Neural network-enhanced real-time impedance flow cytometry for single-cell intrinsic characterization

Area of Science:

• Biophysics
• Cell Biology
• Analytical Chemistry

Background:

• Single-cell impedance flow cytometry (IFC) is a label-free technique for assessing cell electrical properties and heterogeneity.
• Current IFC methods often rely on phenomenological parameters, requiring time-consuming off-line model fitting for intrinsic biophysical metrics.
• There is a need for faster, high-throughput methods for real-time intrinsic cell characterization.

Purpose of the Study:

• To develop a neural network (NN)-enhanced IFC method for real-time, high-throughput single-cell intrinsic characterization.
• To enable rapid online calculation of intrinsic biophysical parameters (radius, cytoplasm conductivity, specific membrane capacitance).
• To achieve accurate cell classification based on these intrinsic parameters.

Main Methods:

• Implementation of a trained neural network (NN) to process IFC data in real-time.
• Online calculation of three intrinsic cell parameters: radius (r), cytoplasm conductivity (σi), and specific membrane capacitance (Csm).
• High-throughput testing on five cell types (four cancer, one lymphocyte) involving 9751 cell samples.

Main Results:

• The NN-enhanced IFC achieved real-time online calculation of intrinsic parameters at 0.3 ms per cell event.
• Demonstrated 91.5% classification accuracy for cell types within the test group.
• Viability assays confirmed that the IFC procedure does not substantially alter cell properties.

Conclusions:

• NN-enhanced IFC offers a significant advancement for rapid, online intrinsic electrical characterization of single cells.
• This technology facilitates high-throughput cell classification and analysis with improved speed and accuracy.
• The method holds promise as a new platform for various cell-based assays and research.