Contrastive virtual staining enhances deep learning‐based <scp>PDAC</scp> subtyping from H&E‐stained tissue cores
Pancreatic ductal adenocarcinoma (PDAC) subtyping typically relies on immunohistochemistry (IHC) staining for critical markers like HNF1A and KRT81, a labor-intensive manual staining process that introduces variability. Virtual staining methods offer promising alternatives by generating synthetic IHC images from routine hematoxylin and eosin (H&E) slides. However, most current approaches evaluate success by image quality measures rather than assessing diagnostically relevant features. Here, we introduce a novel cycleGAN framework utilizing a contrastive-inspired approach trained on semipaired datasets derived from consecutive tissue sections. Our method significantly enhances PDAC subtyping accuracy based on synthetic IHC images generated from standard H&E inputs, improving the classification F1-score from 0.66 to 0.77 for KRT81 and from 0.61 to 0.73 for HNF1A, compared with classification directly on H&E images. This approach also substantially outperforms baseline CycleGAN models. These results underscore the clinical potential of contrastive virtual staining to streamline PDAC diagnostics and improve their robustness. © 2025 The Author(s). The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.
Authors
- Katja Steiger (ORCID: https://orcid.org/0000-0002-7269-5433)
- Marco Nolden (ORCID: https://orcid.org/0000-0001-9629-0564)
- Rickmer Braren (ORCID: https://orcid.org/0000-0001-6039-6957)
- Michael Götz (ORCID: https://orcid.org/0000-0003-0984-224X)
- Jens Kleesiek (ORCID: https://orcid.org/0000-0001-8686-0682)
- Sebastian Ziegler (ORCID: https://orcid.org/0000-0002-1693-5586)
- Shuhan Xiao (ORCID: https://orcid.org/0000-0001-5397-814X)
- Klaus Maier‐Hein (ORCID: https://orcid.org/0000-0002-6626-2463)
- Peter J. Schüffler
- Maximilian Fischer (ORCID: https://orcid.org/0009-0001-8305-6370)
- Philipp Schader
- Alexander Muckenhuber (ORCID: https://orcid.org/0000-0002-0414-1797)
- Hanno Gao
- Jens Thomas Siveke (ORCID: https://orcid.org/0000-0002-8772-4778)
- Peter Neher (ORCID: https://orcid.org/0000-0002-5285-7554)
- Constantin Ulrich (ORCID: https://orcid.org/0000-0003-3002-8170)
- Robin Peretzke (ORCID: https://orcid.org/0000-0002-6187-3636)
- Lukas Endrös (ORCID: https://orcid.org/0009-0007-1817-2598)
- Luay Farah
- Lorenz Feineis
Institutions
- Technische Hochschule Mannheim (DE)
- University of Applied Sciences and Arts of Southern Switzerland (CH)
- Universität Hamburg (DE)
- German Cancer Research Center (DE)
- Universität Ulm (DE)
- Heidelberg University (DE)
- University Hospital Heidelberg (DE)
- University Medical Center Hamburg-Eppendorf (DE)
- Institute for Advanced Study in the Humanities Essen (DE)
- National Center for Tumor Diseases (DE)
- Helmholtz Institute Mainz (DE)
- Essen University Hospital (DE)
- University Hospital Ulm (DE)
- Munich Center for Machine Learning
- University of Duisburg-Essen (DE)
- Technical University of Munich (DE)
Publication Details
- Journal
- The Journal of Pathology
- Published
- 2025-11-04
- DOI
- https://doi.org/10.1002/path.6491
- Primary Topic
- AI in cancer detection
- Type
- article
- Field-Weighted Citation Impact
- 0.00