Breast Cancer Histopathology Research Articles

Breast Cancer Image Multi-Classification Using Random Patch Aggregation and Depth-Wise Convolution based Deep-Net Model

Adapting the profound, deep convolutional neural network models for large image classification can result in the layout of network architectures with a large number of learnable parameters and tuning of those varied parameters can considerably grow the complexity of the model. To address this problem a convolutional Deep-Net Model based on the extraction of random patches and enforcing depth-wise convolutions is proposed for training and classification of widely known benchmark Breast Cancer histopathology images. The classification result of these patches is aggregated using majority vote casting in deciding the final image classification type. It has been observed that the proposed Deep-Net model implementation results when compared with classification results of the VGG Net(16 layers) learned features, outclasses in terms of accuracy when applied to breast tumor Histopathology images. The objective of this work is to examine and comprehensively analyze the sub-class classification performance of the proposed model across all optical magnification frontiers.

Breast Cancer Histopathology Image Super-Resolution Using Wide-Attention GAN With Improved Wasserstein Gradient Penalty and Perceptual Loss

In the realm of image processing, enhancing the quality of the images is known as a super-resolution problem (SR). Among SR methods, a super-resolution generative adversarial network, or SRGAN, has been introduced to generate SR images from low-resolution images. As it is of the utmost importance to keep the size and the shape of the images, while enlarging the medical images, we propose a novel super-resolution model with a generative adversarial network to generate SR images with finer details and higher quality to encourage less blurring. By widening residual blocks and using a self-attention layer, our model becomes robust and generalizable as it is able to extract the most important part of the images before up-sampling. We named our proposed model as wide-attention SRGAN (WA-SRGAN). Moreover, we have applied improved Wasserstein with a Gradient penalty to stabilize the model while training. To train our model, we have applied images from Camylon 16 database and enlarged them by $2\times $ , $4\times $ , $8\times $ , and $16\times $ upscale factors with the ground truth of the size of $256\times 256\times 3$ . Furthermore, two normalization methods, including batch normalization, and weight normalization have been applied and we observed that weight normalization is an enabling factor to improve metric performance in terms of SSIM. Moreover, several evaluation metrics, such as PSNR, MSE, SSIM, MS-SSIM, and QILV have been applied for having a comprehensive objective comparison with other methods, including SRGAN, A-SRGAN, and bicubial. Also, we performed the job of classification by using a deep learning model called ResNeXt-101 ( $32\times 8\text{d}$ ) for super-resolution, high-resolution, and low-resolution images and compared the outcomes in terms of accuracy score. Finally, the results on breast cancer histopathology images show the superiority of our model by using weight normalization and a batch size of one in terms of restoration of the color and the texture details.

Learning to segment images with classification labels.

Two of the most common tasks in medical imaging are classification and segmentation. Either task requires labeled data annotated by experts, which is scarce and expensive to collect. Annotating data for segmentation is generally considered to be more laborious as the annotator has to draw around the boundaries of regions of interest, as opposed to assigning image patches a class label. Furthermore, in tasks such as breast cancer histopathology, any realistic clinical application often includes working with whole slide images, whereas most publicly available training data are in the form of image patches, which are given a class label. We propose an architecture that can alleviate the requirements for segmentation-level ground truth by making use of image-level labels to reduce the amount of time spent on data curation. In addition, this architecture can help unlock the potential of previously acquired image-level datasets on segmentation tasks by annotating a small number of regions of interest. In our experiments, we show using only one segmentation-level annotation per class, we can achieve performance comparable to a fully annotated dataset.

Automated segmentation technique with self‐driven post‐processing for histopathological breast cancer images

Automated segmentation of histopathological images is a challenging task to detect cancerous cells in breast tissue. Recent reviews state high accuracy to segment image, but depends on user input, say window area size, time steps, level set, magnification factor and so on. To extract the region of interest effectively, the subject expert performs post-processing operations several times on the segmentation results with different input values for different parameters say, area opening, fill holes and selects most appropriate enhanced image required for further analysis. The authors proposed an automated segmentation technique followed by self-driven post-processing operations to detect cancerous cells effectively. The post-processing method itself determines the value of different parameters for different operations based on segmented results obtained. The proposed technique has the following features: (i) technique is context sensitive; (ii) no prior setting of time step, weighted area coefficient parameters is required; (iii) magnification independent; (iv) post-processing operations are self-driven which enhance segmentation results adaptively. The experimental results are compared with four state-of-the-art techniques: fuzzy C-means, spatial fuzzy C-means, spatial neutrosophic distance regularised level set and convolutional neural network-based PangNet. Experimental results obtained on two publicly available data sets show that the proposed technique outperforms effectively.

Aggregated Deep Global Feature Representation for Breast Cancer Histopathology Image Classification

Convolutional neural networks (CNNs), successfully used in a great number of medical image analysis applications, have also achieved the state-of-the-art performance in breast cancer histopathology image (BCHI) classification problem recently. However, due to the large varieties among within-class images and insufficient data volume, it is still a challenge to obtain more competitive results by using deep CNN models alone. In this paper, we aim to explore the combination of CNN models with a milestone feature representation method in visual tasks, i.e., vector of locally aggregated descriptors (VLAD), for the BCHI classification, and further propose a novel aggregated deep global feature representation (ADGFR) for this problem. ADGFR adopts the deep features that are extracted from the fully connected layer to form an individual descriptor vector, and augments input images to generate different descriptors for achieving the final aggregated descriptor vector. The individual descriptor vector can effectively keep the global features of benign and malignant image, whose discriminability is further reinforced by the aggregate operation, leading to the more discriminant capability of ADGFR for BCHI. Extensive experiments on the public Break His dataset illuminate that our ADGFR can achieve the optimal classification accuracies of 95.05% at image level and 95.50% at patient level, respectively.

HookNet: Multi-resolution convolutional neural networks for semantic segmentation in histopathology whole-slide images.

We propose HookNet, a semantic segmentation model for histopathology whole-slide images, which combines context and details via multiple branches of encoder-decoder convolutional neural networks. Concentric patches at multiple resolutions with different fields of view, feed different branches of HookNet, and intermediate representations are combined via a hooking mechanism. We describe a framework to design and train HookNet for achieving high-resolution semantic segmentation and introduce constraints to guarantee pixel-wise alignment in feature maps during hooking. We show the advantages of using HookNet in two histopathology image segmentation tasks where tissue type prediction accuracy strongly depends on contextual information, namely (1) multi-class tissue segmentation in breast cancer and, (2) segmentation of tertiary lymphoid structures and germinal centers in lung cancer. We show the superiority of HookNet when compared with single-resolution U-Net models working at different resolutions as well as with a recently published multi-resolution model for histopathology image segmentation. We have made HookNet publicly available by releasing the source code1 as well as in the form of web-based applications2,3 based on the grand-challenge.org platform.

Multifaceted fused-CNN based scoring of breast cancer whole-slide histopathology images

Automating the scoring of Whole-Slide Images (WSIs) is a challenging task because the search space for selecting region of interest (ROI) is huge due to the very large sizes of WSIs. A Multifaceted Fused-CNN (MF-CNN) and a Hybrid-Descriptor are proposed to develop an integrated scoring system for Breast Cancer histopathology WSIs. Suitable color and textural features are identified to help mitotic count based selection of ROIs at lower resolution. To recognize complex patterns, the MF-CNN considers multiple facets of the input image. It counts mitoses, extracts handcrafted features from ROIs and utilizes global texture of the images to form a Hybrid-Descriptor for training a classifier assigning scores to WSIs. The proposed system is evaluated on a publicly available benchmark (TUPAC16) and produced the highest score of 0.582 in terms of Cohen’s Kappa. It surpassed human experts’ level accuracy of ROI selection and can therefore reduce the burden of manual ROI selection for WSIs.

340 Poster - Deep learning enables fully automated mitotic density assessment in breast cancer histopathology

Background: Mitosis counting is an important part of breast cancer grading, yet known to suffer from observer variability. Advances in machine learning enable fully automated analysis of digitized glass slides. The present study evaluated automatic mitosis counting and demonstrated applicability on triple negative breast cancers (TNBC).

An ensemble algorithm for breast cancer histopathology image classification

Breast cancer is considered among prime reasons of deaths among women. Early detection can improve the survival rate and can decrease the mortality rate too. Though it’s a difficult and tedious task for pathologists. The motive of this research is to assess the efficiency and accuracy of various deep learning models and ensemble methods. Authors have proposed Stacked Generalized Ensemble algorithm that classifies the images into benign and malignant. The dataset used in experimentation is H&E breast cancer image dataset. Experimental results show that SGE has outperformed on various deep learning single classifiers. The evaluation criteria used for measuring the efficiency of algorithm is accuracy, precision, recall and F1 measure. A comparative analysis has been done with the existing deep learning methods. All the experimental results demonstrate that the Stacked Generalized Ensemble approach performed exponentially good on the histopathological breast cancer image classification and achieved an accuracy of 97.53%.

Contribution of Fluorodeoxyglucose-Positron Emission Tomography Late Imaging to Diagnosis and Correlation with Histopathology in Invasive Lobular Breast Cancer

Contribution of Fluorodeoxyglucose-Positron Emission Tomography Late Imaging to Diagnosis and Correlation with Histopathology in Invasive Lobular Breast Cancer

Tubule Formation Segmentation Of Histopathological Image Of Breast Cancer By Using Clustering Method

Histopathological assessment is one of the examinations that allows the classification of breast cancer based on its level. Histopathological assessment factors are based on tubule formation, nuclear pleomorphism, and the mitotic count. This study only focused on tubule formation. The tubule formation was represented by a lumen surrounded a nucleus. The segmentation of tubule histopathology of breast cancer method was using a combination of k-means clustering and graph cut. The image data used in this study were 15 images of breast cancer histopathology preparations using 5 variations in the number of clusters (k) in the k-means clustering method. The best results of tubule formation segmentation using k = 4, with an average value of balanced accuracy was 81.08% and the most optimal balanced accuracy results was 94.34%.

Breast Cancer Histopathology Image Classification Using an Ensemble of Deep Learning Models.

Breast cancer is one of the major public health issues and is considered a leading cause of cancer-related deaths among women worldwide. Its early diagnosis can effectively help in increasing the chances of survival rate. To this end, biopsy is usually followed as a gold standard approach in which tissues are collected for microscopic analysis. However, the histopathological analysis of breast cancer is non-trivial, labor-intensive, and may lead to a high degree of disagreement among pathologists. Therefore, an automatic diagnostic system could assist pathologists to improve the effectiveness of diagnostic processes. This paper presents an ensemble deep learning approach for the definite classification of non-carcinoma and carcinoma breast cancer histopathology images using our collected dataset. We trained four different models based on pre-trained VGG16 and VGG19 architectures. Initially, we followed 5-fold cross-validation operations on all the individual models, namely, fully-trained VGG16, fine-tuned VGG16, fully-trained VGG19, and fine-tuned VGG19 models. Then, we followed an ensemble strategy by taking the average of predicted probabilities and found that the ensemble of fine-tuned VGG16 and fine-tuned VGG19 performed competitive classification performance, especially on the carcinoma class. The ensemble of fine-tuned VGG16 and VGG19 models offered sensitivity of for carcinoma class and overall accuracy of . Also, it offered an F1 score of . These experimental results demonstrated that our proposed deep learning approach is effective for the automatic classification of complex-natured histopathology images of breast cancer, more specifically for carcinoma images.

Breast cancer histopathology image classification using kernelized weighted extreme learning machine

AbstractHistopathology is considered as the gold standard for diagnosing breast cancer. Traditional machine learning (ML) algorithm provides a promising performance for cancer diagnosis if the training dataset is balanced. Nevertheless, if the training dataset is imbalanced the performance of the ML model is skewed toward the majority class. It may pose a problem for the pathologist because if the benign sample is misclassified as malignant, then a pathologist could make a misjudgment about the diagnosis. A limited investigation has been done in literature for solving the class imbalance problem in computer‐aided diagnosis (CAD) of breast cancer using histopathology. This work proposes a hybrid ML model to solve the class imbalance problem. The proposed model employs pretrained ResNet50 and the kernelized weighted extreme learning machine for CAD of breast cancer using histopathology. The breast cancer histopathological images are obtained from publicly available BreakHis and BisQue datasets. The proposed method achieved a reasonable performance for the classification of the minority as well as the majority class instances. In comparison, the proposed approach outperforms the state‐of‐the‐art ML models implemented in previous studies using the same training‐testing folds of the publicly accessible BreakHis dataset.

Convolutional Neural Network Based Breast Cancer Histopathology Image Classification.

Breast cancer is a global health concern, with approximately 30 million new cases projected to be reported by 2030. While efforts are being channeled into curative measures, preventive and diagnostic measures also need to be improved to curb the situation. Convolutional Neural Networks (CNNs) are a class of deep learning algorithms that have been widely adopted for the computerized classification of breast cancer histopathology images. In this work, we propose a set of training techniques to improve the performance of CNN-based classifiers for breast cancer identification. We combined transfer learning techniques with data augmentation and whole image training to improve the performance of the CNN classifier. Instead of conventional image patch extraction for training and testing, we employed a high-resolution whole-image training and testing on a modified network that was pre-trained on the Imagenet dataset. Despite the computational complexity, our proposed classifier achieved significant improvement over the previously reported studies on the open-source BreakHis dataset, with an average image level accuracy of about 91% and patient scores as high as 95%.Clinical Relevance- this work improves on the performance of CNN for breast cancer histopathology image classification. An improved Breast cancer image classification can be used for the preliminary examination of tissue slides in breast cancer diagnosis.

The Effect of Systemic Chemotherapy on Ovarian Function: A Prospective Clinical Trial

Premenopausal women with breast cancer are at risk of developing ovarian failure after chemotherapy. The aim of this study was to investigate the negative effects of systemic chemotherapy on ovarian function in premenoupausal women with breast cancer. Thirty-one premenopausal women with operable breast cancer aged between 26-48 years were enrolled in this prospective cohort study to investigate preliminary results. Additional 69 patients' data will be included after the completion of all five measurements. The change in serum Antimullerian Hormone (AMH) levels, mean ovarian volumes (MOV) and antral follicle counts (AFCs) at 3-month intervals were recorded to evaluate ovarian function. Women who had at least one pretreatment and four post-treatment measurements in one year follow-up period were included in the study. Decision of chemotherapy regimen was taken by the Tumor Board. Thirty-one patients had all five AMH, MOV and AFCs results. There was a statistically significant negative correlation between 1st - 5th AMH levels (p=0.006) and 1st - 5th AFCs during the follow-up period (p<0.0001). However pre- and post-chemotherapy measurements of MOVs did not demonstrate any significant correlation (p=0.799). BMI, parity, lactation, histopathology and molecular subtypes of breast cancer, alcohol intake, smoking and type of chemotherapy regimen were not significantly correlated with AMH, AFC and MOV. Pretreatment AMH levels and AFC were shown to have a significant role in early prediction of ovarian-reserve after chemotherapy.

Mitosis detection in breast cancer histopathology images using hybrid feature space.

Breast Cancer grading is a challenging task as regards image analysis, which is normally based on mitosis count rate. The mitotic count provides an estimate of aggressiveness of the tumor. The detection of mitosis is a challenging task because in a frame of slides at X40 magnification, there are hundreds of nuclei containing few mitotic nuclei. However, manual counting of mitosis by pathologists is a difficult and time intensive job, moreover conventional method rely mainly on the shape, color, and/or texture features as well as pathologist experience. The objective of this study is to accept the atypaia-2014 mitosis detection challenge, automate the process of mitosis detection and a proposal of a hybrid feature space that provides better discrimination of mitotic and non-mitotic nuclei by combining color features with morphological and texture features. To exploit color channels, they were first selected, and then normalized and cumulative histograms were computed in wavelet domain. A detailed analysis presented on these features in different color channels of respective color spaces using Random Forest (RF) and Support Vector Machine (SVM) classifiers. The proposed hybrid feature space when used with SVM classifier achieved a detection rate of 78.88% and F-measure of 72.07%. Our results, especially high detection rate, indicate that proposed hybrid feature space model contains discriminant information for mitotic nuclei, being therefore a very capable are for exploration to improve the quality of the diagnostic assistance in histopathology.

Invasive Ductal Carcinoma Detection in Histopathology Images Using Machine Learning Techniques Without Deploying Feature Extraction

Breast cancer is one of the common malignant diseases in female all over the world. Microscopic investigation of tissues in breast is essential for analysis of breast cancer. For detection of breast cancer, pathologist uses various magnificent stages for obtaining accurate diagnosis of biopsy images which is time consuming. Development in digital imaging techniques has helped in assessment of pathology images using machine learning and computerized methods which could computerize a few of the pathology stages in the diagnosis of breast cancer. This kind of automation can be helpful in achieving quick and exact results reducing the observers’ inconsistency, thus increasing the accuracy. In this work, a new method is proposed to categorize breast cancer histopathology images. The objective is to evaluate the robustness and accuracy of a classification system based on machine learning, to automatically identify invasive tumor on digitized images without extracting the features. Here, a new method is presented that employs machine learning classifiers for classification of invasive tumor on whole slide images. The accuracy of different classifiers varies from 80% to 85%, leaving scope for improvement. The aim is to gather different researchers in both machine learning and medical field to proceed toward this Computer Aided Diagnosis (CAD) system for classification of invasive ductal carcinoma (IDC).

A new deep convolutional neural network model for classifying breast cancer histopathological images and the hyperparameter optimisation of the proposed model

Deep learning algorithms have yielded remarkable results in medical diagnosis and image analysis, besides their contribution to improvements in a number of fields such as drug discovery, time-series modelling and optimisation methods. With regard to the analysis of histopathologic breast cancer images, the similarity of those images and the presence of healthy and tumourous tissues in different areas complicate the detection and classification of tumours on whole slide images. An accurate diagnosis in a short time is a need for full treatment in breast cancer. A successful classification on breast cancer histopathological images will overcome the burden on the pathologist and reduce the subjectivity of diagnosis. In this study, we propose a deep convolutional neural network model. The model uses various algorithms (i.e., stochastic gradient descent, Nesterov accelerated gradient, adaptive gradient, RMSprop, AdaDelta and Adam) to compute the initial weight of the network and update the model parameters for faster backpropagation learning. In order to train the model with less hardware in a short time, we used the parallel computing architecture with Cuda-enabled graphics processing unit. The results indicate that the deep convolutional neural network model is an effective classification model with a high performance up to 99.05% accuracy value.

Dual Stage Normalization Approach Towards Classification of Breast Cancer

ABSTRACTBreast cancer is a major concern among women that causes high risk of death. Early diagnosis of such cancer becomes challenging due to alterations in the color of the histopathological breast images. This study uses a publicly available dataset of breast cancer histopathology images. This paper introduces a dual stage normalization approach, to address the color variation problem of biopsy specimen collectively caused by incompatible staining in biopsy process and bizarre imaging quality. The dual stage normalization proposed here consists of a stain normalization unit and a light normalization unit. This system addresses the variations of both imaging and staining of specimen that are caused by a microscopic imaging setup. Later on, eight features have been extracted from the normalized images and used for the classification of breast cancer (benign and malignant). The overall accuracy of the back propagation algorithm (BPA) classifier is obtained as 81.8%. After comparison with other classifier accuracies, BPA classifier is found to be acceptable. Recall and precision values are approximately 89% and 90%, respectively, which is acceptable. The saturation-weighted hue statistics produces balanced and uniform color hues for stain normalization. This statistic is powerful against variations in model parameters and unsusceptible to image subjects and achromatic colors. This normalization technique retains all histological data with an enhanced performance.

Artificial Intelligence-Based Mitosis Detection in Breast Cancer Histopathology Images Using Faster R-CNN and Deep CNNs.

Breast cancer is the leading cause of mortality in women. Early diagnosis of breast cancer can reduce the mortality rate. In the diagnosis, the mitotic cell count is an important biomarker for predicting the aggressiveness, prognosis, and grade of breast cancer. In general, pathologists manually examine histopathology images under high-resolution microscopes for the detection of mitotic cells. However, because of the minute differences between the mitotic and normal cells, this process is tiresome, time-consuming, and subjective. To overcome these challenges, artificial-intelligence-based (AI-based) techniques have been developed which automatically detect mitotic cells in the histopathology images. Such AI techniques accelerate the diagnosis and can be used as a second-opinion system for a medical doctor. Previously, conventional image-processing techniques were used for the detection of mitotic cells, which have low accuracy and high computational cost. Therefore, a number of deep-learning techniques that demonstrate outstanding performance and low computational cost were recently developed; however, they still require improvement in terms of accuracy and reliability. Therefore, we present a multistage mitotic-cell-detection method based on Faster region convolutional neural network (Faster R-CNN) and deep CNNs. Two open datasets (international conference on pattern recognition (ICPR) 2012 and ICPR 2014 (MITOS-ATYPIA-14)) of breast cancer histopathology were used in our experiments. The experimental results showed that our method achieves the state-of-the-art results of 0.876 precision, 0.841 recall, and 0.858 F1-measure for the ICPR 2012 dataset, and 0.848 precision, 0.583 recall, and 0.691 F1-measure for the ICPR 2014 dataset, which were higher than those obtained using previous methods. Moreover, we tested the generalization capability of our technique by testing on the tumor proliferation assessment challenge 2016 (TUPAC16) dataset and found that our technique also performs well in a cross-dataset experiment which proved the generalization capability of our proposed technique.