=Paper=
{{Paper
|id=None
|storemode=property
|title=Extracting the Fine Structure of the Left Cardiac Ventricle in 4D CT Data
|pdfUrl=https://ceur-ws.org/Vol-715/bvm2011_55.pdf
|volume=Vol-715
}}
==Extracting the Fine Structure of the Left Cardiac Ventricle in 4D CT Data==
https://ceur-ws.org/Vol-715/bvm2011_55.pdf
Extracting the Fine Structure of the Left
Cardiac Ventricle in 4D CT Data
A Semi-Automatic Segmentation Pipeline
Juliane Dinse1 , Daniela Wellein1 , Matthias Pfeifle2 , Silvia Born1 , Thilo Noack3 ,
Matthias Gutberlet3 , Lukas Lehmkuhl3 , Oliver Burgert1 , Bernhard Preim4
1
VCM/ICCAS, Universität Leipzig
2
Neurochirurgische Klinik, Universitätsklinikum Tübingen
3
Herzzentrum Leipzig
4
Institut für Simulation und Graphik, Otto-von-Guerike-Universität Magdeburg
daniela.wellein@medizin.uni-leipzig.de
Abstract. We propose a pipeline for the segmentation of the left car-
diac ventricle (LV) in 4D CT data based on the random walker (RW)
algorithm. A segmentation of the LV allows to extract clinical relevant
parameters such as ejection fraction (EF) and volume over time (VoT),
supporting diagnostic and therapy planning. The presented pipeline
works aside approaches incorporating annotated databases, statistical
shape modeling or atlas-based segmentation. We have tested our seg-
mentation approach on six clinical 4D CT datasets including different
pathologies and typical artifacts and compared the segmentation results
to manually segmented slices. We achieve a minimum sensitivity of 86%
and specificity of 96%. The resulting EF and VoT is comparable to
known reference values and reflects the present pathologies correctly.
Additionally, we tested three different routines for thresholding the RW
probability maps. An interview with surgical and radiological experts to-
gether with high sensitivity scores indicates the superiority of the fixed
threshold selection method – especially in the presence of pathologies.
The segmentation is also correct near problematic fine structures such
as cardiac valves, papillary muscles and the apex of the heart.
1 Introduction
A segmentation of the left cardiac ventricle (LV) in 4D CT data is necessary for
a functional analysis of the heart. Clinical parameters are often approximated
and/or only estimated in end-diastole and end-systole. With the opportunity to
measure the ejection fraction (EF) and the volume over time (VoT), diagnos-
tics and therapy planning can be supported. For such an analysis the correct
segmentation of cardiac fine structures such as the valves, the papillary muscle
(PM), the apex and the heart’s inner myocardium, is important.
For the segmentation of the LV various approaches exist. Jolly et al. combine
edge, region and shape information for the segmentation of the left ventricle
in MRI [1]. Zheng et al. use a database of models and an automatic fitting
Pipeline for Segmentation of Left Ventricle 265
to segment all four cardiac chambers in CT data [2]. For the fitting they use
manually defined landmarks to deform the control points of the models according
to the data. Kirisli et al. propose a multi-atlas-based segmentation of the whole
heart in 3D CTA data [3]. They match atlas information and data using a
non-rigid registration framework.
Unlike the presented segmentation approaches, we focus on a correct segmen-
tation of the above mentioned cardiac structures. The proposed semi-automatic
pipeline allows the inspection (and correction) of all intermediate results to in-
corporate expert knowledge of the segmenting radiological and surgical experts.
2 Materials and Methods
We evaluated our segmentation pipeline on six clinical 4D CT datasets with 10
or 20 time steps. The dimensions of the datasets have been 512x512x254±111
and pixel size was 0.36-0.58mm with a slice distance of 0.67-1.00mm. The
pathologies present in the datasets include coronary artery disease, aortic steno-
sis, and restricted, dilated, and hypertrophic cardiomyopathy (cardiac muscle is
too less/too much compacted). Two of the datasets included a metallic mitral
valve replacement, and therefore a high degree of metal artifacts.
The proposed pipeline has been integrated into the volume data processing
and visualization platform VolV [4] which has already been applied to segmen-
tation in neurosurgical intervention planning [5]. The pipeline consists of three
main steps: pre-processing, random walker (RW), and post-processing (Fig. 1).
Pre-processing includes anisotropic filtering [6] and an edge enhancement filter-
ing, implemented as the difference image between the original and the mean
filtered data. The main step of the pipeline is the RW algorithm, which is
formulated on a graph and has strong connections to electrical circuit and po-
tential theory. Its input are labeled voxels belonging to two classes: LV and
non-LV structure, i.e. background. The algorithm analytically determines the
probability that a random walker starting at each unlabeled voxel will first reach
one of the pre-labeled voxels [7]. Output of RW is a probability map in which
each voxel is assigned to the label with the highest probability. To initialize
the pipeline (Fig. 1), the user has to label a few voxels of the LV and all other
anatomical structures (background) only for t = 1. The labeling does not require
any delineation of object boundaries and can be inexact, and, therefore quick.
Fig. 1. The segmentation pipeline for extracting the LV in 4D CT data.
266 J. Dinse et al.
In the post-processing step, the RW probability map is thresholded to generate
the final LV segmentation for the current time step. We used three different
thresholding approaches: a manual threshold selection, fixing the threshold to a
specific probability and calculating the minimum in the histogram of each time
step. Also, a set of labeled seed points for the next time step is generated by
reducing the segmentation results of the LV and the background with multiple
erosions.
For evaluation, three slices in every time step of each of the six datasets
(capturing the valves, the papillary muscles, and the cardiac apex) have been
manually segmented, which leads to an overall number of 3 ∗ 90 = 270 slices.
The sensitivity (SE) and specificity (SP) for each dataset and each threshold-
ing method is computed. The volume over time and the ejection fraction to-
gether with its mean squared error to the manual segmentation (MSEEF ) are
calculated and compared to known reference values. In an interview, four sur-
gical/radiological experts have ranked the segmentation results of the different
thresholding methods and rated them according to the correct detection of the
previously mentioned fine structures.
3 Results
Figure 2 shows the segmentation results for three exemplary datasets in the
coronal view. The SE and SP for the three different thresholding routines are
depicted in Figure 3. The manual thresholding results in an average SE of
91.8% (±6.9) and SP of 97.7% (±1.2). Fixed thresholding achieves a SE of
94.1% (±3.8) and SP of 96.7% (±2.2). The minimum thresholding approach
leads to an average SE of 88.6% (±9.1) and SP of 98.4% (±1.2). In Figure 4
the VoT for three exemplary datasets is shown. EF for all datasets is depicted
in Table 1. MSEEF of our method is 2.2% (±1.4) for manual thresholding,
2.2% (±2.5) for fixed thresholding, and 4.3% (±3.6) for minimum thresholding.
(a) DS4 (b) DS5 (c) DS6
Fig. 2. Segmentation of three exemplary datasets showing different pathologies: (a)
restricted, (b) hypertrophic, and (c) dilated cardiomyopathy. The interior of the LV as
well as fine structures like the aortic outflow tract and papillary muscles are correctly
detected.
Pipeline for Segmentation of Left Ventricle 267
In the interview the clinical experts preferred the segmentation results of
the fixed threshold routine in four out of six datasets. They also confirmed the
overall correct detection of the cardiac fine structures.
4 Discussion
The previously given SE and SP values correlate with the expert ratings, which
indicate, that in our case three manually segmented slices might be enough
to measure the segmentation quality. A high SE and SP with a low standard
deviation lets us conclude, that the performance is constantly high throughout all
tested datasets. The VoT curves (Fig. 4) reflect the general pump (dys-)function
of the present pathologies. In DS4 the restrictive cardiomyopathy leads to a delay
in proper blood filling in the diastole. This results in a decreased slope of the
VoT for the second half of the heart cycle.
The obtained EF values also reflect the underlying pathologies. In DS6 the
EF is below 50% for the fixed thresholding. A value, that corresponds to the
present dilated cardiomyopathy, in which the amount of blood being pumped
out of the LV in the diastole is reduced. In addition, we compared the MSEEF
to the results presented by Zheng et al. [2]. On six datasets with 10 time steps
each, they obtained an MSEEF of 2.3% (±1.6). Our MSEEF of 2.2% for fixed
thresholding is comparable, but one has to mention, that this measure is com-
puted in relation to only three manually segmented slices, which might not be
enough to deduce the volume of the LV.
In some of the datasets, the plane of the mitral valves, that delineates the LV,
is not correctly detected with our pipeline. The dysfunction of the mitral valve
leads to the leaflets hitting the cardiac wall and, therefore, they are not detected
in the segmentation. In addition, the rapid and inconsistent movement of the
Fig. 3. Sensitivity and
specificity for the applied
thresholding routines and
all datasets.
Fig. 4. Volume over time curves for three
exemplary datasets. For present pathologies
see Figure 2.
268 J. Dinse et al.
Table 1. The calculated ejection fraction of all datasets (in %).
Segmentation DS1 DS2 DS3 DS4 DS5 DS6
Manual 47.6 61.5 51.9 63.4 58.2 47.5
Manual Thresh. 48.8 62.1 52.8 60.9 62.2 51.1
Fixed Thresh. 54.5 62.9 52.2 60.6 59.1 46.6
Minimum Thresh. 50.5 61.5 58.3 66.3 68.6 50.8
mitral valve makes them undetectable in some datasets and, the segmentation
leaks into the atrium. Furthermore, the PM could be wrongly segmented as
belonging to the LV, when the connections between myocardium and PM are
too small in the datasets or the contrast between those two structures is too high
(PM not dark enough).
To conclude, our pipeline provides a segmentation of the LV that is indepen-
dent from atlas or large database information. The segmentation quality is high
also in the presence of pathologies and artifacts and is able to correctly detect
cardiac fine structures. These findings could also be confirmed in an interview
with clinical experts.
A promising future work approach would be to use the grey value frequencies
in a weighting function for the RW. This was proposed by Grady and Jolly [8],
who have shown that their approach outperforms purely intensity-based weight-
ing functions.
References
1. Jolly MP. Combining edge, region, and shape information to segment the left ven-
tricle in cardiac MR images. Lect Notes Computer Sci. 2010; p. 482–90.
2. Zheng Y, Barbu A, Georgescu B, et al. Four-chamber heart modeling and auto-
matic segmentation for 3-D cardiac CT volumes using marginal space learning and
steerable features. IEEE Trans Med Imaging. 2008;27(11):1668–81.
3. Kirisli HA, Schaap M, Klein Sea. Fully automatic cardiac segmentation from 3D
CTA data: a multi-atlas based approach. Proc SPIE. 2010;7623:762305–1.
4. Pfeifle M, Born S, Fischer J, et al. VolV - Eine OpenSource-Plattform für die
Medizinische Visualisierung. Proc CURAC. 2007; p. 193–6.
5. Wellein D, Pfeifle M, Althuizes M, et al. A cortex segmentation pipeline. Proc
BVM. 2010; p. 271–5.
6. Perona P, Malik J. Scale-space and edge detection using anisotropic diffusion. IEEE
Trans Pattern Anal Mach Intell. 1990;12(7):629–39.
7. Grady L. Random walks for image segmentation. IEEE Trans Pattern Anal Mach
Intell. 2006;28(11):1768–83.
8. Grady L, Jolly MP. Weights and topology: a study of the effects of graph construc-
tion on 3D image segmentation. Med Image Comput Comput Assist Interv. 2008;
p. 153–61.