Staging liver fibrosis by analysis of non-linear normalization texture in gadolinium-enhanced magnetic resonance imaging

Abstract

A reliable non-invasive approach for accurate staging fibrosis is still under development. In this study, we utilized a computer aided diagnosis (CAD) system to analyze equilibrium phase magnetic resonance (MR) images in a rat model to determine relative accuracy of staging liver fibrosis with CAD. Methods: Experimental rats were injected with a mixture of CCl 4 to generate varying stages of fibrosis and equilibrium phase images of rats were acquired. All rats were grouped based on histological stages (F0, F1, F2, F3 and F4). During the process of CAD, the first step was manual section of region of interest (ROI) with 10 * 10 pixels guided by histology. Next, Lloyd's algorithm and linear normalization were both used to compress ROIs to 256 gray level. Then, 80 texture features based on the gray level co-occurrence matrix were extracted. Lastly, the back-propagation (BP), linear, k-nearest neighbor and support vector machine classifiers were individually used to classify F0-F4.Results: With histological staging as a reference, we found that performance of the BP classifier based on Lloyd's algorithm in staging fibrosis was better than that of linear normalization. Staging accuracy rates of the former were 67. 27%, 80.26% and 79.41% for F0 versus F3, F2 versus F4 and F3 versus F4, respectively, secondary to none of the other classifiers. Conclusions: The assessment of equilibrium phase MR images with CAD can effectively stage rat liver fibrosis. This article highlights progress in developing a reliable non-invasive liver fibrosis staging procedure.

Quantitative diagnosis of rotator cuff tears based on sonographic pattern recognition

“…The gray-scale intensities of echogenicities and statistical correlation between pixel values would provide useful information for differentiating lesion types. In this study, the gray-scale co-occurrence matrices (GLCM) [35, 36], which calculated the second-order statistics of ultrasound texture inside the lesion area, were proposed as features. The statistics revealed the correlations between adjacent pixels with different combinations of gray-scales.…”

“…The matrix element P = [ p ( i , j | d , θ )] represented the frequency of two adjacent pixels with values of i and j at a distance ( d ) and a direction ( θ ). Based on the matrix, 14 GLCM texture features were calculated [36]. Fig 2 illustrates the distance d = 1 and the direction θ = 0°, 45°, 90°, or 135° in the consideration of texture composition.…”

TMM: A comprehensive CAD system for hepatic fibrosis 5‐grade METAVIR staging based on liver MRI

“…GLCM is also studied for comparison. 4,7,11 Table 3 shows the accuracy rate for each stage and all the samples, and Figure 6.…”

“…[10][11][12] Goudarzi 10 extracted 18 GLCM texture features from the region of interest (ROI) of 1.5T/3.0T MRI and applied a two-stage optimized SVM and a left-side method to diagnose fibrosis as non-cirrhosis or cirrhosis with accuracies 68.17% and 71.98% for MR 1.5T and 3T respectively. Mou et al 11 used a rat model of liver fibrosis and combined 80 GLCM features with classifiers, such as back-propagation neural network (BPNN), linear classification, k-nearest neighbor (KNN) and SVM, to classify F0 to F4. They achieved the accuracies of F0 versus F3, F2 versus F4 ,and F3 versus F4 with 67.27%, 80.26% ,and 79.41% respectively.…”

“…According to the experience of radiologists, hepatic fibrosis at different stages leads to different texture patterns 9 . Therefore, the most classical texture feature GLCM is widely used 10–12 . Goudarzi 10 extracted 18 GLCM texture features from the region of interest (ROI) of 1.5T/3.0T MRI and applied a two‐stage optimized SVM and a left‐side method to diagnose fibrosis as non‐cirrhosis or cirrhosis with accuracies 68.17% and 71.98% for MR 1.5T and 3T respectively.…”