent of the remaining curves with the biggest area;
(e) a fifth processor for determining peak values of every pixel in every scan image and excluding Ppks percent of a pixel which peak value is smaller than a big peak value of all pixels in the imaging;
(f) a sixth processor for calculating a dissimilarity value between an AIF and a processed concentration time curve by
disij=√{square root over ((xi−xj)(xi−xj)′)}{square root over ((xi−xj)(xi−xj)′)},
and selecting Psim percent of small dissimilarity values, wherein xi represents the processed concentration time curve and xj represents the AIF curve.
(g) a seventh processor for clustering concentration time curves using a fast-AP clustering algorithm, in which concentration time curves of remaining pixels are simultaneously considered as input data in the cluster analysis.

41 more elements

Dependencies are inferred from the claim text.

patentgrep

Loading patent…

Back to results