Reproduce “WILSON - Multi-magnification composites as a foundation for patient-level pathology analysis and diagnostic text generation” 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

Saghir Alfasly, Wataru Uegami, Sobhan Hemati, Wenchao Han, Xiaojia Tang, Kevin Thompson, Daniel Stone, Ghazal Alabtah and 12 more

WILSON represents pathology slides and patient cases as composites combining selected tissue views at multiple magnifications, which one vision encoder processes across one or several slides. Trained on approximately 189,000 Mayo Clinic slides, it achieved mean macro-F1 of 0.52 across four zero-shot case-level benchmarks, versus MOOZY’s 0.38; the external CPTAC-BRCA difference was within composite-to-composite variation.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.