Reproduce “Benchmarking Pathology Foundation Models for Breast Cancer Survival Prediction” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Fredrik K. Gustafsson, Constance Boissin, Johan Vallon-Christersson, David A. Clifton, Mattias Rantalainen
The authors benchmark 13 pathology image encoders for breast-cancer survival prediction, using frozen patch features, a shared slide-level method, and training on one Swedish cohort before testing on two others. H-optimus-1 ranks first overall, while the 86-million-parameter H0-mini ranks second and exceeds its teacher in three of four settings. These retrospective results do not establish clinical benefit or wider generalization.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.