Prompt-Guided Virtual Multiplex IHC Staining

H&E to Virtual IHC Staining

Upload an H&E pathological image, select the target staining domain, and invoke the backend PGVMS model to generate virtual IHC results. It is recommended to use a 20× magnification image;

H&E Input pathology image
IHC PR / Ki67 / HER2 / ER
PGVMS Generator backend ready
Size Auto-detected

Run virtual staining

No file chosen
Supported formats: JPG, PNG, TIF
Input H&E waiting
Upload an H&E image to preview.
Virtual IHC PGVMS
Generated result will appear here.
Paper

Title: PGVMS: A Prompt-Guided Unified Framework for Virtual Multiplex IHC Staining with Pathological Semantic Learning (TMI 2026)

Abstract: PGVMS is a prompt-guided framework for virtual multiplex IHC staining. Using only uniplex IHC training data, PGVMS digitally transforms H&E images into multiple IHC staining results. It leverages a pathological vision-language model for adaptive semantic guidance, and introduces protein-aware and prototype-consistent learning strategies to improve staining distribution consistency and pathological reliability.

Authors: Fuqiang Chen, Ranran Zhang, Wanming Hu, Deboch Eyob Abera, Yue Peng, Boyun Zheng, Yiwen Sun, Jing Cai, Wenjian Qin

Institution: Shenzhen Institutes of Advanced Technology，Chinese Academy of Sciences

Cite as: F. Chen et al., " PGVMS: A Prompt-Guided Unified Framework for Virtual Multiplex IHC Staining with Pathological Semantic Learning", in IEEE Transactions on Medical Imaging, doi: 10.1109/TMI.2026.3663755.

Chen F, Zhang R, Zheng B, et al. Pathological semantics-preserving learning for H&E-to-IHC virtual staining[C]//International Conference on Medical Image Computing and Computer-Assisted Intervention. Cham: Springer Nature Switzerland, 2024: 384-394.