South Korean Researchers Develop AI Model for Customized Cancer Vaccine

The AI model predicts which neoantigens are likely to trigger strong B cell responses by learning structural interaction patterns between mutant peptides and B cell receptors.

South Korean researchers have developed a new artificial intelligence model aimed at creating customised cancer vaccines by identifying neoantigens unique to individual patients.

The joint research team from the Korea Advanced Institute of Science and Technology and biotechnology firm Neogenlogic said the AI platform helps detect mutation-derived protein fragments that can train a patient’s immune system and help prevent recurrence of cancer, according to Yonhap news agency.

“Neoantigens — mutation-derived protein fragments unique to a patient's tumor — are the 'fingerprints' used by vaccines to train the immune system,” the team, led by professor Choi Jung-kyoon of KAIST, said in a press release.

The researchers said existing cancer vaccines largely focus on activating cytotoxic T cells, while emerging clinical evidence shows that B cell-mediated immune memory plays a critical role in long-term antitumor responses and recurrence prevention.

“While current vaccines focus almost exclusively on activating cytotoxic T cells for immediate attack, emerging clinical evidence highlights that B cell-mediated immune memory is the key to durable, long-term antitumor responses and the prevention of recurrence,” the team said.

According to the researchers, the AI model predicts which neoantigens are likely to trigger strong B cell responses by learning structural interaction patterns between mutant peptides and B cell receptors.

“The study introduces the world's first AI framework capable of predicting B cell immunogenicity alongside T cell responses for the design of personalized cancer vaccines,” the team said.

Choi said the study provided empirical evidence supporting the importance of B cells in cancer vaccine development. “While the academic community was aware that studying B cells is important in developing cancer vaccines, there were no tools to verify the concept,” he told Yonhap News Agency.

Neogenlogic said the technology has been validated using large-scale genomic datasets and clinical trial data, and has been integrated into its proprietary discovery engine DeepNeo.

Choi said the team is preparing an investigational new drug submission with the U.S. Food and Drug Administration and plans to begin clinical trials in 2027. “Together with Neogenlogic, we are translating this academic breakthrough into a clinical-grade platform,” he said. “Our proprietary AI elevates the scientific rigor of neoantigen selection, moving us from theoretical prediction to systematic clinical application.”

The findings were published in the December 3 edition of the peer-reviewed journal Science Advances.

Stay tuned for more such updates on Digital Health News