AI can see the cell-types in a histology slide

Histology slides contain millions of different cells. Identifying the different kinds of cells - immune cells, tumor cells, connective tissue - needs to be done by a trained pathologist. This is very time and cost-intensive.  Owkin has created an AI tool that can now do this. Owkin's HistoPLUS is trained on millions of expert-curated annotations, and it identifies and segments various cell types, including tumor cells, lymphocytes, and macrophages, from digitized H&E slides.

HistoPLUS can also go beyond cell detection, extracting key features like nuclear morphology (roundness, area, aspect ratio) and spatial organization to reveal critical cell-cell interactions, such as neighborhood enrichment and co-occurrence patterns. These insights provide a comprehensive view of the tumor microenvironment, allowing researchers to see biomarkers and cellular dynamics essential for understanding tumor behavior. By making complex pathology data accessible, HistoPLUS empowers precision oncology and advances personalized treatment strategies.