"Improving Treatment Precision" Predicting the Effectiveness of Immunotherapy with AI from Yonsei University College of Medicine

[Financial News] A path has opened to predict how effective immunotherapy will be for cancer patients using artificial intelligence (AI).
Yonsei University College of Medicine announced on the 27th that it has developed an AI model to analyze the suitability of immunotherapy through joint research with the Mayo Clinic and Vanderbilt University Medical Center in the United States.

This model precisely analyzes the pathological images of cancer patients to predict the likelihood of the patient responding to immunotherapy. The results of this study were published in the latest issue of the international journal (npj Digital Medicine).
Immunotherapy works by inducing immune cells in the body to recognize and eliminate cancer cells. However, the effectiveness of the treatment varies greatly depending on the genetic characteristics of the cancer cells.
In particular, patients with gastric and colorectal cancer who can expect the effects of immunotherapy often exhibit a genetic characteristic known as 'high microsatellite instability (MSI-H)'.
MSI-H is a cell type with many mutations, making it easier for immune cells to recognize cancer as a 'foreign invader', resulting in a high treatment response. The problem is that with existing testing methods, MSI-H, which is not spread throughout the entire cancer tissue but only in some parts, is easy to miss. 
The AI model 'MSI-SEER' developed by Professor Jeong Jae-ho's team at Yonsei University College of Medicine overcomes these limitations. It divides cancer tissue pathology images into thousands of small tile images and quantitatively analyzes and visualizes the probability of MSI-H existing in each area. 
AI also provides the reliability of its judgment, allowing medical staff to have objective grounds for interpreting the prediction results. 
This study also reported cases where the AI model identified MSI-H patients who were not diagnosed with existing testing methods. In a test evaluating the clinical applicability conducted by the research team, the AI model detected the presence of MSI-H in gastric and colorectal cancer patients who were previously judged as MSI-H negative and did not use immunotherapy. 
Professor Jeong Jae-ho of the Department of Gastrointestinal Surgery at Yonsei University College of Medicine said, "Accurately identifying the genetic characteristics within cancer cells determines the direction of patient treatment," adding, "This AI model is significant as a diagnostic support tool that helps clinicians make clearer decisions."
Professor Hwang Tae-hyun, who participated in the research from Vanderbilt University Medical Center in the United States, emphasized, "This AI technology is the starting point of true medical AI, where the expertise of doctors and the computational power of artificial intelligence collaborate, beyond simple automatic analysis."
The research team at Yonsei University College of Medicine plans to conduct additional verification and technological advancements to apply this AI model to real clinical environments.

vrdw88@fnnews.com Kang Joong-mo Reporter