Related Experiment Video
Updated: Oct 31, 2025
10:27
Controlled Microfluidic Environment for Dynamic Investigation of Red Blood Cell Aggregation
Published on: June 4, 2015
12.0K
Combining microfluidics with machine learning algorithms for RBC classification in rare hereditary hemolytic anemia
Valeria Rizzuto1,2,3, Arianna Mencattini4,5, Begoña Álvarez-González2,6
1Josep Carreras Leukaemia Research Institute (IJC), 08916, Badalona, Spain.
Scientific Reports
|July 1, 2021
Summary
This study uses a microfluidic device and machine learning to analyze red blood cell shape, mimicking spleen function for rare hereditary hemolytic anemia diagnostics. The platform accurately distinguishes between healthy individuals and RHHA patients.
Area of Science:
- Biomedical Engineering
- Hematology
- Artificial Intelligence
Background:
- The spleen removes defective red blood cells (RBCs) in rare hereditary hemolytic anemia (RHHA).
- Current methods for RHHA diagnosis and monitoring require improvement for accurate RBC characterization.
Purpose of the Study:
- To develop a microfluidic platform combined with machine learning for in vitro RBC analysis.
- To mimic spleen red pulp filtration for studying RBC deformability and shape in RHHA.
- To enable smart decision-making for clinical diagnostics of blood diseases.
Main Methods:
- A microfluidic device was designed to simulate spleen slits, analyzing RBC deformability.
- Video data analysis of RBCs in planar orientation after microconstriction passage.
- Two cooperative machine learning approaches (majority voting, maximum sum of scores) for classification.
Main Results:
- The platform achieved 91% average efficiency in discriminating healthy controls from RHHA patients.
- It distinguished between different RHHA subtypes with 82% efficiency.
- Demonstrated capability for quantitative cell behavior study and smart diagnostic support.
Conclusions:
- The integrated microfluidic and machine learning system offers a novel approach for RHHA diagnosis.
- This technology can aid in the in vitro study and monitoring of blood diseases through RBC shape analysis.
- The platform shows potential for improving clinical diagnostics by accurately characterizing RBCs.
