Automated image analysis software for the study and quantification of retinal glial cells

Microglia, the resident macrophages of the central nervous system, play a key role in immune surveillance, homeostasis, and neurodegenerative processes. Manual microglia analysis is time-consuming and prone to subjective bias, while most automated methods focus on quantification rather than characterization. Additionally, few tools are specifically designed for retinal microglia analysis. In this study, we present a novel automated image analysis software for microglia evaluation, integrating soma detection, quantification, characterization, skeletonization, and arborization measurement-the latter two being automated for the first time. The software was validated against expert manual annotations on 1,702 images fluorescence microscopy images of murine retinal tissue (24,559 cells), assessing its performance across both high- and low-quality images to evaluate its robustness in large-scale datasets. Our software processes images over 1,000 times faster than manual methods while maintaining high accuracy. It successfully analyzes low-quality images, though excluding them improves performance. The algorithm's ability to extract morphological features with high reproducibility enhances dataset usability, optimizing sample use and reducing animal sacrifices. This is the first tool to automate microglial skeletonization and arborization measurement, establishing a new standard for retinal microglia analysis. Its efficiency, scalability, and ability to handle varied image quality make it a valuable resource for large-scale studies. Future applications will focus on leveraging the reference values established in this study to advance neurodegenerative disease analysis, marking a significant step toward more sophisticated microglial research.