SNU develops deep learning technology that helps predict cancer patient’s survival

A computer-generated image illustrating a network graph of a cancer tissue [SEOUL NATIONAL UNIVERSITY]

Deep learning technology that helps predict a cancer patient’s survival rate with greater precision was developed by a Seoul National University (SNU) research team.

The study was published in Nature Biomedical Engineering on Friday. The joint research team encompassing medical and engineering experts at SNU includes Professor Kwon Sung-hoon of the electrical engineering department, Doctor Park Jeong-hwan and Doctor Moon Kyung-Chul of the College of Medicine.

The graph-based deep learning technology, the first of its kind, represents cancer tissues in a network diagram. Unlike the previous deep learning-powered diagnosis methods that can only analyze the shape of each cancer cell, the new technology considers the tumor microenvironment, which refers to the contextual factors related to cancer tissues such as the distance, interaction and correlations between immune and cancer cells, according to the research.

Analyzing such microenvironment is crucial in cancer diagnosis and predicting a cancer patient’s survival rate.

The SNU research team’s newly-developed deep learning technology finds patterns within the tumor microenvironment and measures correlations between cancer and immune cells. Medical staff can utilize the compiled and analyzed data in predicting the survival rate of a cancer patient.

The research team developed software with its deep learning technology in collaboration with Seoul National University Hospital and validated that the correlations among cancer cells, immune cells and blood vessels in a cancer tissue serve as an indicator of a cancer patient’s survival rate.

The technology is expected to be adopted in the next-generation diagnostic and prognostic index for cancer said SNU.

BY SHIN HA-NEE [shin.hanee@joongang.co.kr]