Related Experiment Video
Updated: Oct 14, 2025
08:35
Qualitative and Quantitative Analysis of the Immune Synapse in the Human System Using Imaging Flow Cytometry
Published on: January 7, 2019
11.8K
Imaging flow cytometry data analysis using convolutional neural network for quantitative investigation of
Elizaveta N Mochalova1,2,3, Ivan A Kotov1, Dmitry A Lifanov1
1Nanobiotechnology Laboratory, Moscow Institute of Physics and Technology, Moscow, Russia.
Biotechnology and Bioengineering
|November 9, 2021
Summary
A novel deep learning method using convolutional neural networks (CNNs) accurately quantifies phagocytosis from imaging flow cytometry (IFC) data. This automated approach overcomes manual analysis limitations, improving cancer and autoimmune disease immunotherapy research.
Area of Science:
- Immunology
- Computational Biology
- Biotechnology
Background:
- Macrophages are crucial immune cells involved in Fc receptor-mediated phagocytosis, a process targeted by immunotherapies for cancer and autoimmune diseases.
- Accurate analysis of phagocytosis is essential for improving current treatments and developing new therapeutic strategies.
- Imaging flow cytometry (IFC) offers high-throughput multiparametric analysis but conventional data processing relies on subjective manual methods, potentially losing valuable information.
Purpose of the Study:
- To apply a Faster region-based convolutional neural network (CNN) for accurate, automated quantitative analysis of phagocytosis using IFC data.
- To overcome the limitations of subjective manual analysis in conventional IFC data processing.
Main Methods:
- Utilized a Faster region-based convolutional neural network (CNN) for automated analysis of imaging flow cytometry (IFC) data.
- Employed a model system of erythrocyte phagocytosis by peritoneal macrophages.
- Developed a procedure for high-throughput identification and classification of cells, including macrophages and erythrocytes, with diverse visual characteristics.
Main Results:
- The CNN accurately performed high-throughput processing of IFC datasets, identifying and classifying macrophages and erythrocytes effectively.
- The developed procedure quantified phagocytosed cells, excluding low-probability classifications.
- Demonstrated impressive results in analyzing phagocytosis despite variations in cell shape, size, intensity, and texture.
Conclusions:
- CNN-based approaches offer a powerful tool for in-depth investigation of biological processes like phagocytosis.
- This method enhances the analysis of heterogeneous cellular objects in images, paving the way for new diagnostic and therapeutic capabilities.
- Automated quantitative analysis of phagocytosis using CNNs can significantly advance immunotherapy research.
