Title:	Foundation models meet graph-based learning to advance spatial biology towards patient-specific cancer immunotherapy
DNr:	Berzelius-2025-315
Project Type:	LiU Berzelius
Principal Investigator:	Natasa Sladoje <natasa.sladoje@it.uu.se>
Affiliation:	Uppsala universitet
Duration:	2025-09-18 – 2026-04-01
Classification:	20208
Homepage:	https://www.scilifelab.se/data-driven/ddls-research-school/rs-phd-projects/
Keywords:

Abstract

Immunotherapy has become a life-saving option for advanced cancer patients. However, only a minority of patients develop a durable response. Despite great efforts to explain the variable responses to immunotherapy and to optimize patient selection, the currently used clinical biomarkers demonstrate only modest predictive performance. This project is a key part of our broader initiative to employ innovative, powerful, yet interpretable data-driven analysis methods, significantly advancing our understanding of immune cell inter-relations within the cancer microenvironment. We will leverage state-of-the-art foundation models that excel in digital pathology tasks, to maximize the information gain from in situ multiplex tissue profiling. By integrating correlated structural and molecular analysis into the natural 3D tissue space, we aim to exceed the current state of the art in histopathology. Our approach will enable reliable prediction of patient response to immunotherapy and facilitate optimized personalized cancer treatment.