ISCO: Intelligent Framework for Accurate Segmentation and Comparative Analysis of Organoids

Abstract

Organoids are self-organizing 3D cell clusters that closely mimic the structure and function of in vivo tissues and organs. Quantifying organoid morphology is critical for advancing our understanding of organ development, drug discovery, and toxicity assessment. Recent advances in microscopy have provided powerful tools to capture detailed morphological features of organoids, yet manual image analysis remains labor-intensive and time-consuming. In response, we present a comprehensive microscopy-based analysis pipeline that utilizes SegmentAnything 2.1 to accurately segment individual organoids. Additionally, we introduce a suite of morphological features--including perimeter, area, radius, non-smoothness, and non-circularity--that enable researchers to quantitatively and automatically analyze organoid structures. To further standardize organoid analysis across the field, Intelligent segmentation and comparison of organoids(ISCO), an intelligent AI-driven open-source algorithm, is developed with the aim of establishing a comprehensive toolset for organoid characterization.