Lunit Inc. says its artificial intelligence software can better predict which rare cancer patients will respond to immunotherapy, potentially helping doctors make crucial treatment decisions.
The South Korean company’s findings, presented at a medical conference in Houston, suggest patients with specific AI-detected patterns in their tumor samples had up to 72% lower risk of death when treated with Merck & Co.’s pembrolizumab.
The research, conducted with MD Anderson Cancer Center, analyzed more than 500 tissue samples across 10 different rare tumor types. Patients whose samples showed higher levels of immune cells before treatment had a 51% lower risk of disease progression, according to the study.
Brandon Suh, Lunit’s chief executive officer, indicated the technology could help personalize treatment for patients with limited options. The study focused on pembrolizumab, part of a class of drugs called immune checkpoint inhibitors that has transformed cancer care but doesn’t work for all patients.
The research comes as medical technology companies race to develop AI tools for healthcare, though many such advances still need validation through larger clinical trials before widespread adoption.
