Research Areas Ⅳ
Advancing Standardized Experimental Animal Pathology through Deep Learning and Large-Scale Slide Data Analysis
Histopathological examination remains a critical endpoint in experimental animal research, but interpretation often depends on individual expertise and subjective scoring criteria. This limits reproducibility, cross-study comparison, and the full use of histological information generated from animal experiments.
We aim to advance the standardization and refinement of experimental animal pathology. Deep learning is used not as an end in itself, but as a tool to support this mission. AI-assisted analysis enables large-scale processing of whole-slide images, extraction of consistent morphological features, and objective comparison of disease-associated patterns across models, studies, and institutions.
This approach is also aligned with NAMs and the 3Rs. Large archives of experimental animal slides have already been generated, yet many remain underused after the original study. By transforming accumulated slide archives into reusable and analyzable pathology big data, we seek to increase the scientific value of existing animal studies, reduce unnecessary repetition, and support more responsible animal use.
Through standardized workflows, quantitative metrics, and scalable AI-assisted analysis, we aim to build a more reproducible, data-driven, and educational foundation for experimental animal pathology.
