Label-Free and Intelligent Cell Death Recognition Toward Lung Cancer Chemotherapy.

Area of Science:

• Biomedical Engineering
• Computational Biology
• Oncology

Background:

• Accurate cell death assessment is crucial for lung cancer chemotherapy effectiveness.
• Current methods lack high-throughput, label-free, and intelligent analysis capabilities.

Purpose of the Study:

• To develop an intelligent quantitative detection technique for cell death.
• To enable high-throughput, label-free analysis of cell death heterogeneity in lung cancer.

Main Methods:

• Utilized high-throughput quantitative phase imaging flow cytometry for label-free image acquisition.
• Employed convolutional neural networks (CNNs) for characterizing cell death heterogeneity and quantitative detection.
• Applied clustering for interpretability analysis of CNN models.

Main Results:

• Revealed heterogeneity of cell death through morphological feature analysis.
• Achieved high correlation (0.92 Pearson, 0.91 Cosine Similarity) with autophagy detection.
• Demonstrated an average error of 12.52% for apoptosis detection compared to biochemical methods.

Conclusions:

• The proposed intelligent technique offers a reliable and efficient method for cell death analysis.
• This approach provides a new perspective for studying cell death mechanisms in cancer.
• The method has potential applications in advancing lung cancer chemotherapy research and treatment strategies.