Multi-scale characterization and analysis of cellular viscoelastic mechanical phenotypes by atomic force microscopy.

Area of Science:

• Biophysics
• Cell Biology
• Nanotechnology

Background:

• Cellular viscoelasticity is a biomarker for malignant transformation, crucial for cytological examination.
• Inconsistent measurement methods hinder effective characterization of cellular viscoelastic phenotypes.

Purpose of the Study:

• To develop a consistent method for characterizing cell viscoelasticity using atomic force microscopy (AFM).
• To utilize machine learning to identify optimal measurement parameters for distinguishing between normal and cancerous liver cells based on viscoelasticity.

Main Methods:

• Nanomechanical indentation experiments were performed using AFM on normal and cancerous liver cells.
• Multiple indentation methods and varied parameters were employed to build a viscoelasticity database.
• Machine learning algorithms were trained on the database to analyze cell differences and identify optimal measurement conditions.

Main Results:

• Measurement speed significantly impacts cell viscoelasticity, with 5 μm/s showing the most distinct classification between cell types.
• Machine learning algorithms effectively validated measurement parameters and methods.
• Multiparameter indentation measurements of viscoelasticity proved effective for cell classification.

Conclusions:

• This study establishes a machine learning-assisted approach for optimizing AFM-based viscoelasticity measurements.
• The findings confirm that optimized viscoelasticity measurements can reliably differentiate between normal and cancerous liver cells.
• The research provides a validated methodology for enhancing cytological examinations through precise cell mechanical property analysis.