Automated classification of inflammation in colon histological sections based on digital microscopy and advanced image analysis

Exploration and Enhancement of Classifiers in the Detection of Lung Cancer from Histopathological Images

Lung cancer is a prevalent malignancy that impacts individuals of all genders and is often diagnosed late due to delayed symptoms. To catch it early, researchers are developing algorithms to study lung cancer images. The primary objective of this work is to propose a novel approach for the detection of lung cancer using histopathological images. In this work, the histopathological images underwent preprocessing, followed by segmentation using a modified approach of KFCM-based segmentation and the segmented image intensity values were dimensionally reduced using Particle Swarm Optimization (PSO) and Grey Wolf Optimization (GWO). Algorithms such as KL Divergence and Invasive Weed Optimization (IWO) are used for feature selection. Seven different classifiers such as SVM, KNN, Random Forest, Decision Tree, Softmax Discriminant, Multilayer Perceptron, and BLDC were used to analyze and classify the images as benign or malignant. Results were compared using standard metrics, and kappa analysis assessed classifier agreement. The Decision Tree Classifier with GWO feature extraction achieved good accuracy of 85.01% without feature selection and hyperparameter tuning approaches. Furthermore, we present a methodology to enhance the accuracy of the classifiers by employing hyperparameter tuning algorithms based on Adam and RAdam. By combining features from GWO and IWO, and using the RAdam algorithm, the Decision Tree classifier achieves the commendable accuracy of 91.57%.

In-depth immunophenotyping reveals significant alteration of lymphocytes in buffalo with brucellosis

Water buffalo (Bubalus bubalis) has a prominent position in the livestock industry worldwide but still suffers from limited knowledge on the mechanisms regulating the immune against infections, including brucellosis (BRC), one of the most significant neglected zoonotic diseases of livestock. Seventy-three buffalo were recruited for the study. Thirty-five were naturally infected with Brucella spp. The aims of the study were to (i) verify the cross-reactivity of 16 monoclonal antibodies (mAbs) developed against human, bovine, and ovine antigens; (ii) evaluate lymphocyte subset alterations in BRC positive buffalo; (iii) evaluate the use of the canonical discriminant analysis (CDA), with flow cytometric data, to discriminate BRC positive from negative animals. A new set of eight mAbs (anti CD3e, CD16, CD18, CD45R0, CD79a; CD172a) were shown to cross-react with water buffalo orthologous molecules. BRC positive animals presented a significant (p < 0.0001) decrease in the percentage of PBMC (29.5 vs. 40.3), total, T and B lymphocytes (23.0 vs. 35.5, 19.2 vs. 28.9, 2.6 vs. 5.7, respectively). In contrast, they showed an increase in percentage of granulocytes (65.2 vs. 55.1; p < 0.0001) and B lymphocytes CD21^neg (22.9 vs. 16.1; p = 0.0067), a higher T/B lymphocyte ratio (10.3 vs. 6.4; p = 0.0011) and CD3⁺ /CD21⁺ (14.7 vs. 8.3; p = 0.0005) ratio. The CDA, applied to 33 different flow cytometric traits, allowed the discrimination of all BRC positive from negative buffalo. Although this is a preliminary study, our results show that flow cytometry can be used in a wide range of applications in livestock diseases, including in support of uncertain BRC diagnoses.

Development of convolutional neural networks for recognition of tenogenic differentiation based on cellular morphology

BACKGROUND AND OBJECTIVE

The use of automated systems for image recognition is highly preferred for regenerative medicine applications to evaluate stem cell differentiation early in the culturing state with non-invasive methodologies instead of invasive counterparts. Bone marrow-derived mesenchymal stem cells (BMSCs) are able to differentiate into desired cell phenotypes, and thereby promise a proper cell source for tendon regeneration. The therapeutic success of stem cell therapy requires cellular characterization prior to the implantation of cells. The foremost problem is that traditional characterization techniques require cellular material which would be more useful for cell therapy, complex laboratory procedures, and human expertise. Convolutional neural networks (CNNs), a class of deep neural networks, have recently made great improvements in image-based classifications, recognition, and detection tasks. We, therefore, aim to develop a potential CNN model in order to recognize differentiated stem cells by learning features directly from image data of unlabelled cells.

METHODS

The differentiation of bone marrow mesenchymal stem cells (BMSCs) into tenocytes was induced with the treatment of bone morphogenetic protein-12 (BMP-12). Following the treatment and incubation step, the phase-contrast images of cells were obtained and immunofluorescence staining has been applied to characterize the differentiated state of BMSCs. CNN models were developed and trained with the phase-contrast cell images. The comparison of CNN models was performed with respect to prediction performance and training time. Moreover, we have evaluated the effect of image enhancement method, data augmentation, and fine-tuning training strategy to increase classification accuracy of CNN models. The best model was integrated into a mobile application.

RESULTS

All the CNN models can fit the biological data extracted from immunofluorescence characterization. CNN models enable the cell classification with satisfactory accuracies. The best result in terms of accuracy and training time is achieved by the model proposed based on Inception-ResNet V2 trained from scratch using image enhancement and data augmentation strategies (96.80%, 434.55 sec).

CONCLUSION

Our study reveals that the CNN models show good performance by identifying stem cell differentiation. Importantly this technique provides a faster and real-time tool in comparison to traditional methods enabling the adjustment of culture conditions during cultivation to improve the yield of therapeutic stem cells.

Automated Classification of Lung Cancer Types from Cytological Images Using Deep Convolutional Neural Networks

Lung cancer is a leading cause of death worldwide. Currently, in differential diagnosis of lung cancer, accurate classification of cancer types (adenocarcinoma, squamous cell carcinoma, and small cell carcinoma) is required. However, improving the accuracy and stability of diagnosis is challenging. In this study, we developed an automated classification scheme for lung cancers presented in microscopic images using a deep convolutional neural network (DCNN), which is a major deep learning technique. The DCNN used for classification consists of three convolutional layers, three pooling layers, and two fully connected layers. In evaluation experiments conducted, the DCNN was trained using our original database with a graphics processing unit. Microscopic images were first cropped and resampled to obtain images with resolution of 256 × 256 pixels and, to prevent overfitting, collected images were augmented via rotation, flipping, and filtering. The probabilities of three types of cancers were estimated using the developed scheme and its classification accuracy was evaluated using threefold cross validation. In the results obtained, approximately 71% of the images were classified correctly, which is on par with the accuracy of cytotechnologists and pathologists. Thus, the developed scheme is useful for classification of lung cancers from microscopic images.

计算机深度学习与智能图像诊断对胃高分化腺癌病理诊断的价值

随着计算机技术的发展, 机器学习被深入研究并应用到各个领域, 机器学习在医学中的应用将转换现在的医学模式, 利用机器学习处理医学中庞大数据可提高医生诊断准确率, 指导治疗, 评估预后. 机器学习中的深度学习已广泛应用在病理智能图像诊断方面, 目前在有丝分裂检测, 细胞核的分割和检测, 组织分类中已取得较好成效. 在病理组织学上, 胃高分化腺癌因其组织结构和细胞形态异型性小, 取材标本表浅等原因容易漏诊. 现有的早期胃癌的病理智能图像诊断系统中没有关于腺腔圆度的研究, 圆度测量可以将腺腔结构的不规则, 腺腔扩张等特征转换为具体数值的定量指标, 通过数值大小来进行诊断分析, 为病理诊断提供参考价值.

Staining correction in digital pathology by utilizing a dye amount table

The stained colors of the tissue components are popularly used as features for image analysis. However, variations in the staining condition of the histology slides prompt variations to the color distribution of the stained tissue samples which could impact the accuracy of the analysis. In this paper, we present a method to correct the staining condition of a histology image. In the method, a look-up table (LUT) based on the dye amounts absorbed by the sample is built. The LUT can be built when either (i) the source and reference staining conditions are specified or (ii) when the user simply wants to recreate his/her preferred staining condition without specifying any reference slide. The effectiveness of the present method was evaluated in two aspects: (i) CIELAB color difference of nuclei, cytoplasm, and red blood cells, between the ten different slides of liver tissue, and (ii) classification of the different tissue components. Application of the present staining correction method reduced the color difference between the slides by an average factor of 9.8 and the classification performance of a linear discriminant classifier improved by 16.5% on the average. Results of the paired t test statistical analysis further showed that the reduction in the CIELAB color difference between the slides and the improvement in the classifier's performance when staining correction was implemented is significant at p < 0.001.

Shape in migration: quantitative image analysis of migrating chemoresistant HCT-8 colon cancer cells

Unsuccessful cytotoxic anticancer treatments may contribute to tumor morphologic instability and consequent tissue invasion, promoting the selection of a more malignant phenotype. Indeed, morphological changes have been demonstrated to be more pronounced in strongly vs. weakly metastatic cells. By means of normalized bending energy, we have previously quantitatively defined the link between cell shape modifications and the acquisition of a more malignant phenotype by 5-FU-resistant colon cancer cells (HCT-8FUres). Such changes were significantly correlated with an increase in motility speed. Herein, we propose a method to quantitatively analyze the shape of wild and chemoresistant HCT-8 migration front cells during wound healing assay. We evaluated the reliability of parameters (area/perimeter ratio [A/p], circularity, roundness, fractal dimension, and solidity) in describing the biological behavior of the two cell lines, enabling hence in distinguishing the chemoresistant line from the other one. We found solidity index the parameter that better described the difference between chemoresistant and wild cells. Moreover, solidity is able to capture the differences between chemoresistant and wild cells at each time point of the migration process. Indeed, motility speed was found to be inversely correlated with solidity, a quantitative index of cell deformability. Deformability is an outstanding hallmark of the process leading to metastatic spread; consequently, solidity may be considered a marker of acquired metastatic property.

Probe-based confocal laser endomicroscopy: a new method for quantitative analysis of pit structure in healthy and Crohn's disease patients

BACKGROUND

Probe-based confocal laser endomicroscopy enables microscopic examination of the digestive mucosa.

AIMS

(1) To identify and validate quantitative endomicroscopic criteria for evaluation of the colonic mucosa and (2) to compare these criteria between healthy and Crohn's disease patients in clinical remission.

METHODS

Six healthy controls and ten Crohn's disease patients in clinical remission were included in this prospective study. Methylene blue-stained biopsies of the right colon and corresponding endomicroscopic images were analyzed. Major axis, minor axis, and major axis/minor axis ratio of crypt lumens were quantified.

RESULTS

Quantitative assessment was performed on 21 ± 4 crypt lumens per patient. Major axis/minor axis ratio values measured with endomicroscopy or methylene blue-stained biopsies were linearly correlated (r=0.63, p=0.01). All macroscopically inflamed mucosa had values of major axis/minor axis ratio higher than the median of controls. Interestingly, 50% (3/6) of Crohn's disease patients with macroscopically normal mucosa had also a higher ratio than pooled controls. Histological analysis showed that 6/7 patients with major axis/minor axis ratio superior to 1.7 had microscopic inflammation.

CONCLUSION

Probe-based confocal laser endomicroscopy allows quantitative analysis of colonic pit structure. Endomicroscopic analysis of major axis/minor axis ratio allows the detection of microscopic residual inflammation with greater accuracy than standard endoscopy in Crohn's disease patients in clinical remission.

Recent advances in morphological cell image analysis

This paper summarizes the recent advances in image processing methods for morphological cell analysis. The topic of morphological analysis has received much attention with the increasing demands in both bioinformatics and biomedical applications. Among many factors that affect the diagnosis of a disease, morphological cell analysis and statistics have made great contributions to results and effects for a doctor. Morphological cell analysis finds the cellar shape, cellar regularity, classification, statistics, diagnosis, and so forth. In the last 20 years, about 1000 publications have reported the use of morphological cell analysis in biomedical research. Relevant solutions encompass a rather wide application area, such as cell clumps segmentation, morphological characteristics extraction, 3D reconstruction, abnormal cells identification, and statistical analysis. These reports are summarized in this paper to enable easy referral to suitable methods for practical solutions. Representative contributions and future research trends are also addressed.

Automated colorectal cancer diagnosis for whole-slice histopathology

In this study, we propose a computational diagnosis system for detecting the colorectal cancer from histopathological slices. The computational analysis was usually performed on patch level where only a small part of the slice is covered. However, slice-based classification is more realistic for histopathological diagnosis. The developed method combines both textural and structural features from patch images and proposes a two level classification scheme. In the first level, the patches in slices are classified into possible classes (adenomatous, inflamed, cancer and normal) and the distribution of the patches into these classes is considered as the information representing the slices. Then the slices are classified using a logistic linear classifier. In patch level, we obtain the correct classification accuracies of 94.36% and 96.34% for the cancer and normal classes, respectively. However, in slice level, the accuracies of the 79.17% and 92.68% are achieved for cancer and normal classes, respectively.

Whole slide images for primary diagnostics of gastrointestinal tract pathology: a feasibility study

During the last decade, whole slide images have been used in many areas of pathology such as teaching, research, digital archiving, teleconsultation, and quality assurance testing. However, whole slide images have as yet not much been used for up-front diagnostics because of the lack of validation studies. The aim of this study was, therefore, to test the feasibility of whole slide images for diagnosis of gastrointestinal tract specimens, one of the largest areas of diagnostic pathology. One hundred gastrointestinal tract biopsies and resections that had been diagnosed using light microscopy 1 year before were rediagnosed on whole slide images scanned at ×20 magnification by 5 pathologists (all reassessing their own cases), having the original clinical information available but blinded to their original light microscopy diagnoses. The original light microscopy and whole slide image-based diagnoses were compared and classified as concordant, slightly discordant (without clinical consequences), and discordant. The diagnoses based on light microscopy and the whole slide image-based rediagnoses were concordant in 95% of the cases. Light microscopy and whole slide image diagnosis in the remaining 5% of cases were slightly discordant, none of these were with clinical or prognostic implications. Up-front histopathologic diagnosis of gastrointestinal biopsies and resections can be done on whole slide images.

Introduction A: recent advances in cytometry instrumentation, probes, and methods--review

Cytometric techniques are continually being improved, refined, and adapted to new applications. This chapter briefly outlines recent advances in the field of cytometry with the main focus on new instrumentations in flow and image cytometry as well as new probes suitable for multiparametric analyses. There is a remarkable trend for miniaturizing cytometers, developing label-free and fluorescence-free analytical approaches, and designing "intelligent" probes. Furthermore, new methods for analyzing complex data for extracting relevant information are reviewed.

Improving the visualization and detection of tissue folds in whole slide images through color enhancement

OBJECTIVE

The objective of this paper is to improve the visualization and detection of tissue folds, which are prominent among tissue slides, from the pre-scan image of a whole slide image by introducing a color enhancement method that enables the differentiation between fold and non-fold image pixels.

METHOD

The weighted difference between the color saturation and luminance of the image pixels is used as shifting factor to the original RGB color of the image.

RESULTS

Application of the enhancement method to hematoxylin and eosin (H&E) stained images improves the visualization of tissue folds regardless of the colorimetric variations in the images. Detection of tissue folds after application of the enhancement also improves but the presence of nuclei, which are also stained dark like the folds, was found to sometimes affect the detection accuracy.

CONCLUSION

The presence of tissue artifacts could affect the quality of whole slide images, especially that whole slide scanners select the focus points from the pre-scan image wherein the artifacts are indistinguishable from real tissue area. We have a presented in this paper an enhancement scheme that improves the visualization and detection of tissue folds from pre-scan images. Since the method works on the simulated pre-scan images its integration to the actual whole slide imaging process should also be possible.

Virtual slide telepathology workstation of the future: lessons learned from teleradiology

The clinical reading environment for the 21st century pathologist looks very different than it did even a few short years ago. Glass slides are quickly being replaced by digital "virtual slides," and the traditional light microscope is being replaced by the computer display. There are numerous questions that arise however when deciding exactly what this new digital display viewing environment will be like. Choosing a workstation for daily use in the interpretation of digital pathology images can be a very daunting task. Radiology went digital nearly 20 years ago and faced many of the same challenges so there are lessons to be learned from these experiences. One major lesson is that there is no "one size fits all" workstation so users must consider a variety of factors when choosing a workstation. In this article, we summarize some of the potentially critical elements in a pathology workstation and the characteristics one should be aware of and look for in the selection of one. Issues pertaining to both hardware and software aspects of medical workstations will be reviewed particularly as they may impact the interpretation process.

Automated recognition and counting of the immunoreactive neuroendocrine cells in chronic gastritis (the preliminary study)

The paper presents the designed software CAMI (Computerized Analysis of Microscopic Images) for a digital reconstruction of the diversiform glands seen in chronic inflammatory gastric mucosa, and for automated recognition and quantization of the immunoreactive neuroendocrine (NE) cells appearing within mucosal glands. Digital reconstruction of the individual gastric gland is difficult due to variable shapes of the glandular cross-sections. Fifteen gastric biopsy specimens representing chronic gastritis were stained routinely with H+E and immunohistochemically with 3 NE markers: Chromogranin A, Somatostatin and Serotonin. Two expert pathologists counted manually the NE cells with the light microscope in 4 types of glandular cross-sections: round, short- oblique, long- oblique and longitudinal. The automated counting of the NE cells was performed on the digital images presenting the same microscopic areas which were selected for the manual reading. The first step of image analysis was concerned to the cell extraction and recognition of the cytoplasmic immunoreactivity. The unstained nuclei of the NE cells were spotted by the sequential thresholding algorithm combined with the artificial neural network of Support\Vector Machine (SVM) type. The second step of image analysis comprised reconstruction of the glands. The presumed shape of each gastric gland was defined by the cellular lining of viewed glandular cross-section. The designed algorithm for gland reconstruction was based on the cell masks. The third step of analysis dealt the cell counting. Every recognized gland with the face cells was used for the NE cell evaluation. The results of the automated quantization compared with manual counting results for the number of NE cells showed high concordance in 3 types of glandular cross-sections: round, short- and long- oblique. A difference noticed in the results of the longitudinal glands should be verified in the extended study. The designed software CAMI is more adequate for the gland recognition with an discontinuous gland face seen in the immunohistochemical digital images, which appear to be a difficult problem for the accurate automated analysis of the cellular component of glands.