Computationally enabled polychromatic polarized imaging enables mapping of matrix architectures that promote pancreatic ductal adenocarcinoma dissemination.
Authors
Affiliations (5)
Affiliations (5)
- Department of Biomedical Engineering, University of Minnesota; Center for Multiparametric Imaging of Tumor Immune Microenvironments, University of Minnesota and University of Wisconsin-Madison.
- Center for Multiparametric Imaging of Tumor Immune Microenvironments, University of Minnesota and University of Wisconsin-Madison; Center for Quantitative Cell Imaging, University of Wisconsin-Madison, Madison, WI.
- Marine Biological Laboratory, University of Chicago, Woods Hole, MA.
- Center for Multiparametric Imaging of Tumor Immune Microenvironments, University of Minnesota and University of Wisconsin-Madison; Center for Quantitative Cell Imaging, University of Wisconsin-Madison, Madison, WI; Department of Medical Physics, University of Wisconsin-Madison, Madison, WI.
- Department of Biomedical Engineering, University of Minnesota; Center for Multiparametric Imaging of Tumor Immune Microenvironments, University of Minnesota and University of Wisconsin-Madison; Masonic Cancer Center, University of Minnesota; Dept of Medicine, Div. of Hematology, Oncology, and Transplantation, University of Minnesota; Institute for Engineering in Medicine, University of Minnesota; Stem Cell Institute, University of Minnesota. Electronic address: [email protected].
Abstract
Pancreatic ductal adenocarcinoma (PDA) is an extremely metastatic and lethal disease. In PDA, extracellular matrix (ECM) architectures known as Tumor-Associated Collagen Signatures (TACS) regulate invasion and metastatic spread in both early dissemination and in late-stage disease. As such, TACS has been suggested as a biomarker to aid in pathologic assessment. However, despite its significance, approaches to quantitatively capture these ECM patterns currently require advanced optical systems with signaling processing analysis. Here we present an expansion of polychromatic polarized microscopy (PPM) with inherent angular information coupled to machine learning and computational pixel-wise analysis of TACS. Using this platform, we are able to accurately capture TACS architectures in H&E stained histology sections directly through PPM contrast. Moreover, PPM facilitated identification of transitions to dissemination architectures, i.e., transitions from sequestration through expansion to dissemination from both PanINs and throughout PDA. Lastly, PPM evaluation of architectures in liver metastases, the most common metastatic site for PDA, demonstrates TACS-mediated focal and local invasion as well as identification of unique patterns anchoring aligned fibers into normal-adjacent tumor, suggesting that these patterns may be precursors to metastasis expansion and local spread from micrometastatic lesions. Combined, these findings demonstrate that PPM coupled to computational platforms is a powerful tool for analyzing ECM architecture that can be employed to advance cancer microenvironment studies and provide clinically relevant diagnostic information.