The segmentation of scarred and nonscarred myocardium in Cardiac Magnetic Resonance (CMR) is obtained using different features and feature combinations in a Bayes classifier. The used features are found as a local average of intensity values and the underlying texture information in scarred and nonscarred myocardium. The segmentation classifier was trained and tested with different experimental setups and parameter combinations and was cross validated due to limited data. The experimental results show that the intensity variations are indeed an important feature for good segmentation, and the average area under the Receiver Operating Characteristic (ROC) curve, that is, the AUC, is 91.58 ± 3.2%. The segmentation using texture features also gives good segmentation with average AUC values at 85.89 ± 5.8%, that is, lower than the direct current (DC) feature. However, the texture feature gives robust performance compared to a local mean (DC) feature in a test set simulated from the original CMR data. The segmentation of scarred myocardium is comparable to manual segmentation in all the cross validation cases.
After a myocardial infarction (MI), the myocardium does not function properly due to scarring of the tissue. Late Gadolinium Enhanced Cardiac Magnetic Resonance (LGE-CMR) imaging is used for assessing morphology of myocardium after an MI.
Segmenting the scarred areas from the healthy myocardium is an important prerequisite for various diagnostic analyses. For example, the scar size is largely responsible in left ventricular remodeling [
Earlier work focuses mostly on manual or semiautomatic methods for segmenting the scarred area [
Corsi et al. [
Dikici et al. [
The approach by Tao et al. [
Recent studies showed that the scar tissue is heterogeneous in nature and that the mortality of patients with reduced LVEF depends on the heterogeneity of scar tissue [
Section
This section explains the material and methods used in this work. Section
The Department of Cardiology in Stavanger University Hospital provided the LGE-CMR images for our experiments, and the experiments were conducted in MATLAB. The LGE-CMR images is a group of 24 patients, all with high risk of getting arrhythmia, and they were stored according to the Digital imaging and communications in medicine (DICOM) format with 512 × 512 pixel resolution. The number of image slices with visible scar in each patient varies approximately from 5 to 12 depending on the size of scar and heart. Only short-axis CMR images were used in our experiments. The LGE-CMR images were acquired from 1.5 Tesla Philips Intera machine. Images were obtained with a pixel size of 0.82
Cropped short-axis CMR image showing manual segmentation of myocardium and scar tissues. The green and blue dots in the image are manually marked (by cardiologist) coordinates to segment the myocardium and scar. The magenta and yellow contours generated by cubic spline interpolations of the above coordinates show the myocardium and scar tissues, respectively.
The segmentation of the scarred and nonscarred myocardium is obtained by using Bayes decision theory. The class specific probability density function (PDF) modeling specific feature vector values discriminating the scar or the healthy myocardial areas is estimated as
For a specific image,
As discussed in Section
The detailed process of extracting the DC and texture features is discussed in the following Sections
Historically, in electronics field, the
Sparse representations and learned dictionaries have been shown to work well for texture classification by Skretting and Husoy in [
A dictionary
Dictionary learning is the task of learning or training a dictionary on a available training set such that it adapts well to represent that specific class of signals. The training vectors and the sparse coefficients are arranged as columns in the matrices
The RLS-DLA algorithm presented in [
With the inclusion of an adaptive forgetting factor
In Frame Texture Classification Method (FTCM) presented by Skretting and Husoy [
Texture feature extraction requires testing and training phases and the whole CMR data set is divided into training and testing images. In training phase, texture feature,
A pixel should give less error or residual to the dictionary it belongs. Finally, the residuals
The main aim of our experiments was to compute different features and compare their ability in segmenting the scar from the healthy myocardium using Bayes classifier. The objectives that were explored in our experiments are the discriminative power of the features, the robustness of the features, and examples of the segmented images for illustration. The subsequent subsections discuss about the experimental setup.
Our experiments consist of three cases or setups. Each experimental case used different parameters for finding the DC and texture features. Table
The three experimental cases used for segmentation of scarred and nonscarred myocardium.
Experimental parameters | Cases | ||
---|---|---|---|
Case (i) | Case (ii) | Case (iii) | |
Window size |
|
|
|
Sparsity | 2 | 4 | 2 |
No. of training groups |
3 (18) | 3 (18) | 2 (12) |
No. of testing groups |
1 (6) | 1 (6) | 2 (12) |
In all CMR Images, we take into account only myocardium segmented by cardiologists. The training phase involves extraction of the DC and texture features in all the three cases. The process was illustrated in Sections
Two sets of training vectors were generated from scar and healthy myocardium. The neighborhood size 3 × 3 and 5 × 5 were used to form training vectors as explained in Section
The texture features were calculated using the same training and test set employed in finding the DC features. Two sets of training vectors were generated from the scarred and nonscarred myocardium segmented by cardiologists. The training vector and test vectors were generated in the same way as in the DC feature experiment using the same neighborhood sizes. Consider the pixels on the border zones, their neighborhood extends into other regions that are not under consideration. If we use training vectors from border regions, then the dictionaries might learn the texture properties of other regions along with the texture properties they were intended to learn. So, the training vectors for the pixels whose neighborhood span other regions were not considered in our experiments. This is depicted in Figure
The training vector of a pixel is extracted as long as its neighborhood is within one texture area. The neighborhood of pixels
In the testing phase, the DC and texture features were generated as in the training phase. As described in Section
The DC and texture features
The standard deviation of AUC values was computed to examine the robustness of the DC and texture features. The standard deviation of AUC value was computed from the mean true positives and true negative for all the 24 patients. The sensitivity (true positives) and specificity (true negatives) of each patient in all the cross validations of case (i), case (ii), and case (iii) had to be averaged before finding the standard deviation due to the multiple testing of each patient in all the cases. The confidence interval in Figure
Referring to the ROC analysis, the discriminative power of the four feature combinations and the robustness of the DC and texture feature are discussed in the subsequent sections.
The average ROC curves of the three cases are shown in Figure
Comparison of the average AUC values from the ROC analysis of the four different feature combinations in the three experimental cases.
Testing | ||||
---|---|---|---|---|
Average of AUC values | ||||
Cases (window size, sparsity) | dc |
|
|
|
Case (i) ( |
0.92 | 0.92 | 0.92 | 0.85 |
Case (ii) ( |
0.92 | 0.92 | 0.92 | 0.86 |
Case (iii) ( |
0.91 | 0.90 | 0.91 | 0.86 |
ROC curves for the three experimental cases and feature combinations: (i) dc, (ii) dc,
ROC curves of the DC, dc, and texture feature,
The average AUC values of texture feature,
Figure
Figure
Patient specific scaling of intensity values might be a problem as for the robustness of the analyzed features. Hence, we focused to analyze the robustness of DC and texture features individually. Table
Comparison of the standard deviation AUC values of 24 patients from the fourfold cross validation of four different feature combinations in the three experimental cases.
Standard deviation of AUC values of 24 patients | ||||
---|---|---|---|---|
Cases (window size, sparsity) | dc |
|
|
|
Case (i) ( |
0.032 | 0.031 | 0.035 | 0.072 |
Case (ii) ( |
0.032 | 0.032 | 0.034 | 0.074 |
Case (iii) ( |
0.035 | 0.039 | 0.041 | 0.070 |
In order to show that the texture feature performance is robust on the outliers, we simulated a test set by scaling the original intensities of CMR slices. The CMR images used in our work are from the same MRI device. Even though the MRI machine automatically tries to produce MRI images to be in the same scale of intensities, the CMR images of all the patients used in our work vary from one patient to another. With our CMR data, the DC feature performed better than the texture feature. The AUC values for the simulated test were calculated using the already trained classifier in case (iii). The AUC values of the simulated testing set of a patient are plotted against the scaling factor in Figure
AUC values of 8 patients whose CMR slices are scaled at different factors.
The segmentation of the scarred and nonscarred myocardium using the DC and texture feature was compared to Dikici et al. [
Figure
Similarity measure—Dice index [
The segmentation plots of the scarred myocardium in the CMR images of the patients implanted with ICD using the DC value and texture feature,
The segmentation results of the scarred myocardium in the CMR image of a patient implanted with ICD using the DC and texture feature,
The DC feature segments the scarred myocardium comparatively better than the texture feature on our CMR data, and the combination of the DC and texture features does not seem to improve the overall performance compared to the DC feature alone. The DC feature also needs fewer training patients than the texture feature method seems to need. However, the texture feature from the learned dictionaries and sparse representation is shown to be much more robust when it comes to scaling variations, that is, variations in the intensity value range. The DC feature seems to reflect the way cardiologist perceive the scarred myocardium in CMR images whereas our results indicate that the texture features can be explored to investigate the heterogeneous nature of the scarred myocardium. Our belief that the texture feature can be used to explore the properties of the scarred myocardium got strengthened with the cardiac segments experiment based on the probability mapping of the scarred myocardium in our recent paper [
The authors declare that they have no conflict of Interests.