Lung cancer is the leading cause of cancer-related mortality in France, with over 30,000 deaths annually. Immunotherapy, which stimulates the immune response against tumor cells, offers tremendous hope for combating the disease. Unfortunately, only about half of eligible patients respond to this treatment, and it remains challenging to identify responders in advance to administer the most appropriate therapy from the outset.
Three researchers from Pr[AI]rie — Emmanuel Barillot (Institut Curie-INSERM), Thomas Walter (Mines ParisTech), and Irène Buvat (INSERM-Institut Curie) — along with a pneumooncologist (Nicolas Girard, Institut Curie), have been working together to improve patient care. Their approach integrates all available information, including clinical data, MRI and CT imaging, anatomical pathology (tumor slice images), and genomics (gene mutations and expression).
The initial results, published in the journal Nature Communications, combine molecular, cellular, tissue-level, and overall patient data into a statistical learning model to predict patient response to immunotherapy. The project is funded by ARC (via the SIGN’IT call for proposals) and the Pr[AI]rie Institute.