Atlas H&E-TME: Scalable AI-Based Tissue Profiling at Expert Pathologist-Level Accuracy

Atlas H&E-TME is an AI-based system designed for scalable analysis of hematoxylin and eosin (H&E) whole-slide images, achieving expert pathologist-level accuracy in predicting tissue quality, regions, and cell types across various cancer types. It provides over 4,500 quantitative readouts per slide, validated through a dual framework that combines an IHC-informed multi-pathologist consensus protocol with extensive benchmarking against over 200,000 H&E annotations from 1,500+ cases. This development enables more accurate and scalable tissue profiling, which is crucial for advancing tissue-based biomarkers in clinical and translational research.