Feasibility study of stain-free classification of cell apoptosis based on diffraction imaging flow cytometry and...

Area of Science:

• Cell Biology
• Biophysics
• Computational Biology

Background:

• Apoptosis staging is crucial for understanding cellular health and disease.
• Traditional methods often rely on staining, which can be time-consuming and introduce artifacts.
• Developing stain-free methods for cell classification is a significant advancement.

Purpose of the Study:

• To explore the feasibility of a stain-free method for predicting and classifying cells in different apoptosis stages.
• To utilize diffraction images and supervised machine learning for cell analysis.
• To validate a novel polarization diffraction imaging flow cytometry (p-DIFC) technique.

Main Methods:

• Apoptosis was induced in K562 and HL-60 cells using cis-platinum (DDP) and hydrogen peroxide, respectively.
• Polarization diffraction imaging flow cytometry (p-DIFC) was employed to capture diffraction images of cells.
• Texture features from diffraction images were extracted using Gray-level co-occurrence matrix (GLCM) for machine learning classification.

Main Results:

• The p-DIFC method achieved over 90% accuracy in classifying cells across different apoptosis stages in independent test datasets.
• Logistic regression with ridge estimators was effective for cell classification.
• The method demonstrated feasibility in both K562 and HL-60 cell apoptosis models.

Conclusions:

• The stain-free p-DIFC system shows significant potential for predicting and classifying cells in various apoptosis stages.
• This approach offers a promising alternative to traditional staining methods in cell analysis.
• The integration of diffraction imaging and machine learning advances cell-based diagnostics.