An AI-based imaging flow cytometry approach to study erythrophagocytosis

Abstract

Erythrophagocytosis is a process consisting of recognition, engulfment and digestion by phagocytes of antibody-coated or damaged erythrocytes. Understanding the dynamics that are behind erythrophagocytosis is fundamental to comprehend this cellular process under specific circumstances. Several techniques have been used to study phagocytosis. Among these, an interesting approach is the use of Imaging Flow Cytometry (IFC) to distinguish internalization and binding of cells or particles. However, this method requires laborious analysis. Here, we introduce a novel approach to analyze the phagocytosis process by combining Artificial Intelligence (AI) with IFC. Our study demonstrates that this approach is highly suitable to study erythrophagocytosis, categorizing internalized, bound and non-bound erythrocytes. Validation experiments showed that our pipeline performs with high accuracy and reproducibility.