Automated Analysis and Classification of Histological Tissue Features by Multi-Dimensional Microscopic Molecular Profiling

Multiplex immunohistochemistry and high-throughput image analysis for evaluation of spatial tumor immune cell markers in human breast cancer

BACKGROUND

The clinicopathological significance of spatial tumor-infiltrating lymphocytes (TILs) subpopulations is not well studied due to lack of high-throughput scalable methodology for studies with large human sample sizes.

OBJECTIVE

Establishing a cyclic fluorescent multiplex immunohistochemistry (mIHC/IF) method coupled with computer-assisted high-throughput quantitative analysis to evaluate associations of six TIL markers (CD3, CD8, CD20, CD56, FOXP3, and PD-L1) with clinicopathological factors of breast cancer.

METHODS

Our 5-plex mIHC/IF staining was shown to be reliable and highly sensitive for labeling three biomarkers per tissue section. Through repetitive cycles of 5-plex mIHC/IF staining, more than 12 biomarkers could be detected per single tissue section. Using open-source software CellProfiler, the measurement pipelines were successfully developed for high-throughput multiplex evaluation of intratumoral and stromal TILs.

RESULTS

In analyses of 188 breast cancer samples from the Nashville Breast Health Study, high-grade tumors showed significantly increased intratumoral CD3+CD8+ CTL density (P= 0.0008, false discovery rate (FDR) adjusted P= 0.0168) and intratumoral PD-L1 expression (P= 0.0061, FDR adjusted P= 0.0602) compared with low-grade tumors.

CONCLUSIONS

The high- and low-grade breast cancers exhibit differential immune responses which may have clinical significances. The multiplexed imaging quantification strategies established in this study are reliable, cost-efficient and applicable in regular laboratory settings for high-throughput tissue biomarker studies, especially retrospective and population-based studies using archived paraffin tissues.

Automated Quantitative Image Evaluation of Antigen Retrieval Methods for 17 Antibodies in Placentation and Implantation Diagnostic and Research

Immunostaining in clinical routine and research highly depends on standardized staining methods and quantitative image analyses. We qualitatively and quantitatively compared antigen retrieval methods (no pretreatment, pretreatment with pepsin, and heat-induced pretreatment with pH 6 or pH 9) for 17 antibodies relevant for placenta and implantation diagnostics and research. Using our newly established, comprehensive automated quantitative image analysis approach, fluorescent signal intensities were evaluated. Automated quantitative image analysis found that 9 out of 17 antibodies needed antigen retrieval to show positive staining. Heat induction proved to be the most efficient form of antigen retrieval. Eight markers stained positive after pepsin digestion, with β-hCG and vWF showing enhanced staining intensities. To avoid the misinterpretation of quantitative image data, the qualitative aspect should always be considered. Results from native placental tissue were compared with sections of a placental invasion model based on thermo-sensitive scaffolds. Immunostaining on placentas in vitro leads to new insights into fetal development and maternal pathophysiological pathways, as pregnant women are justifiably excluded from clinical studies. Thus, there is a clear need for the assessment of reliable immunofluorescent staining and pretreatment methods. Our evaluation offers a powerful tool for antibody and pretreatment selection in placental research providing objective and precise results.

Semi-Automated Cell and Tissue Analyses Reveal Regionally Specific Morphological Alterations of Immune and Neural Cells in a Porcine Middle Cerebral Artery Occlusion Model of Stroke

Histopathological analysis of cellular changes in the stroked brain provides critical information pertaining to inflammation, cell death, glial scarring, and other dynamic injury and recovery responses. However, commonly used manual approaches are hindered by limitations in speed, accuracy, bias, and the breadth of morphological information that can be obtained. Here, a semi-automated high-content imaging (HCI) and CellProfiler histological analysis method was developed and used in a Yucatan miniature pig permanent middle cerebral artery occlusion (pMCAO) model of ischemic stroke to overcome these limitations. Evaluation of 19 morphological parameters in IBA1⁺ microglia/macrophages, GFAP⁺ astrocytes, NeuN⁺ neuronal, FactorVIII⁺ vascular endothelial, and DCX⁺ neuroblast cell areas was conducted on porcine brain tissue 4 weeks post pMCAO. Out of 19 morphological parameters assessed in the stroke perilesional and ipsilateral hemisphere regions (38 parameters), a significant change in 3838 measured IBA1⁺ parameters, 3438 GFAP⁺ parameters, 3238 NeuN⁺ parameters, 3138 FactorVIII⁺ parameters, and 2838 DCX⁺ parameters were observed in stroked vs. non-stroked animals. Principal component analysis (PCA) and correlation analyses demonstrated that stroke-induced significant and predictable morphological changes that demonstrated strong relationships between IBA1⁺, GFAP⁺, and NeuN⁺ areas. Ultimately, this unbiased, semi-automated HCI and CellProfiler histopathological analysis approach revealed regional and cell specific morphological signatures of immune and neural cells after stroke in a highly translational porcine model. These identified features can provide information of disease pathogenesis and evolution with high resolution, as well as be used in therapeutic screening applications.

Characterizing Immune Responses in Whole Slide Images of Cancer With Digital Pathology and Pathomics

Purpose of Review

Our goal is to show how readily available Pathomics tissue analytics can be used to study tumor immune interactions in cancer. We provide a brief overview of how Pathomics complements traditional histopathologic examination of cancer tissue samples. We highlight a novel Pathomics application, Tumor-TILs, that quantitatively measures and generates maps of tumor infiltrating lymphocytes in breast, pancreatic, and lung cancer by leveraging deep learning computer vision applications to perform automated analyses of whole slide images.

Recent Findings

Tumor-TIL maps have been generated to analyze WSIs from thousands of cases of breast, pancreatic, and lung cancer. We report the availability of these tools in an effort to promote collaborative research and motivate future development of ensemble Pathomics applications to discover novel biomarkers and perform a wide range of correlative clinicopathologic research in cancer immunopathology and beyond.

Summary

Tumor immune interactions in cancer are a fascinating aspect of cancer pathobiology with particular significance due to the emergence of immunotherapy. We present simple yet powerful specialized Pathomics methods that serve as powerful clinical research tools and potential standalone clinical screening tests to predict clinical outcomes and treatment responses for precision medicine applications in immunotherapy.

In Situ Classification of Cell Types in Human Kidney Tissue Using 3D Nuclear Staining

To understand the physiology and pathology of disease, capturing the heterogeneity of cell types within their tissue environment is fundamental. In such an endeavor, the human kidney presents a formidable challenge because its complex organizational structure is tightly linked to key physiological functions. Advances in imaging-based cell classification may be limited by the need to incorporate specific markers that can link classification to function. Multiplex imaging can mitigate these limitations, but requires cumulative incorporation of markers, which may lead to tissue exhaustion. Furthermore, the application of such strategies in large scale 3-dimensional (3D) imaging is challenging. Here, we propose that 3D nuclear signatures from a DNA stain, DAPI, which could be incorporated in most experimental imaging, can be used for classifying cells in intact human kidney tissue. We developed an unsupervised approach that uses 3D tissue cytometry to generate a large training dataset of nuclei images (NephNuc), where each nucleus is associated with a cell type label. We then devised various supervised machine learning approaches for kidney cell classification and demonstrated that a deep learning approach outperforms classical machine learning or shape-based classifiers. Specifically, a custom 3D convolutional neural network (NephNet3D) trained on nuclei image volumes achieved a balanced accuracy of 80.26%. Importantly, integrating NephNet3D classification with tissue cytometry allowed in situ visualization of cell type classifications in kidney tissue. In conclusion, we present a tissue cytometry and deep learning approach for in situ classification of cell types in human kidney tissue using only a DNA stain. This methodology is generalizable to other tissues and has potential advantages on tissue economy and non-exhaustive classification of different cell types.

Automated Identification and Quantification of Signals in Multichannel Immunofluorescence Images: The SignalFinder-IF Platform

Multimarker fluorescence analysis of tissue specimens offers the opportunity to probe the expression levels and locations of multiple markers in a single sample. Software is needed to fully capitalize on the advantages of this technology for sensitive, quantitative, and multiplexed data collection. A major challenge has been the automated identification and quantification of signals. We report on the software SignalFinder-IF, which meets that need. SignalFinder-IF uses a newly developed algorithm called Segment-Fit Thresholding, which showed robust performance for automated signal identification in side-by-side comparisons with several current methods. Two utilities provided with SignalFinder-IF enable downstream analyses. The first allows the quantification and mapping of relationships between an unlimited number of markers through user-defined sequences of AND, OR, and NOT operators. The second produces composite pictures of the signals or colocalization analysis on brightfield hematoxylin and eosin images, which is useful for understanding the morphologies and locations of cells meeting specific marker criteria. SignalFinder-IF enables high-throughput, rigorous analyses of whole-slide, multimarker data, and it promises to open new possibilities in many research and clinical applications.

Unsupervised quantification of tissue immunofluorescence in animal models of multiple sclerosis - Instructions for use

BACKGROUND

In the analysis of animal models of CNS diseases such as experimental autoimmune encephalomyelitis (EAE), immunostaining and histopathology are important readouts. However, the complex morphological features of a tissue staining are often reduced to a single measure which relies on tedious manual planimetry. Furthermore, the measure itself and co-variables such as the region being analysed are chosen in a human decision-making process, which introduces bias.

NEW METHOD

First aim of the present study is to provide an open-source workflow for the high-throughput, unsupervised quantification of different stainings in the spinal cord. We evaluate different EAE models, spinal cord regions and different time points of disease. By applying random forest classification, we compare different measures.

RESULTS

Exemplified for glial reactivity, we show that measures and variables interact and that their values are non-normally distributed, hampering the common use of parametric tests. Furthermore, we demonstrate that one-dimensional measures are insufficient descriptors for immunofluorescence data in EAE and thus need to be considered as partly invalid.

COMPARISON WITH EXISTING METHODS

We show in a systematic analysis of EAE studies that currently published immunohistological outcomes are highly incompatible regarding methodology and statistics. Furthermore, they lack the report of important information necessary for reproducibility and do not use unsupervised automatic analysis.

CONCLUSIONS

Our results discover relevant caveats in the currently used methods of immunofluorescence analysis. The provided step-by-step instructions and open-source code are intended to serve as a framework for sensitive, unbiased immunofluorescence analysis of tissue sections in translational research.

Layer-based visualization and biomedical information exploration of multi-channel large histological data

BACKGROUND AND OBJECTIVE

Modern microscopes can acquire multi-channel large histological data from tissues of human beings or animals, which contain rich biomedical information for disease diagnosis and biological feature analysis. However, due to the large size, fuzzy tissue structure, and complicated multiple elements integrated in the image color space, it is still a challenge for current software systems to effectively calculate histological data, show the inner tissue structures and unveil hidden biomedical information. Therefore, we developed new algorithms and a software platform to address this issue.

METHODS

This paper presents a multi-channel biomedical data computing and visualization system that can efficiently process large 3D histological images acquired from high-resolution microscopes. A novelty of our system is that it can dynamically display a volume of interest and extract tissue information using a layer-based data navigation scheme. During the data exploring process, the actual resolution of the loaded data can be dynamically determined and updated, and data rendering is synchronized in four display windows at each data layer, where 2D textures are extracted from the imaging volume and mapped onto the displayed clipping planes in 3D space.

RESULTS

To test the efficiency and scalability of this system, we performed extensive evaluations using several different hardware systems and large histological color datasets acquired from a CryoViz 3D digital system. The experimental results demonstrated that our system can deliver interactive data navigation speed and display detailed imaging information in real time, which is beyond the capability of commonly available biomedical data exploration software platforms.

CONCLUSION

Taking advantage of both CPU (central processing unit) main memory and GPU (graphics processing unit) graphics memory, the presented software platform can efficiently compute, process and visualize very large biomedical data and enhance data information. The performance of this system can satisfactorily address the challenges of navigating and interrogating volumetric multi-spectral large histological image at multiple resolution levels.

Imitating Pathologist Based Assessment With Interpretable and Context Based Neural Network Modeling of Histology Images

Convolutional neural networks (CNNs) have gained steady popularity as a tool to perform automatic classification of whole slide histology images. While CNNs have proven to be powerful classifiers in this context, they fail to explain this classification, as the network engineered features used for modeling and classification are ONLY interpretable by the CNNs themselves. This work aims at enhancing a traditional neural network model to perform histology image modeling, patient classification, and interpretation of the distinctive features identified by the network within the histology whole slide images (WSIs). We synthesize a workflow which (a) intelligently samples the training data by automatically selecting only image areas that display visible disease-relevant tissue state and (b) isolates regions most pertinent to the trained CNN prediction and translates them to observable and qualitative features such as color, intensity, cell and tissue morphology and texture. We use the Cancer Genome Atlas’s Breast Invasive Carcinoma (TCGA-BRCA) histology dataset to build a model predicting patient attributes (disease stage and node status) and the tumor proliferation challenge (TUPAC 2016) breast cancer histology image repository to help identify disease-relevant tissue state (mitotic activity). We find that our enhanced CNN based workflow both increased patient attribute predictive accuracy (~2% increase for disease stage and ~10% increase for node status) and experimentally proved that a data-driven CNN histology model predicting breast invasive carcinoma stages is highly sensitive to features such as color, cell size, and shape, granularity, and uniformity. This work summarizes the need for understanding the widely trusted models built using deep learning and adds a layer of biological context to a technique that functioned as a classification only approach till now.

Histological coherent Raman imaging: a prognostic review

Histopathology plays a central role in diagnosis of many diseases including solid cancers. Efforts are underway to transform this subjective art to an objective and quantitative science. Coherent Raman imaging (CRI), a label-free imaging modality with sub-cellular spatial resolution and molecule-specific contrast possesses characteristics which could support the qualitative-to-quantitative transition of histopathology. In this work we briefly survey major themes related to modernization of histopathology, review applications of CRI to histopathology and, finally, discuss potential roles for CRI in the transformation of histopathology that is already underway.

Single-Cell Computational Strategies for Lineage Reconstruction in Tissue Systems

Function at the organ level manifests itself from a heterogeneous collection of cell types. Cellular heterogeneity emerges from developmental processes by which multipotent progenitor cells make fate decisions and transition to specific cell types through intermediate cell states. Although genetic experimental strategies such as lineage tracing have provided insights into cell lineages, recent developments in single-cell technologies have greatly increased our ability to interrogate distinct cell types, as well as transitional cell states in tissue systems. From single-cell data that describe these intermediate cell states, computational tools have been developed to reconstruct cell-state transition trajectories that model cell developmental processes. These algorithms, although powerful, are still in their infancy, and attention must be paid to their strengths and weaknesses when they are used. Here, we review some of these tools, also referred to as pseudotemporal ordering algorithms, and their associated assumptions and caveats. We hope to provide a rational and generalizable workflow for single-cell trajectory analysis that is intuitive for experimental biologists.

Multiplexed Single-Cell Imaging: Past, Present, and Future

Jia-Ren Lin from the Laboratory of Systems Pharmacology at Harvard Medical School was awarded best poster at the annual Society of Biomolecular Imaging and Informatics meeting held in Boston, September 2016. His work focuses on single-cell imaging, especially on developing new methods for simultaneously detecting many antigens, named cyclic immunofluorescence (CycIF). This method could be applied in different stages of drug development, from discovery phase, preclinical research to clinical research. The current works and future directions of CycIF method are summarized in the following overview.

Cyclic Immunofluorescence (CycIF), A Highly Multiplexed Method for Single-cell Imaging

Cyclic Immunofluorescence (CycIF) is a public-domain method for performing highly multiplexed immunofluorescence imaging using a conventional epifluorescence microscope. It uses simple reagents and existing antibodies to construct images with up to 30 channels by sequential 4- to 6-channel imaging followed by fluorophore inactivation. Three variant methods are described, the most generally useful of which involves staining fixed cells with antibodies directly conjugated to Alexa Fluor dyes and imaging in four colors, inactivating fluorophores using a mild base in the presence of hydrogen peroxide and light, and then performing another round of staining and imaging. Cell morphology is preserved through multiple rounds of CycIF, and signal-to-noise ratios appear to increase. Unlike antibody-stripping methods, CycIF is gentle and optimized for monolayers of cultured cells. A second protocol involves indirect immunofluorescence and a third enables chemical inactivation of genetically encoded fluorescent proteins, allowing multiplex immunofluorescence to be combined with live-cell analysis of cells expressing fluorescent reporter proteins. © 2016 by John Wiley & Sons, Inc.

Impaired coordination between signaling pathways is revealed in human colorectal cancer using single-cell mass cytometry of archival tissue blocks

Cellular heterogeneity poses a substantial challenge to understanding tissue-level phenotypes and confounds conventional bulk analyses. To analyze signaling at the single-cell level in human tissues, we applied mass cytometry using cytometry time of flight to formalin-fixed, paraffin-embedded (FFPE) normal and diseased intestinal specimens. This technique, called FFPE-DISSECT (disaggregation for intracellular signaling in single epithelial cells from tissue), is a single-cell approach to characterizing signaling states in embedded tissue samples. We applied FFPE-DISSECT coupled to mass cytometry and found differential signaling by tumor necrosis factor-α in intestinal enterocytes, goblet cells, and enteroendocrine cells, implicating the downstream RAS-RAF-MEK pathway in determining goblet cell identity. Application of this technique and computational analyses to human colon specimens confirmed the reduced differentiation in colorectal cancer (CRC) compared to normal colon and revealed increased intratissue and intertissue heterogeneity in CRC with quantitative changes in the regulation of signaling pathways. Specifically, coregulation of the kinases p38 and ERK, the translation regulator 4EBP1, and the transcription factor CREB in proliferating normal colon cells was lost in CRC. Our data suggest that this single-cell approach, applied in conjunction with genomic annotation, enables the rapid and detailed characterization of cellular heterogeneity from clinical repositories of embedded human tissues. This technique can be used to derive cellular landscapes from archived patient samples (beyond CRC) and as a high-resolution tool for disease characterization and subtyping.