LG AI Research and Vanderbilt University Medical Center unveiled the results of their joint cancer agentic artificial intelligence research at the American Association for Cancer Research 2026, which is being held in San Diego from Friday to Wednesday.
The two sides were to present the findings in a joint session titled “Human-AI cooperation: AI as a decision-making partner for medical professionals,” highlighting how AI can support clinical decision-making.
The system, built on LG’s pathology AI model Exaone Path, is designed to streamline the entire cancer treatment workflow — from tissue analysis to treatment planning — within a single day.
Unlike conventional approaches, the agentic AI operates through a multiagent framework in which specialized AI agents collaborate step-by-step across the treatment process. These include cancer tissue image analysis, tumor mapping, cross-validation between AI predictions and clinical data, drug response evaluation, treatment strategy formulation and clinical decision support.
“The collaborative model — where AI processes large-scale data and medical professionals make the final decisions — can deliver significant gains in clinical settings,” said Hwang Tae-hyun, a professor at Vanderbilt University Medical Center.
The research team plans to expand the system beyond gastric cancer to other major cancers, including colorectal and lung cancer.
LG AI Research and Vanderbilt are also exploring partnerships with global pharmaceutical companies and university hospitals to deploy the technology in real-world clinical environments.
“LG aims to build a ‘brain’ where AI agents collaborate with medical professionals to advance personalized cancer treatment,” said Jang Jong-seong, head of the Bio-Intelligence Lab at LG AI Research. “We aim to shorten the current four-week process of diagnosis and treatment planning to a single day.”
The initiative reflects LG’s broader push into AI and biotechnology, identified by Chairman Koo Kwang-mo as key pillars for future growth and for enhancing quality of life.
By Lee Han-gyoul (stlee0329@heraldcorp.com)
