skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.


Title: Are we there yet? A machine learning architecture to predict organotropic metastases
Abstract Background & Aims Cancer metastasis into distant organs is an evolutionarily selective process. A better understanding of the driving forces endowing proliferative plasticity of tumor seeds in distant soils is required to develop and adapt better treatment systems for this lethal stage of the disease. To this end, we aimed to utilize transcript expression profiling features to predict the site-specific metastases of primary tumors and second, to identify the determinants of tissue specific progression. Methods We used statistical machine learning for transcript feature selection to optimize classification and built tree-based classifiers to predict tissue specific sites of metastatic progression. Results We developed a novel machine learning architecture that analyzes 33 types of RNA transcriptome profiles from The Cancer Genome Atlas (TCGA) database. Our classifier identifies the tumor type, derives synthetic instances of primary tumors metastasizing to distant organs and classifies the site-specific metastases in 16 types of cancers metastasizing to 12 locations. Conclusions We have demonstrated that site specific metastatic progression is predictable using transcriptomic profiling data from primary tumors and that the overrepresented biological processes in tumors metastasizing to congruent distant loci are highly overlapping. These results indicate site-specific progression was organotropic and core features of biological signaling pathways are identifiable that may describe proliferative plasticity in distant soils.  more » « less
Award ID(s):
1946937
PAR ID:
10303109
Author(s) / Creator(s):
; ; ; ; ; ; ;
Date Published:
Journal Name:
BMC Medical Genomics
Volume:
14
Issue:
1
ISSN:
1755-8794
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. null (Ed.)
    Abstract Development of an assay to predict response to chemotherapy has remained an elusive goal in cancer research. We report a phenotypic chemosensitivity assay for epithelial ovarian cancer based on Doppler spectroscopy of infrared light scattered from intracellular motions in living three-dimensional tumor biopsy tissue measured in vitro. The study analyzed biospecimens from 20 human patients with epithelial ovarian cancer. Matched primary and metastatic tumor tissues were collected for 3 patients, and an additional 3 patients provided only metastatic tissues. Doppler fluctuation spectra were obtained using full-field optical coherence tomography through off-axis digital holography. Frequencies in the range from 10 mHz to 10 Hz are sensitive to changes in intracellular dynamics caused by platinum-based chemotherapy. Metastatic tumor tissues were found to display a biodynamic phenotype that was similar to primary tissue from patients who had poor clinical outcomes. The biodynamic phenotypic profile correctly classified 90% [88–91% c.i.] of the patients when the metastatic samples were characterized as having a chemoresistant phenotype. This work suggests that Doppler profiling of tissue response to chemotherapy has the potential to predict patient clinical outcomes based on primary, but not metastatic, tumor tissue. 
    more » « less
  2. null (Ed.)
    Breast cancer cells can metastasize either as single cells or as clusters to distant organs from the primary tumor site. Cell clusters have been shown to possess higher metastatic potential compared to single cells. The organ microenvironment is critical in regulating the ultimate phenotype, specifically, the dormant versus proliferative phenotypes, of these clusters. In the context of breast cancer brain metastasis (BCBM), tumor cell cluster–organ microenvironment interactions are not well understood, in part, due to the lack of suitable biomimetic in vitro models. To address this need, herein, we report a biomaterial-based model, utilizing hyaluronic acid (HA) hydrogels with varying stiffnesses to mimic the brain microenvironment. Cell spheroids were used to mimic cell clusters. Using 100–10 000 MDA-MB-231Br BCBM cells, six different sizes of cell spheroids were prepared to study the impact of cluster size on dormancy. On soft HA hydrogels (∼0.4 kPa), irrespective of spheroid size, all cell spheroids attained a dormant phenotype, whereas on stiff HA hydrogels (∼4.5 kPa), size dependent switch between the dormant and proliferative phenotypes was noted ( i.e. , proliferative phenotype ≥5000 cell clusters < dormant phenotype), as tested via EdU and Ki67 staining. Furthermore, we demonstrated that the matrix stiffness driven dormancy was reversible. Such biomaterial systems provide useful tools to probe cell cluster–matrix interactions in BCBM. 
    more » « less
  3. Abstract Cancer metastasis is the leading cause of death for those afflicted with cancer. In cancer metastasis, the cancer cells break off from the primary tumor, penetrate nearby blood vessels, and attach and extravasate out of the vessels to form secondary tumors at distant organs. This makes extravasation a critical step of the metastatic cascade. Herein, with a focus on triple‐negative breast cancer, the role that the prospective secondary tumor microenvironment's mechanical properties play in circulating tumor cells' extravasation is reviewed. Specifically, the effects of the physically regulated vascular endothelial glycocalyx barrier element, vascular flow factors, and subendothelial extracellular matrix mechanical properties on cancer cell extravasation are examined. The ultimate goal of this review is to clarify the physical mechanisms that drive triple‐negative breast cancer extravasation, as these mechanisms may be potential new targets for anti‐metastasis therapy. 
    more » « less
  4. Abstract Colorectal cancer and other cancers often metastasize to the liver in later stages of the disease, contributing significantly to patient death. While the biomechanical properties of the liver parenchyma (normal liver tissue) are known to affect tumor cell behavior in primary and metastatic tumors, the role of these properties in driving or inhibiting metastatic inception remains poorly understood, as are the longer-term multicellular dynamics. This study adopts a multi-model approach to study the dynamics of tumor-parenchyma biomechanical interactions during metastatic seeding and growth. We employ a detailed poroviscoelastic model of a liver lobule to study how micrometastases disrupt flow and pressure on short time scales. Results from short-time simulations in detailed single hepatic lobules motivate constitutive relations and biological hypotheses for a minimal agent-based model of metastatic growth in centimeter-scale tissue over months-long time scales. After a parameter space investigation, we find that the balance of basic tumor-parenchyma biomechanical interactions on shorter time scales (adhesion, repulsion, and elastic tissue deformation over minutes) and longer time scales (plastic tissue relaxation over hours) can explain a broad range of behaviors of micrometastases, without the need for complex molecular-scale signaling. These interactions may arrest the growth of micrometastases in a dormant state and prevent newly arriving cancer cells from establishing successful metastatic foci. Moreover, the simulations indicate ways in which dormant tumors could “reawaken” after changes in parenchymal tissue mechanical properties, as may arise during aging or following acute liver illness or injury. We conclude that the proposed modeling approach yields insight into the role of tumor-parenchyma biomechanics in promoting liver metastatic growth, and advances the longer term goal of identifying conditions to clinically arrest and reverse the course of late-stage cancer. 
    more » « less
  5. Tumor microenvironment is a complex niche consisting of cancer cells and stromal cells in a network of extracellular matrix proteins and various soluble factors. Dynamic interactions among cellular and non-cellular components of the tumor microenvironment regulate tumor initiation and progression. Fibroblasts are the most abundant stromal cell type and dynamically interact with cancer cells both in primary tumors and in metastases. Cancer cells activate resident fibroblasts to produce and secrete soluble signaling molecules that support proliferation, migration, matrix invasion, and drug resistance of cancer cell and tumor angiogenesis. In recent years, various forms of three-dimensional tumor models have been developed to study tumor–stromal interactions and to identify anti-cancer drugs that block these interactions. There is currently a technological gap in development of tumor models that are physiologically relevant, scalable, and allow convenient, on-demand addition of desired components of the tumor microenvironment. In this review, we discuss three studies from our group that focus on developing bioengineered models to study tumor-stromal signaling. We will present these studies chronologically and based on their increasing complexity. We will discuss the validation of the models using a CXCL12-CXCR4 chemokine-receptor signaling present among activated fibroblasts and breast cancer cells in solid tumors, highlight the advantages and shortcomings of the models, and conclude with our perspectives on their applications. Impact statement Tumor stroma plays an important role in progression of cancers to a fatal metastatic disease. Modern treatment strategies are considering targeting tumor stroma to improve outcomes for cancer patients. A current challenge to develop stroma-targeting therapeutics is the lack of preclinical physiologic tumor models. Animal models widely used in cancer research lack human stroma and are not amenable to screening of chemical compounds for cancer drug discovery. In this review, we outline in vitro three-dimensional tumor models that we have developed to study the interactions among cancer cells and stromal cells. We describe development of the tumor models in a modular fashion, from a spheroid model to a sophisticated organotypic model, and discuss the importance of using correct physiologic models to recapitulate tumor-stromal signaling. These biomimetic tumor models will facilitate understanding of tumor-stromal signaling biology and provide a scalable approach for testing and discovery of cancer drugs. 
    more » « less