Generating realistic tissue images with annotations is a challenging task that is important in many computational histopathology applications. Synthetically generated images and annotations are valuable for training and evaluating algorithms in this domain. To address this, we propose an interactive framework generating pairs of realistic colorectal cancer histology images with corresponding glandular masks from glandular structure layouts. The framework accurately captures vital features like stroma, goblet cells, and glandular lumen. Users can control gland appearance by adjusting parameters such as the number of glands, their locations, and sizes. The generated images exhibit good Frechet Inception Distance (FID) scores compared to the state-of-the-art image-to-image translation model. Additionally, we demonstrate the utility of our synthetic annotations for evaluating gland segmentation algorithms. Furthermore, we present a methodology for constructing glandular masks using advanced deep generative models, such as latent diffusion models. These masks enable tissue image generation through a residual encoder-decoder network.
Automatic nuclear instance segmentation is a crucial task in computational pathology as this information is required for extracting cell-based features in down-stream analysis. However, instance segmentation is a challenging task due to the nature of histology images which show large variations and irregularities in nuclei appearances and arrangements. Various deep learning-based methods have tried to tackle these challenges, however, most of these methods fail to segment the nuclei instances in crowded scenes accurately, or they are not fast enough for practical usage. In this paper, we propose a double-stage neural network for nuclear instance segmentation which leverages the power of an interactive model, NuClick, for accurate instance segmentation by replacing the user input with a nuclei detection module, YOLOv5. We optimized the proposed method to be lightweight and fast and show that it can achieve promising results when tested on the largest publicly available nuclei dataset.
Breast cancer is the dominant cancer among women as it accounts for about one-quarter of all cancer cases in females. The digitized images of Hematoxylin and Eosin (H&E) stained slides of breast cancer specimens carry valuable diagnostic information. However, inspecting these slides manually is a non-trivial task prone to subjective interpretation. Digital pathology (DP) and artificial intelligence (AI) open an opportunity for objective interpretation of the image data. It is challenging to automate the segmentation process in the whole slide images due to the visual complexity of tissue appearance without the need for tedious and time-consuming fine annotations. Many algorithms classify the tissue regions into different types instead of segmenting them, as the classification algorithms require coarse annotations that are easier to acquire. In this paper, we propose a new segmentation framework that combines the simple non-iterative clustering algorithm with a standard convolutional neural network (CNN) classifier to segment whole slide images into different tissue types. In addition, a graph-based post-processing step is applied to improve the framework segmentation performance further. The results show promising improvement to the CNN classifier based coarse segmentation, which would give better feasibility to quantify and study tissues’ mutual relationships.
Oral dysplasia is a pre-malignant stage of oral epithelial carcinomas, e.g., oral squamous cell carcinoma, where significant changes in tissue layers and cells can be observed under the microscope. However, malignancy can be reverted or cured using proper medication or surgery if the grade of malignancy is assessed properly. The assessment of correct grade is therefore critical in patient management as it can change the treatment decisions and prognosis for the dysplastic lesion. This assessment is highly challenging due to considerable inter- and intraobserver variability in pathologists’ agreement, which highlights the need for an automated grading system that can predict more accurate and reliable grade. Recent advancements have made it possible for digital pathology (DP) and artificial intelligence (AI) to join forces from the digitization of tissue slides into images and using those images to train and predict more accurate grades using complex AI models. In this regard, we propose a novel morphometric approach exploiting the architectural features in dysplastic lesions i.e., irregular epithelial stratification where we measure the widths of different layers of the epithelium from the boundary layer i.e., keratin projecting inwards to the epithelium and basal layers to the rest of the tissue section from a clinically significant viewpoint.
There are two main types of lung cancer: small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC), which are grouped accordingly due to similarity in behaviour and response to treatment. The main types of NSCLC are lung adenocarcinoma (LUAD), which accounts for about 40% of all lung cancers and lung squamous cell carcinoma (LUSC), which accounts for about 25-30% of all lung cancers. Due to their differences, automated classification of these two main subtypes of NSCLC is a critical step in developing a computer aided diagnostic system. We present an automated method for NSCLC classification, that consists of a two-part approach. Firstly, we implement a deep learning framework to classify input patches as LUAD, LUSC or non-diagnostic (ND). Next, we extract a collection of statistical and morphological measurements from the labeled whole-slide image (WSI) and use a random forest regression model to classify each WSI as lung adenocarcinoma or lung squamous cell carcinoma. This task is part of the Computational Precision Medicine challenge at the MICCAI 2017 conference, where we achieved the greatest classification accuracy with a score of 0.81.
Analysis of tumour cells is essential for morphological characterisation which is useful for disease prognosis and survival prediction. Visual assessment of tumour cell morphology by expert human observers for prognostic purposes is subjective and potentially a tedious process. In this paper, we propose an automated and objective method for tumour cell analysis in whole slide images (WSI) of lung adenocarcinoma. Tumour cells are first extracted at higher magnification and then morphological, texture and spatial distribution features are computed for each cell. We investigated the biological impact of the nuclear features in the context of tumour grading. Results show that some of these features are correlated with tumour grade. We examine some of these features on the WSI where these features shows different distribution depends on the tumour grade.
The assessment of megakaryocytes (MKs) in bone marrow trephine images is an important step in the classification of different subtypes of myeloproliferative neoplasms (MPNs). In general, bone marrow trephine images include several types of cells mixed together, which make it quite difficult to visually identify MKs. In order to aid hematopathologists in the identification and study of MKs, we develop an image processing framework with supervised machine learning approaches and a novel circumscribing active contour model to identify potential MKs and then to accurately delineate the corresponding nucleus and membrane. Specifically, a number of color and texture features are used in a nave Bayesian classifier and an Adaboost classifier to locate the regions with a high probability of depicting MKs. A region-based active contour is used on the candidate MKs to accurately delineate the boundaries of nucleus and membrane. The proposed circumscribing active contour model employs external forces not only based on pixel intensities, but also on the probabilities of depicting MKs as computed by the classifiers. Experimental results suggest that the machine learning approach can detect potential MKs with an accuracy of more than 75%. When our circumscribing active contour model is employed on the candidate MKs, the nucleus and membrane boundaries are segmented with an accuracy of more than 80% as measured by the Dice similarity coefficient. Compared to traditional region-based active contours, the use of additional external forces based on the probability of depicting MKs improves segmentation performance and computational time by an average 5%.
The first step prior to most analyses on most histopathology images is the detection of area of interest. In this work, we present a superpixel-based approach for glandular structure detection in colon histology images. An image is first segmented into superpixels with the constraint on the presence of glandular boundaries. Texture and color information is then extracted from each superpixel to calculate the probability of that superpixel belonging to glandular regions, resulting in a glandular probability map. In addition, we present a novel texture descriptor derived from a region covariance matrix of scattering coefficients. Our approach shows encouraging results for the detection of glandular structures in colon tissue samples.
Stain separation is the process whereby a full colour histology section image is transformed into a series of single channel images, each corresponding to a given stain's expression. Many algorithms in the field of digital pathology are concerned with the expression of a single stain, thus stain separation is a key preprocessing step in these situations. We present a new versatile method of stain separation. The method uses Independent Component Analysis (ICA) to determine a set of statistically independent vectors, corresponding to the individual stain expressions. In comparison to other popular approaches, such as PCA and NNMF, we found that ICA gives a superior projection of the data with respect to each stain. In addition, we introduce a correction step to improve the initial results provided by the ICA coefficients. Many existing approaches only consider separation of two stains, with primary emphasis on Haematoxylin and Eosin. We show that our method is capable of making a good separation when there are more than two stains present. We also demonstrate our method's ability to achieve good separation on a variety of different stain types.
One of the main factors for high workload in pulmonary pathology in developing countries is the relatively large proportion of tuberculosis (TB) cases which can be detected with high throughput using automated approaches. TB is caused by Mycobacterium tuberculosis, which appears as thin, rod-shaped acid-fast bacillus (AFB) in Ziehl-Neelsen (ZN) stained sputum smear samples. In this paper, we present an algorithm for automatic detection of AFB in digitized images of ZN stained sputum smear samples under a light microscope. A key component of the proposed algorithm is the enhancement of raw input image using a novel anisotropic tubular filter (ATF) which suppresses the background noise while simultaneously enhancing strong anisotropic features of AFBs present in the image. The resulting image is then segmented using color features and candidate AFBs are identified. Finally, a support vector machine classifier using morphological features from candidate AFBs decides whether a given image is AFB positive or not. We demonstrate the effectiveness of the proposed ATF method with two different feature sets by showing that the proposed image analysis pipeline results in higher accuracy and F1-score than the same pipeline with standard median filtering for image enhancement.
The recent development of multivariate imaging techniques, such as the Toponome Imaging System (TIS), has facilitated the analysis of multiple co-localisation of proteins. This could hold the key to understanding complex phenomena such as protein-protein interaction in cancer. In this paper, we propose a Bayesian framework for cell level network analysis allowing the identification of several protein pairs having significantly higher co-expression levels in cancerous tissue samples when compared to normal colon tissue. It involves segmenting the DAPI-labeled image into cells and determining the cell phenotypes according to their protein-protein dependence profile. The cells are phenotyped using Gaussian Bayesian hierarchical clustering (GBHC) after feature selection is performed. The phenotypes are then analysed using Difference in Sums of Weighted cO-dependence Profiles (DiSWOP), which detects differences in the co-expression patterns of protein pairs. We demonstrate that the pairs highlighted by the proposed framework have high concordance with recent results using a different phenotyping method. This demonstrates that the results are independent of the clustering method used. In addition, the highlighted protein pairs are further analysed via protein interaction pathway databases and by considering the localization of high protein-protein dependence within individual samples. This suggests that the proposed approach could identify potentially functional protein complexes active in cancer progression and cell differentiation.
We present a method for fast, approximate registration of whole-slide images (WSIs) of histopathology serial sections. Popular histopathology slide registration methods in the existing literature tend towards intensity-based approaches.1, 2 Further input, in the form of an approximate initial transformation to be applied to one of the two WSIs, is then usually required, and this transformation needs to be optimised. Such a transformation is not readily available in this context and thus there is a need for fast approximation of these parameters. Fast registration is achieved by comparison of the external boundaries of adjacent tissue sections, using local curvature on multiple scales to assess similarity. Our representation of curvature is a modified version of the Curvature Scale Space (CSS)3 image. We substitute zero crossings with signed local absolute maxima of curvature to improve the registration's robustness to the subtle morphological differences of adjacent sections. A pairwise matching is made between curvature maxima at scales increasing exponentially, the matching minimizes the distance between maxima pairs at each scale. The boundary points corresponding to the matched maxima
pairs are used to estimate the desired transformation. Our method is highly robust to translation, rotation, and linear scaling, and shows good performance in cases of moderate non-linear scaling. On our set of test images the algorithm shows improved reliability and processing speed in comparison to existing CSS based registration methods.
Wavelet packets are well-known for their ability to compactly represent textures consiting of oscillatory patterns such as fingerprints or striped cloth. In this paper, we report recent work on representing both periodic and granular types of texture using adaptive wavelet basis functions. The discrimination power of a wavelet packet subband can be defined as its ability to differentiate between any two texture classes in the transform domain, consequently leading to better classification results. The problem of adaptive wavelet basis selection for texture analysis can, therefore, be solved by using a dynamic programming approach to find the best basis from a library of orthonormal basis functions with respect to a discriminant measure. We present a basis selection algorithm which extends the concept of 'Local Discrminant Basis' (Saito and Coifman, 1994) to two dimensions. The problem of feature selection is addressed by sorting the features according to their relevance as described by the discriminant measure, which has a significant advantage over other feature selection methods that both basis selection and reduction of dimensionality of the feature space can be done simultaneously. We show that wavelet packets are good at representing not only oscillatory patterns but also granular textures. Comparative results are presented for four different distance metrics: Kullback-Leibler (KL) divergence, Jensen-Shannon (JS) divergence, Euclidean distance, and Hellinger distance. Initial experimental results show that Hellinger and Euclidean distance metrics may perform better as compared to other cost functions.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.