Harnessing single cell image-transcriptomics and bespoke AI for a deeper understanding of tumour-microenvironment interactions

Ovarian cancer is a deadly disease with a substantial unmet need for more targeted therapies. Changes in epithelial organisation and tumour microenvironment (TME) alter cellular communications and influence patient response to treatment. However, the exact mechanisms involved are not well understood because of the complexity of tumour-TME interactions, and the difficulty in modelling them using current approaches. The overarching goal of this fellowship proposal is to build integrative AI models of how the diverse cell types in the TME interact and how epithelial organisation influences these interactions. I will use multiplexed tissue imaging to tag tens of proteins in the same tumour simultaneously to visualise spatial protein activities across millions of cells. Building on my highly interdisciplinary background in AI, cancer biology, and image-based modelling, I will characterise TME phenotypes and their therapeutic relevance from multiplexed images. Novel methods will be developed to integrate single-cell and spatial transcriptomics data to determine genetic and molecular factors underlying these phenotypes. I will further identify biomarkers and phenotypes that stratify patient responses to therapy using AI. The developed methodologies will provide new conceptual frameworks for modelling the TME, and cellular interactions and greatly advance our understanding of their role in resistance and relapse.