Korean researchers develop AI-based real-time gastric cancer mapping technology

A team of Korean researchers has developed a method to estimate the extent of gastric cancer in real time during surgery using artificial intelligence.

The National Cancer Center (NCC) said Wednesday that a research team led by Yoon Hong-man, head of its Gastric Cancer Center, in collaboration with a team led by Professor Kim Hyung-min of the Department of Chemistry at Kookmin University, has developed a technology capable of distinguishing gastric cancer tissue intraoperatively in real time.

The National Cancer Center said Wednesday that a research team led by Dr. Yoon Hong-man developed a method to improve surgical precision by enabling real-time identification of gastric tumor locations. (Courtesy of NCC)

The technology combines AI with autofluorescence spectroscopy, a method that analyzes the faint light naturally emitted by biological tissues. The approach enables rapid and accurate identification of tumor location and boundaries. However, conventional autofluorescence spectroscopy has faced limitations, including variability in measurement results depending on equipment and sensitivity to environmental conditions. In addition, overlapping light signals have made precise differentiation difficult.

To address these challenges, the researchers developed a customized environmental control system that maintains constant temperature, humidity and lighting conditions. Data collected under these controlled conditions were used to train a deep learning-based AI model, enabling precise separation of complex mixed light signals.

As a result, the newly developed technology distinguished cancerous tissue from normal tissue with 88.1 percent accuracy, demonstrating a near real-time capability to identify tumor boundaries during surgery.

Dr. Yoon said, “This is an innovative study that overcomes the limitations of conventional light-based analysis through AI and environmental control technologies. It is expected to significantly assist surgeons by enabling immediate identification of tumor margins in the operating room.”

The study was published in the international journal Analytical Chemistry under the title “A Robust Spectro-Histopathological Diagnostic Method for Improved Gastric Cancer Detection."