Search

Crucial transitions in cancer-including tumor initiation, local expansion, metastasis, and therapeutic resistance-involve complex interactions between cells within the dynamic tumor ecosystem. Transformative single-cell genomics technologies and spatial multiplex in situ methods now provide an opportunity to interrogate this complexity at unprecedented resolution. The Human Tumor Atlas Network (HTAN), part of the National Cancer Institute (NCI) Cancer Moonshot Initiative, will establish a clinical, experimental, computational, and organizational framework to generate informative and accessible three-dimensional atlases of cancer transitions for a diverse set of tumor types. This effort complements both ongoing efforts to map healthy organs and previous large-scale cancer genomics approaches focused on bulk sequencing at a single point in time. Generating single-cell, multiparametric, longitudinal atlases and integrating them with clinical outcomes should help identify novel predictive biomarkers and features as well as therapeutically relevant cell types, cell states, and cellular interactions across transitions. The resulting tumor atlases should have a profound impact on our understanding of cancer biology and have the potential to improve cancer detection, prevention, and therapeutic discovery for better precision-medicine treatments of cancer patients and those at risk for cancer.

SUMMARY The emergence of megascale single-cell multiplex tissue imaging (MTI) datasets necessitates reproducible, scalable, and robust tools for cell phenotyping and spatial analysis. We developed open-source, graphics processing unit (GPU)-accelerated tools for intensity normalization, phenotyping, and microenvironment characterization. We deploy the toolkit on a human breast cancer (BC) tissue microarray stained by cyclic immunofluorescence and present the first cross-validation of breast cancer cell phenotypes derived by using two different MTI platforms. Finally, we demonstrate an integrative phenotypic and spatial analysis revealing BC subtype-specific features., In brief Burlingame et al. describe a GPU-accelerated workflow for normalization, phenotyping, and spatial analysis of single-cell multiplex tissue imaging (MTI) data. This workflow is deployed on breast cancer (BC) tissues to derive a cell type dictionary, which is validated between MTI platforms. Tissue architecture is used to discriminate between BC subtypes., Graphical Abstract

Emerging multiplexed imaging platforms provide an unprecedented view of an increasing number of molecular markers at subcellular resolution and the dynamic evolution of tumor cellular composition. As such, they are capable of elucidating cell-to-cell interactions within the tumor microenvironment that impact clinical outcome and therapeutic response. However, the rapid development of these platforms has far outpaced the computational methods for processing and analyzing the data they generate. While being technologically disparate, all imaging assays share many computational requirements for post-collection data processing. As such, our Image Analysis Working Group (IAWG), composed of researchers in the Cancer Systems Biology Consortium (CSBC) and Physical Sciences - Oncology Network (PS-ON), convened a workshop on “Computational Challenges Shared by Diverse Imaging Platforms” to characterize these common issues and a follow-up hackathon to implement solutions for a selected subset of them. Here, we delineate these areas that reflect major axes of research within the field, including image registration, segmentation of cells and subcellular structures, and identification of cell types from their morphology. We further describe the logistical organization of these events, believing our lessons learned can aid others in uniting the imaging community around self-identified topics of mutual interest, in designing and implementing operational procedures to address those topics and in mitigating issues inherent in image analysis (e.g., sharing exemplar images of large datasets and disseminating baseline solutions to hackathon challenges through open-source code repositories).

Metastatic progression defines the final stages of tumor evolution and underlies the majority of cancer-related deaths. The heterogeneity in disseminated tumor cell populations capable of seeding and growing in distant organ sites contributes to the development of treatment resistant disease. We recently reported the identification of a novel tumor-derived cell population, circulating hybrid cells (CHCs), harboring attributes from both macrophages and neoplastic cells, including functional characteristics important to metastatic spread. These disseminated hybrids outnumber conventionally defined circulating tumor cells (CTCs) in cancer patients. It is unknown if CHCs represent a generalized cancer mechanism for cell dissemination, or if this population is relevant to the metastatic cascade. Herein, we detect CHCs in the peripheral blood of patients with cancer in myriad disease sites encompassing epithelial and non-epithelial malignancies. Further, we demonstrate that in vivo-derived hybrid cells harbor tumor-initiating capacity in murine cancer models and that CHCs from human breast cancer patients express stem cell antigens, features consistent with the potential to seed and grow at metastatic sites. Finally, we reveal heterogeneity of CHC phenotypes reflect key tumor features, including oncogenic mutations and functional protein expression. Importantly, this novel population of disseminated neoplastic cells opens a new area in cancer biology and renewed opportunity for battling metastatic disease.

Accurate diagnosis of metastatic cancer is essential for prescribing optimal control strategies to halt further spread of metastasizing disease. While pathological inspection aided by immunohistochemistry staining provides a valuable gold standard for clinical diagnostics, deep learning methods have emerged as powerful tools for identifying clinically relevant features of whole slide histology relevant to a tumor’s metastatic origin. Although deep learning models require significant training data to learn effectively, transfer learning paradigms provide mechanisms to circumvent limited training data by first training a model on related data prior to fine-tuning on smaller data sets of interest. In this work we propose a transfer learning approach that trains a convolutional neural network to infer the metastatic origin of tumor tissue from whole slide images of hematoxylin and eosin (H&E) stained tissue sections and illustrate the advantages of pre-training network on whole slide images of primary tumor morphology. We further characterize statistical dissimilarity between primary and metastatic tumors of various indications on patch-level images to highlight limitations of our indication-specific transfer learning approach. Using a primary-to-metastatic transfer learning approach, we achieved mean class-specific areas under receiver operator characteristics curve (AUROC) of 0.779, which outperformed comparable models trained on only images of primary tumor (mean AUROC of 0.691) or trained on only images of metastatic tumor (mean AUROC of 0.675), supporting the use of large scale primary tumor imaging data in developing computer vision models to characterize metastatic origin of tumor lesions.

Metastatic progression defines the final stages of tumor evolution and underlies the majority of cancer-related deaths. The heterogeneity in disseminated tumor cell populations capable of seeding and growing in distant organ sites contributes to the development of treatment resistant disease. We recently reported the identification of a novel tumor-derived cell population, circulating hybrid cells (CHCs), harboring attributes from both macrophages and neoplastic cells, including functional characteristics important to metastatic spread. These disseminated hybrids outnumber conventionally defined circulating tumor cells (CTCs) in cancer patients. It is unknown if CHCs represent a generalized cancer mechanism for cell dissemination, or if this population is relevant to the metastatic cascade. Herein, we detect CHCs in the peripheral blood of patients with cancer in myriad disease sites encompassing epithelial and non-epithelial malignancies. Further, we demonstrate that in vivo-derived hybrid cells harbor tumor-initiating capacity in murine cancer models and that CHCs from human breast cancer patients express stem cell antigens, features consistent with the ability to seed and grow at metastatic sites. Finally, we reveal heterogeneity of CHC phenotypes reflect key tumor features, including oncogenic mutations and functional protein expression. Importantly, this novel population of disseminated neoplastic cells opens a new area in cancer biology and renewed opportunity for battling metastatic disease.Simple SummaryThere is an incomplete understanding of circulating neoplastic cell populations and the fundamental mechanisms that drive dissemination, immune evasion, and growth —all critical information to more effectively prevent and treat cancer progression. A novel disseminated tumor cell population, circulating hybrid cells, are detected across many cancer types and carry functional tumor-initiating properties. Additionally, circulating hybrid cells are found at significantly higher levels than conventionally defined circulating tumor cells. Our study demonstrates that neoplastic hybrid cells harbor phenotypic and genetic characteristics of tumor and immune cells, display stem features, and are a generalizable phenomenon in solid tumors. Circulating hybrid cells therefore have relevance as a novel biomarker and open a new field of study in malignancy.

Deep learning systems have emerged as powerful mechanisms for learning domain translation models. However, in many cases, complete information in one domain is assumed to be necessary for sufficient cross-domain prediction. In this work, we motivate a formal justification for domain-specific information separation in a simple linear case and illustrate that a self-supervised approach enables domain translation between data domains while filtering out domain-specific data features. We introduce a novel approach to identify domainspecific information from sets of unpaired measurements in complementary data domains by considering a deep learning cross-domain autoencoder architecture designed to learn shared latent representations of data while enabling domain translation. We introduce an orthogonal gate block designed to enforce orthogonality of input feature sets by explicitly removing non-sharable information specific to each domain and illustrate separability of domain-specific information on a toy dataset.

SummaryMechanisms of therapeutic resistance manifest in metastatic cancers as tumor cell intrinsic alterations and extrinsic microenvironmental influences that can change during treatment. To support the development of methods for the identification of these mechanisms in individual patients, we present here an Omic and Multidimensional Spatial (OMS) Atlas generated from four serial biopsies of a metastatic breast cancer patient during 3.5 years of therapy. This resource links detailed, longitudinal clinical metadata including treatment times and doses, anatomic imaging, and blood-based response measurements to exploratory analytics including comprehensive DNA, RNA, and protein profiles, images of multiplexed immunostaining, and 2- and 3-dimensional scanning electron micrographs. These data reveal aspects of therapy-associated heterogeneity and evolution of the cancer’s genome, signaling pathways, immune microenvironment, cellular composition and organization, and ultrastructure. We present illustrative examples showing how integrative analyses of these data provide insights into potential mechanisms of response and resistance, and suggest novel therapeutic vulnerabilities.

Spatially-resolved molecular profiling by immunostaining tissue sections is a key feature in cancer diagnosis, subtyping, and treatment, where it complements routine histopathological evaluation by clarifying tumor phenotypes. In this work, we present a deep learning-based method called speedy histological-to-immunofluorescent translation (SHIFT) which takes histologic images of hematoxylin and eosin (H&E)-stained tissue as input, then in near-real time returns inferred virtual immunofluorescence (IF) images that estimate the underlying distribution of the tumor cell marker pan-cytokeratin (panCK). To build a dataset suitable for learning this task, we developed a serial staining protocol which allows IF and H&E images from the same tissue to be spatially registered. We show that deep learning-extracted morphological feature representations of histological images can guide representative sample selection, which improved SHIFT generalizability in a small but heterogenous set of human pancreatic cancer samples. With validation in larger cohorts, SHIFT could serve as an efficient preliminary, auxiliary, or substitute for panCK IF by delivering virtual panCK IF images for a fraction of the cost and in a fraction of the time required by traditional IF.

Recent advances in multiplexed imaging technologies promise to improve the understanding of the functional states of individual cells and the interactions between the cells in tissues. This often requires compilation of results from multiple samples. However, quantitative integration of information between samples is complicated by variations in staining intensity and background fluorescence that obscure biological variations. Failure to remove these unwanted artifacts will complicate downstream analysis and diminish the value of multiplexed imaging for clinical applications. Here, to compensate for unwanted variations, we automatically identify negative control cells for each marker within the same tissue and use their expression levels to infer background signal level. The intensity profile is normalized by the inferred level of the negative control cells to remove between-sample variation. Using a tissue microarray data and a pair of longitudinal biopsy samples, we demonstrated that the proposed approach can remove unwanted variations effectively and shows robust performance., Chang et al. develop an analytical method called RESTORE to control for variations due to technical artifacts in multiplexed imaging. They test their method on a CycIF stained tissue microarray dataset and biopsies processed at different times. Their method can improve the applicability of imaging techniques in diagnostics and inference using unbiased clustering methods.

Pathologists rely on clinical information, tissue morphology, and sophisticated molecular diagnostics to accurately infer the metastatic origin of secondary liver cancer. In this paper, we introduce a deep learning approach to identify spatially localized regions of cancerous tumor within hematoxylin and eosin stained tissue sections of liver cancer and to generate predictions of the cancer’s metastatic origin. Our approach achieves an accuracy of 90.2% when classifying metastatic origin of whole slide images into three distinct classes, which compares favorably to an established clinical benchmark by three board-certified pathologists whose accuracies ranged from 90.2% to 94.1% on the same prediction task. This approach illustrates the potential impact of deep learning systems to leverage morphological and structural features of H&E stained tissue sections to guide pathological and clinical determination of the metastatic origin of secondary liver cancers.

Spatially-resolved molecular profiling by immunostaining tissue sections is a key feature in cancer diagnosis, subtyping, and treatment, where it complements routine histopathological evaluation by clarifying tumor phenotypes. In this work, we present a deep learning-based method called speedy histological-to-immunofluorescent translation (SHIFT) which takes histologic images of hematoxylin and eosin-stained tissue as input, then in near-real time returns inferred virtual immunofluorescence (IF) images that accurately depict the underlying distribution of phenotypes without requiring immunostaining of the tissue being tested. We show that deep learning-extracted feature representations of histological images can guide representative sample selection, which improves SHIFT generalizability. SHIFT could serve as an efficient preliminary, auxiliary, or substitute for IF by delivering multiplexed virtual IF images for a fraction of the cost and in a fraction of the time required by nascent multiplexed imaging technologies.KEY POINTSSpatially-resolved molecular profiling is an essential complement to histopathological evaluation of cancer tissues.Information obtained by immunofluorescence imaging is encoded by features in histological images.SHIFT leverages previously unappreciated features in histological images to facilitate virtual immunofluorescence staining.Feature representations of images guide sample selection, improving model generalizability.

Purpose: Pathologists rely on relevant clinical information, visual inspection of stained tissue slide morphology, and sophisticated molecular diagnostics to accurately infer the biological origin of secondary metastatic cancer. While highly effective, this process is expensive in terms of time and clinical resources. We seek to develop and evaluate a computer vision system designed to reasonably infer metastatic origin of secondary liver cancer directly from digitized histopathological whole slide images of liver biopsy. Approach: We illustrate a two-stage deep learning approach to accomplish this task. We first train a model to identify spatially localized regions of cancerous tumor within digitized hematoxylin and eosin (H&E)-stained tissue sections of secondary liver cancer based on a pathologist’s annotation of several whole slide images. Then, a second model is trained to generate predictions of the cancers’ metastatic origin belonging to one of three distinct clinically relevant classes as confirmed by immunohistochemistry. Results: Our approach achieves a classification accuracy of 90.2% in determining metastatic origin of whole slide images from a held-out test set, which compares favorably to an established clinical benchmark by three board-certified pathologists whose accuracies ranged from 90.2% to 94.1% on the same prediction task. Conclusions: We illustrate the potential impact of deep learning systems to leverage morphological and structural features of H&E-stained tissue sections to guide pathological and clinical determination of the metastatic origin of secondary liver cancers.