An international research team led by the Korea Advanced Institute of Science and Technology (KAIST) has unveiled a groundbreaking technology that enables the non-invasive observation of cancer tissues in three dimensions. By combining cutting-edge optical imaging with artificial intelligence, the team has created realistic, virtually stained 3D images without the need for traditional biopsies or physical staining.
Collaborative Innovation Across Institutions
This pioneering effort was led by Professor YongKeun Park from KAIST’s Department of Physics, in collaboration with Professor Su-Jin Shin at Yonsei University Gangnam Severance Hospital, Professor Tae Hyun Hwang at Mayo Clinic, and the AI research team at Tomocube. Together, they developed a technique capable of visualizing the intricate 3D structure of cancer tissue without relying on conventional staining methods.
Moving Beyond Traditional Pathology
For more than two centuries, pathology has depended on the microscopic examination of thinly sliced, stained tissue samples. However, this method only reveals limited two-dimensional cross-sections of the tissue, restricting the understanding of cellular arrangements and spatial relationships.
To address this limitation, the researchers employed holotomography (HT), a sophisticated optical imaging technique that measures the 3D refractive index of tissues. They then applied a deep learning algorithm to convert this data into virtual Hematoxylin& Eosin (H&E) images—the gold standard in pathological staining. Hematoxylin stains cell nuclei blue, while eosin stains the cytoplasm pink, offering crucial contrast for cellular structure analysis.
AI-Powered Precision and Versatility
As per Medical Xpress, the virtual images generated using this method closely match traditionally stained tissue images in appearance and diagnostic value. The research team demonstrated the method’s high accuracy and confirmed its consistent performance across a variety of organs and tissue types, highlighting its robustness and adaptability for clinical use.
Proven Potential for Real-World Application
To validate the technology’s feasibility, the researchers collaborated with hospitals and research centers in both Korea and the United States. Using holotomography equipment developed by Tomocube, they successfully tested the system in practical settings, indicating strong potential for widespread clinical adoption.
A Leap Toward 3D Pathological Analysis
“This research represents a major leap forward by expanding pathological analysis from two dimensions to three,” said Professor Park. “We anticipate this technology will be instrumental in biomedical research and clinical diagnosis—especially in examining tumor boundaries and the spatial distribution of cells in the tumor microenvironment.”
The findings were published in the journal Nature Communications, marking a significant milestone in the future of non-invasive cancer diagnosis.