Liver Tissue Classification Research Articles

Machine Learning-Based Classification of Abnormal Liver Tissues Using Relative Permittivity.

The search for non-invasive, fast, and low-cost diagnostic tools has gained significant traction among many researchers worldwide. Dielectric properties calculated from microwave signals offer unique insights into biological tissue. Material properties, such as relative permittivity (εr) and conductivity (σ), can vary significantly between healthy and unhealthy tissue types at a given frequency. Understanding this difference in properties is key for identifying the disease state. The frequency-dependent nature of the dielectric measurements results in large datasets, which can be postprocessed using artificial intelligence (AI) methods. In this work, the dielectric properties of liver tissues in three mouse models of liver disease are characterized using dielectric spectroscopy. The measurements are grouped into four categories based on the diets or disease state of the mice, i.e., healthy mice, mice with non-alcoholic steatohepatitis (NASH) induced by choline-deficient high-fat diet, mice with NASH induced by western diet, and mice with liver fibrosis. Multi-class classification machine learning (ML) models are then explored to differentiate the liver tissue groups based on dielectric measurements. The results show that the support vector machine (SVM) model was able to differentiate the tissue groups with an accuracy up to 90%. This technology pipeline, thus, shows great potential for developing the next generation non-invasive diagnostic tools.

Harmonization of radiomic feature distributions: impact on classification of hepatic tissue in CT imaging.

Following the craze for radiomic features (RF), their lack of reliability raised the question of the generalizability of classification models. Inter-site harmonization of images therefore becomes a central issue. We compared RF harmonization processing designed to detect liver diseases in CT images. We retrospectively analyzed 76 multi-center portal CT series of non-diseased (NDL) and diseased liver (DL) patients. In each series, we positioned volumes of interest in spleen and liver, then extracted 9 RF (histogram and texture). We evaluated two RF harmonization approaches. First, in each series, we computed the Z-score of liver measurements based on those computed in the spleen. Second, we evaluated the ComBat method according to each imaging center; parameters were computed in the spleen and applied to the liver. We compared RF distributions and classification performances before/after harmonization. We classified NDL versus spleen and versus DL tissues. The RF distributions were all different between liver and spleen (p < 0.05). The Z-score harmonization outperformed for the detection of liver versus spleen: AUC = 93.1% (p < 0.001). For the detection of DL versus NDL, in a case/control setting, we found no differences between the harmonizations: mean AUC = 73.6% (p = 0.49). Using the whole datasets, the performances were improved using ComBat (p = 0.05) AUC = 82.4% and degraded with Z-score AUC = 67.4% (p = 0.008). Data harmonization requires to first focus on data structuring to not degrade the performances of subsequent classifications. Liver tissue classification after harmonization of spleen-based RF is a promising strategy for improving the detection of DL tissue. • Variability of acquisition parameter makes radiomics of CT features non-reproducible. • Data harmonization can help circumvent the inter-site variability of acquisition protocols. • Inter-site harmonization must be carefully implemented and requires designing consistent data sets.

Shear wave imaging and classification using extended Kalman filter and decision tree algorithm.

Shear wave ultrasound elastography is a quantitative imaging approach in soft tissues based on viscosity-elastic properties. Complex shear modulus (CSM) estimation is an effective solution to analyze tissues' physical properties for elasticity and viscosity based on the wavenumber and attenuation coefficient. CSM offers a way to detect and classify some types of soft tissues. However, CSM-based elastography inherits some obstacles, such as estimation precision and calculation complexity. This work proposes an approach for two-dimensional CSM estimation and soft tissue classification using the Extended Kalman Filter (EKF) and Decision Tree (DT) algorithm, named the EKF-DT approach. CSM estimation is obtained by applying EKF to exploit shear wave propagation at each spatial point. Afterward, the classification of tissues is done by a direct and efficient decision tree algorithm categorizing three types of normal, cirrhosis, and fibrosis liver tissues. Numerical simulation scenarios have been employed to illustrate the recovered quality and practicality of the proposed method's liver tissue classification. With the EKF, the estimated wave number and attenuation coefficient are close to the ideal values, especially the estimated wave number. The states of three liver tissue types were automatically classified by applying the DT coupled with two proposed thresholds of elasticity and viscosity: (2.310 kPa, 1.885 Pa.s) and (3.620 kPa 3.146 Pa.s), respectively. The proposed method shows the feasibility of CSM estimation based on the wavenumber and attenuation coefficient by applying the EKF. Moreover, the DT can automate the classification of liver tissue conditions by proposing two thresholds. The proposed EKF-DT method can be developed by 3D image reconstruction and empirical data before applying it in medical practice.

Automatic Classification Algorithm for Diffused Liver Diseases Based on Ultrasound Images

Diffuse liver diseases such as fatty liver and cirrhosis, are leading causes of disability and fatality across the world. Early diagnosis of these diseases is extremely important to save lives and improve the effectiveness of treatment. This study proposes a non-invasive method for diagnosing liver diseases using ultrasound images, by classifying liver tissue as normal, steatosis, or cirrhosis, using feature extraction, feature selection, and classification. First, the correlation, homogeneity, variance, entropy, contrast, energy, long run emphasis, run percentage, and standard deviation are determined. Second, the most efficient features are selected based on the Fisher discriminant and manual selection methods. Third, three voting-based sub-classifiers are used, namely, the normal/steatosis, normal/cirrhosis, and steatosis/cirrhosis classifiers. The final liver tissue classification is based on the majority function. Our classification method provides two key contributions: combination of two different feature selection methods, avoiding the limitations of each method while benefiting from their strengths; and classifier categorization into three sub-classifiers, where the overall classification is based on the decision of each individual sub-classifier. We obtained recognition accuracies for the normal/steatosis, normal/cirrhosis, and steatosis/cirrhosis classifiers as 95%, 95.74%, and 94.23%, respectively, and an overall recognition accuracy of 95%, which outperforms other methods.

Layer Embedding Analysis in Convolutional Neural Networks for Improved Probability Calibration and Classification.

In this project, our goal is to develop a method for interpreting how a neural network makes layer-by-layer embedded decisions when trained for a classification task, and also to use this insight for improving the model performance. To do this, we first approximate the distribution of the image representations in these embeddings using random forest models, the output of which, termed embedding outputs, are used for measuring how the network classifies each sample. Next, we design a pipeline to use this layer embedding output to calibrate the original model output for improved probability calibration and classification. We apply this two-steps method in a fully convolutional neural network trained for a liver tissue classification task on our institutional dataset that contains 20 3D multi-parameter MR images for patients with hepatocellular carcinoma, as well as on a public dataset with 131 3D CT images. The results show that our method is not only able to provide visualizations that are easy to interpret, but that the embedded decision-based information is also useful for improving model performance in terms of probability calibration and classification, achieving the best performance compared to other baseline methods. Moreover, this method is computationally efficient, easy to implement, and robust to hyper-parameters.

Liver tissue classification of en face images by fractal dimension-based support vector machine.

Full-field optical coherence tomography (FF-OCT) has been reported with its label-free subcellular imaging performance. To realize quantitive cancer detection, the support vector machine model of classifying normal and cancerous human liver tissue is proposed with en face tomographic images. Twenty samples (10 normal and 10 cancerous) were operated from humans and composed of 285 en face tomographic images. Six histogram features and one proposed fractal dimension parameter that reveal the refractive index inhomogeneities of tissue were extracted and made up the training set. The other different 16 samples (8 normal and 8 cancerous) were imaged (190 images) and employed as the test set with the same features. First, a subcellular-resolution tomographic image library for four histopathological areas in liver tissue was established. Second, the area under the receiver operating characteristics of 0.9378, 0.9858, 0.9391, 0.9517 for prediction of the cancerous hepatic cell, central vein, fibrosis, and portal vein were measured with the test set. The results indicate that the proposed classifier from FF-OCT images shows promise as a label-free assessment of quantified tumor detection, suggesting the fractal dimension-based classifier could aid clinicians in detecting tumor boundaries for resection in surgery in the future.

Liver Tissue Classification Using an Auto-context-based Deep Neural Network with a Multi-phase Training Framework.

In this project, our goal is to classify different types of liver tissue on 3D multi-parameter magnetic resonance images in patients with hepatocellular carcinoma. In these cases, 3D fully annotated segmentation masks from experts are expensive to acquire, thus the dataset available for training a predictive model is usually small. To achieve the goal, we designed a novel deep convolutional neural network that incorporates auto-context elements directly into a U-net-like architecture. We used a patch-based strategy with a weighted sampling procedure in order to train on a sufficient number of samples. Furthermore, we designed a multi-resolution and multi-phase training framework to reduce the learning space and to increase the regularization of the model. Our method was tested on images from 20 patients and yielded promising results, outperforming standard neural network approaches as well as a benchmark method for liver tissue classification.

Liver tissue classification in patients with hepatocellular carcinoma by fusing structured and rotationally invariant context representation.

This work addresses multi-class liver tissue classification from multi-parameter MRI in patients with hepatocellular carcinoma (HCC), and is among the first to do so. We propose a structured prediction framework to simultaneously classify parenchyma, blood vessels, viable tumor tissue, and necrosis, which overcomes limitations related to classifying these tissue classes individually and consecutively. A novel classification framework is introduced, based on the integration of multi-scale shape and appearance features to initiate the classification, which is iteratively refined by augmenting the feature space with both structured and rotationally invariant label context features. We study further the topic of rotationally invariant label context feature representations, and introduce a method for this purpose based on computing the energies of the spherical harmonic decompositions computed at different frequencies and radii. We test our method on full 3D multi-parameter MRI volumes from 47 patients with HCC and achieve promising results.

The Identification and Differentiation of Secondary Colorectal Cancer in Human Liver Tissue Using X-ray Fluorescence, Coherent Scatter Spectroscopy, and Multivariate Analysis

Secondary colorectal liver cancer is the most widespread malignancy in patients with colorectal cancer. The aim of this study is to identify and differentiate between normal liver tissue and malignant secondary colorectal liver cancer tissue using X-ray scattering and X-ray fluorescence spectroscopy to investigate the best combination of data that can be used to enable classification of these two tissue types. X-ray fluorescence (XRF) and coherent scatter data were collected for 24 normal and 24 tumor matched pair tissue samples. The levels of 12 elements (P, S, K, Ca, Cr, Fe, Cu, Zn, As, Se, Br, and Rb) were measured in all samples. When comparisons were made between normal and tumor tissues, statistically significant differences were determined for K (p = 0.046), Ca (p = 0.040), Cr (p = 0.011), Fe, Cu, Zn, Br, and Rb (p < 0.01). However, for P, S, As, and Se, no statistically significant differences were found (p > 0.05). For the coherent scatter spectra collected, three peaks due to adipose, fibrous content, and water content of tissue were observed. The amplitude, full width half-maximum, and area under both fibrous content and water content peaks were found to be significantly higher in secondary colorectal liver tumors compared with surrounding normal liver tissue (p < 0.05). However, no significant differences were found for the adipose peak parameters (p > 0.05). Soft independent modeling of class analogy was performed using the XRF, coherent scatter, and elemental ratio data separately, and the accuracy of the classification of 20 unknown samples was found to be 50, 30, and 80%, respectively. Further analysis has shown that using a combination of the XRF and coherent scatter data in a single combined model gave improved normal and tumor liver tissue classification, with an accuracy that was found to be 85%.

Differential diagnosis of CT focal liver lesions using texture features, feature selection and ensemble driven classifiers

The aim of the present study is to define an optimally performing computer-aided diagnosis (CAD) architecture for the classification of liver tissue from non-enhanced computed tomography (CT) images into normal liver (C1), hepatic cyst (C2), hemangioma (C3), and hepatocellular carcinoma (C4). To this end, various CAD architectures, based on texture features and ensembles of classifiers (ECs), are comparatively assessed. Number of regions of interests (ROIs) corresponding to C1-C4 have been defined by experienced radiologists in non-enhanced liver CT images. For each ROI, five distinct sets of texture features were extracted using first order statistics, spatial gray level dependence matrix, gray level difference method, Laws' texture energy measures, and fractal dimension measurements. Two different ECs were constructed and compared. The first one consists of five multilayer perceptron neural networks (NNs), each using as input one of the computed texture feature sets or its reduced version after genetic algorithm-based feature selection. The second EC comprised five different primary classifiers, namely one multilayer perceptron NN, one probabilistic NN, and three k-nearest neighbor classifiers, each fed with the combination of the five texture feature sets or their reduced versions. The final decision of each EC was extracted by using appropriate voting schemes, while bootstrap re-sampling was utilized in order to estimate the generalization ability of the CAD architectures based on the available relatively small-sized data set. The best mean classification accuracy (84.96%) is achieved by the second EC using a fused feature set, and the weighted voting scheme. The fused feature set was obtained after appropriate feature selection applied to specific subsets of the original feature set. The comparative assessment of the various CAD architectures shows that combining three types of classifiers with a voting scheme, fed with identical feature sets obtained after appropriate feature selection and fusion, may result in an accurate system able to assist differential diagnosis of focal liver lesions from non-enhanced CT images.

A STUDY OF ULTRASONIC LIVER IMAGES CLASSIFICATION WITH ARTIFICIAL NEURAL NETWORKS BASED ON FRACTAL GEOMETRY AND MULTIRESOLUTION ANALYSIS

In this study, we evaluate the accuracy of classifiers for classification of ultrasonic liver tissues. Two different statistic classifiers and three various artificial neural networks are included: Bayes classifier, k-nearest neighbor classifier, Back-propagation neural networks, probabilistic neural network and modified probabilistic neural network. These five different classifiers were investigated to determine their ability to classify various categories of ultrasonic liver images. The investigation was performed on the basis of the same feature vector. For statistic classifiers the classification accuracy is at most 90.7% and with artificial neural networks the accuracy is at least 92%. The experimental results illustrated that artificial neural networks are an attractive alternative to conventional statistic techniques when dealing with classification task. Moreover, the feature vector based on fractal geometry and wavelet transform can provide good discriminant ability for ultrasonic liver images under study.

Bounded-depth threshold circuits for computer-assisted CT image classification

We present a stochastic algorithm that computes threshold circuits designed to discriminate between two classes of computed tomography (CT) images. The algorithm employs a partition of training examples into several classes according to the average grey scale value of images. For each class, a sub-circuit is computed, where the first layer of the sub-circuit is calculated by a new combination of the Perceptron algorithm with a special type of simulated annealing. The algorithm is evaluated for the case of liver tissue classification. A depth-five threshold circuit (with pre-processing: depth-seven) is calculated from 400 positive (abnormal findings) and 400 negative (normal liver tissue) examples. The examples are of size n=14,161 (119 ×119) with an 8 bit grey scale. On test sets of 100 positive and 100 negative examples (all different from the learning set) we obtain a correct classification close to 99%. The total sequential run-time to compute a depth-five circuit is about 75 h up to 230 h on a SUN Ultra 5/360 workstation, depending on the width of the threshold circuit at depth-three. In our computational experiments, the depth-five circuits were calculated from three simultaneous runs for depth-four circuits. The classification of a single image is performed within a few seconds.

Liver tissue classification by bounded-depth threshold circuits

We present a new stochastic learning algorithm and first results of computational experiments on fragments of liver CT images. The algorithm is designed to compute a depth-three threshold circuit, where the first layer is calculated by an extension of the Perceptron algorithm by a special type of simulated annealing. The fragments of CT images are of size 119×119 with eight bit grey levels. From 348 positive (focal liver tumours) and 348 negative examples a number of hypotheses of the type w1x1+⋯+wnxn≥ϑ were calculated for n=14161. The threshold functions at levels two and three were determined by computational experiments. The circuit was tested on various sets of 50+50 additional positive and negative examples. For depth-three circuits, we obtained a correct classification of about 97%. The input to the algorithm is derived from the DICOM standard representation of CT images. The simulated annealing procedure employs a logarithmic cooling schedule c(k)=Γ/ln(k+2), where Γ is a parameter that depends on the underlying configuration space. In our experiments, the parameter Γ is chosen according to estimations of the maximum escape depth from local minima of the associated energy landscape.

The analysis and classification of small-scale tissue structures using the generalized spectrum

Conventional ultrasonic imaging systems primarily use backscattered signals for creating qualitative images that reveal large-scale structures, such as tissue boundaries. Efforts to extract additional quantitative information have resulted in limited success. The nonstationarities of the scatterers comprising biological tissue often violate conditions for applying common signal characterization and estimation methods. In addition, ultrasonic tissue properties (such as attenuation, velocity, scatterer density, size, and structure) ambiguously encode information into the backscattered signal, making it difficult or impossible to extract and quantify a single property. This paper presents the generalized spectrum (GS) as a method for analyzing and quantifying the properties of small-scale resolvable structures (on the order of 1 to 4 mm), which result from tissue structures such as lobules, ducts, and vessels. The GS extends the capabilities of power spectral density to include meaningful phase information that results from small-scale structure. The relevant properties of the GS include its ability to reduce the effects’ diffuse scatterers (speckle), and permit normalization schemes that significantly limit the effects of system response and attenuation from the overlying tissue. An implementation of the GS in an analysis and classification of normal and metastic liver tissues is also described.

In Vivo Study of Online Liver Tissue Classification Based on Envelope Power Spectrum Analysis

An ultrasonic imaging and signal analysis system, combining a 3.5 MHz linear array B-scanner and an array processor, has been realized for a clinical liver tissue classification study. Based on the complete unprocessed rf data set of a real time B-image, local envelope power spectra averaged over interactively defined regions are computed. These spectra are partially corrected for several tissue type independent effects that affect comparability. A library containing the in vivo data of 65 patients with known liver tissue states in three groups (normal, cirrhotic, fat) was built up. An algorithm based on numerical optimization of parametrized spectrum characteristics combined with a k-nearest-neighbors classifier, applied to the library, leads to correct classification rates of 85.8 to 87.5 percent in discriminating pathological from normal tissue states. The method can be used for k-weighted color coded tissue type imaging with acceptable spatial resolution.

In vivo study of online liver tissue classification based on envelope power spectrum analysis.

An ultrasonic imaging and signal analysis system, combining a 3.5 MHz linear array B-scanner and an array processor, has been realized for a clinical liver tissue classification study. Based on the complete unprocessed rf data set of a real time B-image, local envelope power spectra averaged over interactively defined regions are computed. These spectra are partially corrected for several tissue type independent effects that affect comparability. A library containing the in vivo data of 65 patients with known liver tissue states in three groups (normal, cirrhotic, fat) was built up. An algorithm based on numerical optimization of parametrized spectrum characteristics combined with a k-nearest-neighbors classifier, applied to the library, leads to correct classification rates of 85.8 to 87.5 percent in discriminating pathological from normal tissue states. The method can be used for k-weighted color coded tissue type imaging with acceptable spatial resolution.

Color image analysis for liver tissue classification

Automatic tissue characterization systems are in great demand by pathologists. However, the existing methods are either too simple to classify a complicated liver tissue image or are dependent on heavy human intervention and very time consuming. We have developed a highly parallel and effective system based on color image segmentation to analyze liver tissue images. To simplify the tissue classification problem, the system first utilizes the achromatic information (the intensity) to segment the tissue image coarsely, then makes use of the chromatic information to classify the segmented regions into four different tissue classes. Thus, the proposed method includes an unsupervised probabilistic relaxation segmentation process and a supervised Bayes classification process. Because the invariant gray level and color properties of the liver tissue image are fully utilized, the difficult classification problem can be fulfilled well at a reasonable computational cost. The proposed method also shows reliable liver tissue classification results from different test sample sets.

Multi-threshold dimension vector for texture analysis and its application to liver tissue classification

A new feature set derived from the fractal geometry, called the multi-threshold dimension vector (MTDV), is proposed for texture analysis. The major properties of the approach are: (i) it is easy to compute; (ii) it is geometrically invariant, i.e. it is invariant under the transformations of translation, rotation, reflection, and scaling; and (iii) it is invariant under the linear gray-level transformation. These properties have enhanced the practical applications of the features. The MTDV is applied to the classification of ultrasonic liver images and produces about 88% of the correct classification rate. This suggests that the MTDV is an excellent tool for medical image processing.

Two‐stage liver tissue classification through fractal geometry

Abstract A fractal analysis procedure for the classification of ultrasonic images of liver tissue is proposed. Based on the two properties of fractals (self‐similarity and self‐affinity), it is possible to recognize not only normal and abnormal liver tissues but also different kinds of abnormalities. Two‐stage procedures are performed in the classification. In the first stage, the mass density feature vectors are estimated by the property of self‐similarity, classification of normal and abnormal ultrasonic liver tissues can be obtained from the differences between their mass density feature vectors. In the second stage, based on the property of self‐affinity, the fractal Brownian motion model is adopted to represent the imaging surface. By evaluating the fractal Brownian feature vectors in about 27 abnormal ultrasonic liver images, it is found that it is possible to classify different kinds of diseases using the model. Moreover, the measurement of lacunarities can be added to improve the method to classify tissues more accurately.

In vivo study of quasi real-time liver tissue classification based on power spectrum analysis

In vivo study of quasi real-time liver tissue classification based on power spectrum analysis