{"ID":2833392,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2512.03577","arxiv_id":"2512.03577","title":"Cross-Stain Contrastive Learning for Paired Immunohistochemistry and Histopathology Slide Representation Learning","abstract":"Universal, transferable whole-slide image (WSI) representations are central to computational pathology. Incorporating multiple markers (e.g., immunohistochemistry, IHC) alongside H\u0026E enriches H\u0026E-based features with diverse, biologically meaningful information. However, progress is limited by the scarcity of well-aligned multi-stain datasets. Inter-stain misalignment shifts corresponding tissue across slides, hindering consistent patch-level features and degrading slide-level embeddings. To address this, we curated a slide-level aligned, five-stain dataset (H\u0026E, HER2, KI67, ER, PGR) to enable paired H\u0026E-IHC learning and robust cross-stain representation. Leveraging this dataset, we propose Cross-Stain Contrastive Learning (CSCL), a two-stage pretraining framework with a lightweight adapter trained using patch-wise contrastive alignment to improve the compatibility of H\u0026E features with corresponding IHC-derived contextual cues, and slide-level representation learning with Multiple Instance Learning (MIL), which uses a cross-stain attention fusion module to integrate stain-specific patch features and a cross-stain global alignment module to enforce consistency among slide-level embeddings across different stains. Experiments on cancer subtype classification, IHC biomarker status classification, and survival prediction show consistent gains, yielding high-quality, transferable H\u0026E slide-level representations. The code and data are available at https://github.com/lily-zyz/CSCL.","short_abstract":"Universal, transferable whole-slide image (WSI) representations are central to computational pathology. Incorporating multiple markers (e.g., immunohistochemistry, IHC) alongside H\u0026E enriches H\u0026E-based features with diverse, biologically meaningful information. However, progress is limited by the scarcity of well-align...","url_abs":"https://arxiv.org/abs/2512.03577","url_pdf":"https://arxiv.org/pdf/2512.03577v1","authors":"[\"Yizhi Zhang\",\"Lei Fan\",\"Zhulin Tao\",\"Donglin Di\",\"Yang Song\",\"Sidong Liu\",\"Cong Cong\"]","published":"2025-12-03T09:00:27Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[]","has_code":false,"code_links":[{"ID":606318,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2833392,"paper_url":"https://arxiv.org/abs/2512.03577","paper_title":"Cross-Stain Contrastive Learning for Paired Immunohistochemistry and Histopathology Slide Representation Learning","repo_url":"https://github.com/lily-zyz/CSCL","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}