🧠 AI🟢 BullishImportance 7/10

Enabling clinical use of foundation models in histopathology

arXiv – CS AI|Audun L. Henriksen, Ole-Johan Skrede, Lisa van der Schee, Enric Domingo, Sepp De Raedt, Ily\'a Kostolomov, Jennifer Hay, Karolina Cyll, Wanja Kildal, Joakim Kalsnes, Robert W. Williams, Manohar Pradhan, John Arne Nesheim, Hanne A. Askautrud, Maria X. Isaksen, Karmele Saez de Gordoa, Miriam Cuatrecasas, Joanne Edwards, TransSCOT group, Arild Nesbakken, Neil A. Shepherd, Ian Tomlinson, Daniel-Christoph Wagner, Rachel S. Kerr, Tarjei Sveinsgjerd Hveem, Knut Liest{\o}l, Yoshiaki Nakamura, Marco Novelli, Masaaki Miyo, Sebastian Foersch, David N. Church, Miangela M. Lacle, David J. Kerr, Andreas Kleppe|February 27, 2026 at 05:00 AM|6 views

🤖AI Summary

Researchers developed a method to improve foundation models in medical histopathology by introducing robustness losses during training, reducing sensitivity to technical variations while maintaining accuracy. The approach was tested on over 27,000 whole slide images from 6,155 patients across eight popular foundation models, showing improved robustness and prediction accuracy without requiring retraining of the foundation models themselves.

Key Takeaways

→Foundation models in histopathology currently capture unwanted technical variations that bias predictions in clinical applications.
→Novel robustness losses during downstream training significantly reduce sensitivity to scanner-specific and pre-analytic variations.
→The method was validated on 27,042 whole slide images from 6,155 patients across eight different foundation models.
→Prediction accuracy improved by focusing training on biologically relevant features rather than technical artifacts.
→The approach enables robust computational pathology models for real-world clinical practice without foundation model retraining.