From left: Jooyun Park, doctoral student in the Department of Physics at KAIST, and Professor YongKeun Park. Top left: Professor Soo-Jin Shin of Yonsei University’s Gangnam Severance Hospital, and Professor Taehyun Hwang of Vanderbilt University School of Medicine. (Yonhap)
DAEJEON, South Korea, May 27 (Korea Bizwire) — An international research team led by KAIST has developed a groundbreaking non-invasive imaging technique that enables three-dimensional observation of cancer tissue without the need for traditional staining or incisions.
KAIST announced on May 26 that Professor YongKeun Park’s team in the Department of Physics, in collaboration with researchers from Yonsei University’s Gangnam Severance Hospital, the Mayo Clinic in the U.S., and AI specialists from Tomocube Inc., has successfully created realistic, 3D virtual histological images using a combination of advanced optical imaging and artificial intelligence.
The technique leverages holotomography, an optical method that maps the 3D refractive index of biological tissues. By integrating this with a deep-learning algorithm, the team generated virtual hematoxylin and eosin (H&E)-stained images — a standard in pathology — without physically altering the tissue.
Korean Researchers Develop Non-Invasive 3D Method to Analyze Cancer Tissue (Image supported by ChatGPT)
Traditionally, tissue samples are thinly sliced and chemically stained before examination under a microscope, revealing only flat, two-dimensional cross-sections. This new method allows clinicians and researchers to visualize the full 3D architecture of cancerous tissue, including the spatial relationships between cells, with high resolution and without destroying the sample.
The technique was validated using colon and gastric cancer samples, demonstrating its ability to render intricate microstructures in vivid detail. The researchers anticipate the innovation could revolutionize next-generation non-invasive diagnostics and enhance biomedical research.
“This is a highly meaningful achievement that expands pathological analysis from two dimensions to three,” said Professor Park. “It opens the door to detailed spatial analysis of tumor boundaries and surrounding microenvironments, which could greatly advance both research and clinical diagnosis.”
The study, with KAIST doctoral student Jooyun Park as first author, was published online in Nature Communications on May 22, 2025.
Kevin Lee (kevinlee@koreabizwire.com)
