Automated spheroid generation, drug application and efficacy screening using a deep learning classification: a...

Area of Science:

• Biotechnology
• Cell Biology
• Drug Discovery

Background:

• Three-dimensional (3D) cell cultures, specifically spheroids, mimic in vivo tissue environments, offering a more accurate model for cell behavior than 2D cultures.
• The labor-intensive nature of generating, treating, and analyzing spheroids hinders their widespread application in drug and compound research.

Purpose of the Study:

• To develop a fully automated system for high-throughput screening of drugs and compounds using 3D cell cultures.
• To enhance the efficiency and applicability of spheroid-based assays in pharmaceutical research and personalized medicine.

Main Methods:

• A novel automated pipetting robot capable of seeding hanging drops for spheroid formation.
• Automated spheroid treatment with drugs and subsequent image-based viability analysis using a deep learning convolutional neural network (CNN).
• CNN model training and validation using viability flow cytometry data for accurate classification of spheroid responses.

Main Results:

• The automated system successfully seeds, treats, and analyzes spheroids in a high-throughput manner.
• A deep learning CNN model accurately classifies spheroid viability ('unaffected', 'mildly affected', 'affected') after drug exposure.
• The approach enables efficient examination of drug combinatorics and new compound efficacy in 3D cell cultures.

Conclusions:

• The automated robotic system significantly improves the efficiency of drug screening in 3D cell cultures.
• This technology facilitates the examination of drug efficacy and supports the development of personalized therapeutic strategies for solid malignancies.
• The system offers a valuable tool for advancing drug discovery and precision oncology research.