A Semi-Automated Microfluidic Platform Employing Machine Learning Analysis to Study Adhesion Kinetics in Acute...

Area of Science:

• Biomedical Engineering
• Cell Biology
• Drug Discovery

Background:

• Current drug screening methods often neglect the tumor microenvironment's influence on cellular behavior.
• Microfluidic shear flow assays offer sensitive single-cell analysis within a simulated microenvironment.
• Cell adhesion in these assays serves as a critical indicator of cellular state and viability.

Purpose of the Study:

• To develop a semi-automated microfluidic shear flow platform for enhanced drug screening.
• To integrate hardware operation with a machine-learning analysis pipeline for efficient data processing.
• To improve the accuracy and speed of analyzing tumor cell communication and viability.

Main Methods:

• Development of a bright-field microscopy-enabled, semi-automated shear flow platform.
• Implementation of a machine-learning-based analysis pipeline for assay data.
• Utilizing cell adhesion as a functional readout for cellular state and viability.

Main Results:

• The platform provides consistent, high-quality results rapidly.
• Machine learning algorithm achieved 98.3% precision and 99.1% recall.
• Demonstrated high tracking specificity and object discrimination capabilities.

Conclusions:

• The developed platform offers a significant advancement in microfluidic drug screening.
• Semi-automation and machine learning enhance efficiency and accuracy in analyzing tumor cell behavior.
• This technology holds promise for more effective drug discovery by accounting for the tumor microenvironment.