Medical Imaging & Image Processing Lab M I I P Center for Informatics Science Nile University, Egypt Special Topic In Cardiac MR Imaging SENC Abdallah G.
Download ReportTranscript Medical Imaging & Image Processing Lab M I I P Center for Informatics Science Nile University, Egypt Special Topic In Cardiac MR Imaging SENC Abdallah G.
Slide 1
Medical Imaging & Image Processing Lab
M
I
I
P
Center for Informatics Science
Nile University, Egypt
Special Topic In Cardiac MR Imaging
SENC
Abdallah G. Motaal
Research Assistant,
MSc. Candidate
Nile University, Egypt
Slide 2
Layout
• Background
• Cardiac Magnetic Resonance Imaging
• Strain Encoded Imaging
Theory
Anatomy Image Enhancement
Strain values correction
Slide 3
Background
•
Magnetic resonance imaging (MRI) is a safe,
noninvasive imaging technique that creates detailed
images of organs and tissues.
• MRI uses radio waves and magnets to create images of
organs and tissues, unlike computed tomography or
conventional x rays,
MRI imaging does not use ionizing radiations.
Slide 4
MR Cardiac Imaging
• Cardiac MR is used to image the heart while it is
beating, producing both still and moving pictures of the
heart and major blood vessels.
• Cardiologists use cardiac MR to get images and to look
at the structure and function of the different region of
the heart. These images can help them decide how best
to treat patients with heart problems.
Slide 5
Slide 6
Cardiac Imaging (2)
• Cardiac MRI is a common test for diagnosing
and evaluating a number of diseases and
conditions, including:
Coronary artery disease
Heart valve problems
Congenital heart defects
Cardiomyopathies
Cardiac Tumors
Image with a cardiac infarction
Slide 7
What is SENC
• SENC is a technique that is used to quantify regional
function of heart by direct encoding of regional strain of
the heart into the acquired image.
• The technique measure the strain in the direction
orthogonal to the image plane.
• Therefore, in case of short-axis images, only the
longitudinal compression of the myocardium from base
to apex is measured. On the other hand, circumferential
shortening of the myocardium can be measured in the
long-axis views of the heart (such as the four-chamber
view).
Slide 8
Short Axis
Slide 9
Theory
• Tagging
– Tags are temporary non-invasive markers
whose deformations follow the motion of
underlying tissue.
Diastole
Systole
Diastole
Systole
Cardiac image before and after applying SPAMM tagging
Slide 10
Theory (2)
• Tagging is done in the direction parallel to the image plane.
• During cardiac cycle, these tag lines follow the motion of the
heart.
Before Contraction
After Contraction
• So by knowing the change in the tagging frequency we can
get the change in length, so we can calculate the strain.
Slide 11
Theory (3)
Because of stretching and contraction of the heart, the tagging
frequency ranges from wl and wh.
Slide 12
Because of the contraction of the heart, the tagging frequency will
change.
M(z)
S(w)
-wh
-wl
0
Myocardium
Static Tissue
wl
wh
Slide 13
Color Mapping
• The normal longitudinal strain ranges from 5 to -25.
• For qualitative assessment, strain values are represented by
colors.
5
-25
Slide 14
Time
Slide 15
Simulator
Slide 16
% Strain
Time
Slide 17
Anatomy Image Enhancement
• By adding both the L.T and H.T we will get an
informative anatomical Image. But the elevated noise
leads to poor contrast to noise ratio(CNR).
Slide 18
• Bayes classifier was used to increase the
CNR.
Based on statistical model of both background and
tissue signal
A. Motaal, M. Al-Attar, N. Osman ,A. Fahmy
CIBEC ’ 08
IEEE Proceedings
Slide 19
Interleaving Strain Error Correction
Slide 20
Non - Interleaved
T Seconds
T Seconds
L.T 1
L.T 2
L.T 3
L.T 4
L.T 5
H.T 1
H.T 2
The Total Time Taken is 2*T Seconds
Interleaved
T Seconds
L.T 1
L.T 3
H.T 2
L.T 5
H.T 4
The Total Time Taken is only T Seconds
H.T 3
H.T 4
H.T 5
Slide 21
One-heartbeat Interleaved SENC
Sequence
trigger
Problem: interframes motion of the heart
Slide 22
Proposed Correction Algorithm
1
L.T
2
H.T
3
L.T
4
H.T
n-1
H.T
............. ..
n
L.T
Δ1
Δ2
Δ3
Δ4
............. ..
Δn1
Δn
Correction Algorithm
1
L.T ‘
2
H.T ’
3
L.T ‘
4
H.T ‘
............. ..
n
L.T ‘
n-1
H.T ‘
Estimate w
w1
Compare Wn-1
w2
w3
Wn-1
NO
Diff
<β
Yes
STOP
Wn
Derivative
Slide 23
Phantom Experiment
• Numerical Computer Simulation was done to
generate SENC images with different strain
curves, and the proposed algorithms were tested
on them
Slide 24
Results
After Correction
Strain
Strain
Before Correction`
Time Frame
Time Frame
Slide 25
Results (cont.)
Slide 26
Summary
SENC is a technique for measuring regional
function of the heart.
We developed a Bayesian method to improve
the anatomical image reconstructed from
SENC.
We propose an algorithm to correct strain
measurements in interleaved acquisition.
The technique enables the reduction of the
required timeframes; hence, faster and better
quality imaging.
Slide 27
Thank You !
[email protected]
Medical Imaging & Image Processing Lab
M
I
I
P
Center for Informatics Science
Nile University, Egypt
Special Topic In Cardiac MR Imaging
SENC
Abdallah G. Motaal
Research Assistant,
MSc. Candidate
Nile University, Egypt
Slide 2
Layout
• Background
• Cardiac Magnetic Resonance Imaging
• Strain Encoded Imaging
Theory
Anatomy Image Enhancement
Strain values correction
Slide 3
Background
•
Magnetic resonance imaging (MRI) is a safe,
noninvasive imaging technique that creates detailed
images of organs and tissues.
• MRI uses radio waves and magnets to create images of
organs and tissues, unlike computed tomography or
conventional x rays,
MRI imaging does not use ionizing radiations.
Slide 4
MR Cardiac Imaging
• Cardiac MR is used to image the heart while it is
beating, producing both still and moving pictures of the
heart and major blood vessels.
• Cardiologists use cardiac MR to get images and to look
at the structure and function of the different region of
the heart. These images can help them decide how best
to treat patients with heart problems.
Slide 5
Slide 6
Cardiac Imaging (2)
• Cardiac MRI is a common test for diagnosing
and evaluating a number of diseases and
conditions, including:
Coronary artery disease
Heart valve problems
Congenital heart defects
Cardiomyopathies
Cardiac Tumors
Image with a cardiac infarction
Slide 7
What is SENC
• SENC is a technique that is used to quantify regional
function of heart by direct encoding of regional strain of
the heart into the acquired image.
• The technique measure the strain in the direction
orthogonal to the image plane.
• Therefore, in case of short-axis images, only the
longitudinal compression of the myocardium from base
to apex is measured. On the other hand, circumferential
shortening of the myocardium can be measured in the
long-axis views of the heart (such as the four-chamber
view).
Slide 8
Short Axis
Slide 9
Theory
• Tagging
– Tags are temporary non-invasive markers
whose deformations follow the motion of
underlying tissue.
Diastole
Systole
Diastole
Systole
Cardiac image before and after applying SPAMM tagging
Slide 10
Theory (2)
• Tagging is done in the direction parallel to the image plane.
• During cardiac cycle, these tag lines follow the motion of the
heart.
Before Contraction
After Contraction
• So by knowing the change in the tagging frequency we can
get the change in length, so we can calculate the strain.
Slide 11
Theory (3)
Because of stretching and contraction of the heart, the tagging
frequency ranges from wl and wh.
Slide 12
Because of the contraction of the heart, the tagging frequency will
change.
M(z)
S(w)
-wh
-wl
0
Myocardium
Static Tissue
wl
wh
Slide 13
Color Mapping
• The normal longitudinal strain ranges from 5 to -25.
• For qualitative assessment, strain values are represented by
colors.
5
-25
Slide 14
Time
Slide 15
Simulator
Slide 16
% Strain
Time
Slide 17
Anatomy Image Enhancement
• By adding both the L.T and H.T we will get an
informative anatomical Image. But the elevated noise
leads to poor contrast to noise ratio(CNR).
Slide 18
• Bayes classifier was used to increase the
CNR.
Based on statistical model of both background and
tissue signal
A. Motaal, M. Al-Attar, N. Osman ,A. Fahmy
CIBEC ’ 08
IEEE Proceedings
Slide 19
Interleaving Strain Error Correction
Slide 20
Non - Interleaved
T Seconds
T Seconds
L.T 1
L.T 2
L.T 3
L.T 4
L.T 5
H.T 1
H.T 2
The Total Time Taken is 2*T Seconds
Interleaved
T Seconds
L.T 1
L.T 3
H.T 2
L.T 5
H.T 4
The Total Time Taken is only T Seconds
H.T 3
H.T 4
H.T 5
Slide 21
One-heartbeat Interleaved SENC
Sequence
trigger
Problem: interframes motion of the heart
Slide 22
Proposed Correction Algorithm
1
L.T
2
H.T
3
L.T
4
H.T
n-1
H.T
............. ..
n
L.T
Δ1
Δ2
Δ3
Δ4
............. ..
Δn1
Δn
Correction Algorithm
1
L.T ‘
2
H.T ’
3
L.T ‘
4
H.T ‘
............. ..
n
L.T ‘
n-1
H.T ‘
Estimate w
w1
Compare Wn-1
w2
w3
Wn-1
NO
Diff
<β
Yes
STOP
Wn
Derivative
Slide 23
Phantom Experiment
• Numerical Computer Simulation was done to
generate SENC images with different strain
curves, and the proposed algorithms were tested
on them
Slide 24
Results
After Correction
Strain
Strain
Before Correction`
Time Frame
Time Frame
Slide 25
Results (cont.)
Slide 26
Summary
SENC is a technique for measuring regional
function of the heart.
We developed a Bayesian method to improve
the anatomical image reconstructed from
SENC.
We propose an algorithm to correct strain
measurements in interleaved acquisition.
The technique enables the reduction of the
required timeframes; hence, faster and better
quality imaging.
Slide 27
Thank You !
[email protected]