Biophysical cytometry for classification of circulating pancreatic cancer cells in their epithelial, mesenchymal and hybrid states

Tuesday, October 7, 2025

4:30 PM - 4:50 PM EDT

Location: Meeting Room #4&5

Presenting Author(s)

NS

Nathan Swami

Professor
University of Virginia
Charlottesville, Virginia, United States

Non-Presenting Author(s)

JJ

Javad Jarmoshti

PhD Candidate
University of Virginia
Charlottesville, Virginia, United States
NZ

Negin Zenali

University of Virginia
Charlottesville, Virginia, United States
Abdullah-Bin Siddique (he/him/his)

PhD Candidate
University of Virginia
Charlottesville, Virginia, United States

In fifteen words or less, explain the significance of this contribution (Novel Aspect).: Multiparametric biophysical properties enable machine learning-based single-cell classification of the heterogeneity of cancer cell states

Abstract Text: Cellular biophysical metrics exhibit systematic alterations that can be used to identify and separate live cell subpopulations for their targeting during drug screening. Specifically for solid tumors, stiffening of the tumor microenvironment leads to metastasis of specific cell subpopulations, thereby leading to invasion by highly deformable cell subpopulations that can often be identified by their expression of mesenchymal markers. However, hybrid cancer cells expressing epithelial and mesenchymal markers present the greatest phenotypic plasticity for invasion and re-colonization within secondary organs, based on recent clinical correlations of disease metrics to surface markers on circulating tumor cells (CTCs). Since these hybrid CTCs cannot be easily detected by current immunofluorescence methods, we explore the utilization of single-cell biophysical metrics arising from mechanical and electrical physiologies, due to their role in stiffening of the tumor microenvironment. Our hypothesis is that circulating tumor cells expressing epithelial (E-CTCs), mesenchymal (M-CTCs), and hybrid (H-CTCs) markers can be distinguished on biophysical metrics, with their deformability level systematically varying in proportion to their expression level of mesenchymal and epithelial markers. A viscoelastic microfluidic extensional flow device is used to measure electrical physiologies based on the impedance frequency response and mechanical deformation of the same single-cell event based on shape anisotropy acquired by high-speed imaging. In this manner, multiparametric biophysical properties can be utilized to classify heterogeneous samples of cancer cells that are spread over a progression of states, by using a support vector machine model (SVM) that is trained with control epithelial and mesenchymal cells. This is validated using various pancreatic cancer cell types that are induced using growth factors and drug treatment towards epithelial to mesenchymal transition (EMT) and mesenchymal to epithelial transition (MET). We envision that these biophysical metrics can be used in future work in conjunction with cell surface markers to detect hybrid metastatic cancer cells expressing epithelial and mesenchymal states.