AI-Based Prognostic Prediction Model Developed for Hard-to-Treat Gallbladder Cancer

Research Team Led by Park Ju-Kyung and Kim Hongbeom at Samsung Medical Center
Discovery of New Biomarkers Through AI-Based Tumor Microenvironment Analysis
Precise Prognostic Classification Using Three Key Risk Factors

Samsung Medical Center has developed a model that uses artificial intelligence (AI) to predict the recurrence and survival probability of gallbladder cancer patients. This model is expected to be useful in establishing personalized treatment strategies for gallbladder cancer, which is known for its difficulty in prognostic prediction.

A research team that discovered new biomarkers for gallbladder cancer by analyzing the tumor microenvironment using AI. Samsung Medical Center

On May 20, Samsung Medical Center announced that a research team led by Professors Park Ju-Kyung, Lee Kyutaek, and Choi Younghoon from the Department of Gastroenterology, Professor Kim Hongbeom from the Department of Hepatobiliary and Pancreatic Surgery, and Dr. Kim Hyemin from the Institute for Future Medicine confirmed the possibility of predicting patient prognosis by analyzing the tumor microenvironment (TME) of gallbladder cancer using AI-based spatial analysis technology. The research findings were published in the latest issue of the International Journal of Surgery.

Gallbladder cancer typically has almost no symptoms in its early stages, so it is often detected at an advanced stage. According to the Central Cancer Registry, the 5-year relative survival rate for gallbladder and other biliary tract cancers is 29%, which is the second lowest after pancreatic cancer.

The research team developed a prognostic prediction model based on data from 225 patients who underwent surgery for gallbladder cancer and validated its performance with an external cohort of 41 patients. The AI quantitatively analyzed key indicators of the tumor microenvironment, including the density of tumor-infiltrating lymphocytes (TIL), the number of tertiary lymphoid structures (TLS), and the density of fibroblasts around the cancer cells.

The analysis identified the main risk factors associated with worse prognosis as follows: low TIL density, a small number of TLS, and high fibroblast density. The greater the number of risk factors present, the shorter the overall survival (OS) and disease-free survival (DFS) periods.

Patients without any risk factors had an 87% lower risk of recurrence and an 80% lower risk of death compared to those with all three risk factors.

Professor Kim Hongbeom stated, "Gallbladder cancer is a cancer type for which prognostic prediction is particularly difficult," adding, "AI-based precision analysis can help establish patient treatment strategies." Professor Park Ju-Kyung said, "This study demonstrated the potential of AI to serve as a digital biomarker for predicting prognosis by analyzing the biological characteristics of cancer."

This content was produced with the assistance of AI translation services.