Related Experiment Video
Updated: Jun 9, 2025
05:49
Mechano-Node-Pore Sensing: A Rapid, Label-Free Platform for Multi-Parameter Single-Cell Viscoelastic Measurements
Published on: December 2, 2022
2.6K
Neural Network-Enabled Multiparametric Impedance Signal Templating for High throughput Single-Cell Deformability
Javad Jarmoshti1, Abdullah-Bin Siddique1, Aditya Rane2
1Electrical & Computer Engineering, University of Virginia, Charlottesville, VA, 22904, USA.
Small (Weinheim an Der Bergstrasse, Germany)
|October 23, 2024
Summary
This study introduces a new method using impedance cytometry and neural networks to measure cell deformability and electrical properties in real-time. This allows for rapid, accurate sorting of live cell subpopulations for drug screening and cancer research.
Area of Science:
- Biophysics
- Cell Biology
- Machine Learning
Background:
- Cellular biophysical metrics change during metastasis and immune activation, offering potential for live cell subpopulation identification.
- Image-based cytometry quantifies cell deformability but requires complex image reconstruction, limiting real-time sorting applications.
- Impedance cytometry offers functional insights into cell viability and internal structure, complementing imaging techniques.
Purpose of the Study:
- To develop a rapid, inline method for quantifying single-cell biophysical metrics for cell sorting.
- To utilize impedance cytometry signals with a neural network for accurate cell deformability measurements.
- To enable multiparametric classification of live cancer cells and associated fibroblasts.
Main Methods:
- A multilayer perceptron neural network was employed for signal templating using raw impedance data from cells under extensional flow.
- The neural network was trained with image-derived metrics to establish net electrical anisotropy, quantifying cell deformability.
- Support vector machine models were used for multiparametric classification based on deformability and electrical physiology metrics.
Main Results:
- The neural network approach accurately quantifies cell deformability across a wide range of anisotropies, minimizing errors from cell size variations.
- The combined deformability and electrical physiology metrics enabled effective classification of live pancreatic cancer cells versus cancer-associated fibroblasts.
- The method demonstrates potential for real-time, high-throughput analysis and sorting of live cell subpopulations.
Conclusions:
- A novel impedance cytometry and neural network approach enables accurate, real-time measurement of cell deformability and electrical properties.
- This technique facilitates multiparametric cell classification, crucial for applications like targeted drug screening and cancer diagnostics.
- The developed method overcomes limitations of traditional imaging cytometry for inline cell sorting and analysis.
