Abstract
Generating specialized histopathological stains like Immunohistochemistry
(IHC) and Periodic Acid-Schiff (PAS) in a laboratory is
costly and time-consuming, making purely digital virtual staining a critical
solution. However, acquiring perfectly paired training data is physically
impossible, and standard diffusion models often alter the underlying
tissue architecture when applied to challenging unpaired datasets.
To address this, a novel morphology-preserving virtual staining framework
is proposed for unpaired H&E-to-PAS and H&E-to-IHC translation.
For the PAS domain, a luminance-guided post-processing step is
introduced to recover fine structural boundaries. For the IHC domain, a
two-step inference pipeline guides the generative process using a nuclear
mask to protect spatial fidelity, followed by an automated refinement step
to remove structural artifacts. Quantitative and qualitative evaluations
demonstrate that the proposed approach improves morphological metrics
compared to baselines. Furthermore, incorporating these generated
virtual stains into a downstream subcancer classification task improves
overall accuracy, demonstrating the practical and clinical utility of the
proposed method.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/COMPAYL_051.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
BibTex
@InProceedings{SpaGiu_MorphologyPreserving_MICCAISAT2026,
author = { Spathis, Giuseppe AND Chtourou, Imen AND Descombes, Xavier AND Ambrosetti, Damien AND Toni, Giorgio AND Hannetel, Paul},
title = { { Morphology-Preserving Virtual Staining from H&E to PAS and IHC using Enhanced Diffusion Models } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17251},
month = {pending},
page = {pending}
}