[Newsletter] Label-free AI-driven Virtual 3D H&E Staining
Newsletter
Innovate the way we see life | Tomocube Inc. | July 2025

Revealing 3D microanatomical structures of unlabeled thick cancer tissues using holotomography and virtual H&E staining
What if a virtual staining framework could unveil a detailed 3D representation of cancer tissue—without the need for a serial thin sectioning procedure?
This month, we're highlighting a new methodology that integrates our holotomography technology with deep learning to create 3D virtual H&E images from label-free, thick cancer tissue. By using holotomography as the input modality, this approach achieves a level of accuracy and spatial information beyond what conventional methods can offer.
This is an exciting step forward, opening up new possibilities in histopathology.
Explore the full details in the recent Nature Communications article to discover a powerful approach that overcomes two predominant challenges in histology: the limitations of 2D analysis and the need to preserve both structure and context while preparing samples non-invasively and efficiently.
Virtual 3D H&E Staining: Visualizing the Impossible
Key highlights of the article:
- 3D imaging of thick tissue. Holotomography accurately reconstructs the 3D architecture of colon cancer tissues up to 50 µm thick — 12.5 times thicker than conventional slides.
- AI-powered virtual staining. A deep learning framework translates label-free refractive index images into 3D virtual H&E, revealing subcellular details without actual dyes or sectioning.
- Solving key limitations. This approach overcomes major drawbacks of the conventional method in histology: the poor visualization of fine structures in thick tissues due to uneven penetration of chemical H&E dyes and the disruption of spatial continuity due to serial thin sectioning. These issues compromise data integrity, a problem further compounded by labor-intensive, time-consuming workflows.
- Enabling deeper insight. New method delivers volumetric, artifact-free visualization to boost both efficiency and diagnostic value and paves the way for breakthroughs in cancer research and digital pathology.
3D microanatomical rendering and quantitative analysis of a whole 50 μm-thick colon cancer tissue slide.
This study builds on a growing trend of integrating advanced imaging technology and AI in biomedical research. For deeper insight, explore these two related articles that highlight advancements across cancer discovery and the application of holotomography:
- Wang, L., Li, M., & Hwang, T. H. The 3D revolution in cancer discovery. Cancer Discovery, 14(4), 625–629 (2024).
- Park, J., Bai, B., Ryu, D. et al. Artificial intelligence-enabled quantitative phase imaging methods for life sciences. Nat Methods 20, 1645–1660 (2023).
[Video] Brain Tissue Microstructure Visualization
Brain tissue microstructures, taken label-free by the Tomocube HT-X1