Methods of Machine Learning-Based Chimeric Antigen Receptor Immunological Synapse Quality Quantification.

Area of Science:

• Immunotherapy
• Biotechnology
• Machine Learning

Background:

• Chimeric Antigen Receptor (CAR)-T cell therapy is effective against refractory blood cancers, with six CAR-T drugs approved by the FDA.
• Effective CAR-T cell function relies on forming a robust immunological synapse (IS) with target tumor cells.
• CAR IS quality is a potential predictive biomarker for CAR-T immunotherapy efficacy, but its clinical quantification is challenging.

Purpose of the Study:

• To present a user-friendly, step-by-step methodology for predicting CAR-modified cell efficacy using machine learning-based CAR IS quality quantification.
• To provide detailed guidance on imaging CAR IS, defining focal planes, segmenting images, and applying ML algorithms for IS quality assessment.

Main Methods:

• Utilizing a glass-supported planar lipid bilayer system for imaging CAR-T cell immunological synapses (IS).
• Implementing image segmentation techniques to isolate and analyze CAR IS.
• Applying machine learning algorithms for quantitative assessment of CAR IS quality.

Main Results:

• Demonstrated an accessible approach for predicting CAR-T cell efficacy through ML-based CAR IS quality quantification.
• Detailed protocol covers CAR IS imaging, focal plane definition, image segmentation, and ML-driven quality assessment.

Conclusions:

• The developed approach significantly improves the accuracy and efficiency of predicting CAR-T cell immunotherapy outcomes.
• This methodology offers a practical tool for researchers to enhance CAR IS prediction in preclinical and clinical settings.