1. Automatic Basal Slice Detection for Cardiac Analysis
Mahsa Paknezhad1, Stephanie Marchesseau2, Michael S. Brown1
1National University of Singapore, 2A*STAR-NUS Clinical Imaging Research Centre, Singapore
INTRODUCTION
Identification of the basal slice in cardiac imaging is a key step to measuring
the ejection fraction (EF) of the left ventricle. Previous approaches assumed
that the basal slice is the first short-axis slice below the mitral valve.
However, guidelines published in 2013 by the society for cardiovascular
magnetic resonance indicate that the basal slice is the uppermost short-axis
slice with more than 50% myocardium surrounding the blood cavity.
Consequently, these existing methods are at times identifying the incorrect
short-axis slice.
GOALS
Since correct identification of the basal slice under these guidelines is
challenging due to the poor image quality and blood movement during
image acquisition. This work proposes an automatic algorithm that focuses
on the two-chamber view slice to find the basal slice.
METHOD
An active shape model [2] was trained and used to automatically segment
the two-chamber view through the whole cardiac cycle.
1. Training the active shape model
RESULTS
The method was applied to clinical data from 51 cases from three different
scanners. For each test case, the basal slice was detected in the end-systolic
and the end-diastolic phases, giving a total of 102 basal slices. Overall, the
proposed algorithm selected the same basal slices as the expert selection for
47 out of the 51 subjects for end-systole and 43 out of the 51 subjects for
end-diastole. The average time to detect the basal slices for the end-diastole
and the end-systole was about 24 seconds.
REFERENCES
[1] Schulz-Menger, J., Bluemke, D. A., Bremerich, J., Flamm, S. D., Fogel, M. A., Friedrich,
M. G., and Nagel, E., “Standardized Image Interpretation and Post Processing in
Cardiovascular Magnetic Resonance: Society for Cardiovascular Magnetic Resonance
(SCMR),” JCMR 15(35), 10-1186 (2013).
[2] Cootes, T. F., Taylor, C. J., Cooper, D. H., and Graham, J., “Active Shape Models - Their
Training and Application,” Computer vision and image understanding 61(1), 38-59
(1995).
time
3. Apply the ASM on the test two-chamber view image sequence
5. Create the temporal binary profile for each short-axis slice
temporal profiles for the basal short-axis slices
SM
AM
PCA
Eigenvectors up to
98% cumulative
energy
Mean
PCA
For each contour point
Align
contour points
50 MRI scans
of the heart
Intensity
Data
+
Eigenvectors up to
98% cumulative
energy
Mean
+
Training the
ASM
2. Initialization of the ASM on the cardiac cycle
Find the location to initialize ASM
ASM with
limited iterations
two-chamber
view image
sequence
ASM independent
segmentation
ASM
Segmentation of the two-chamber view sequence
NCC
>NCC
>
segmented
sequence
time 25time 11time 1
segmented
Two-chamber
view
short-axis
view slice
time
1
11
25
1-D binary profile
Area(pixels)
end-systolic end-diastolic
4. Estimate the end-diastolic and the end-systolic phases
1234
5
1
2
3
4
5
two-chamber view at end-diastole 1-D profiles for the basal
short-axis slices
time
Shot-axissliceid
binary profileintensity profile
ED
ED
ED
ED
ED
Thebasalsliceatend-diastole
6. Select the basal slice using the temporal profiles while considering the
estimated end-diastolic and the end-systolic phases
using previous
segmentation
results
1 5 7 11 17 19 23 25
1 5 7 11 17 19 23 25