Despite striking results, clinical outcome with immune checkpoint inhibitors remains too often uncertain. This joint-project aims at generating dense longitudinal data in lung cancer patients undergoing anti-PD1 or anti-PDL1 therapy, alone or in combination with other anticancer agents. Mathematical modelling with mechanistic learning algorithms will be used next to decipher the mechanisms underlying response or resistance to immunotherapy. Ultimately, this project should help to better understand the mechanisms underlying resistance to immune checkpoint inhibitors and identify a serial of actionable items to increase the efficacy of immunotherapy.