Multiscale discovery of cellular and microanatomical determinants of metastasis

Metastasis is the major cause of cancer-related deaths. About 30% of breast cancer patients develop metastasis. While some tumor-cell intrinsic determinants of metastasis are known, the impact of the microenvironmental context and tissue organization remains largely unclear. Single-cell technologies capture the heterogeneity of tumor, immune, and stromal cells but are mostly applied to primary tumors, and fail to reconstruct the local and tissue-level architecture of metastasis. We hypothesize that metastases are determined by the complex interplay of tumor cells and their TME and that cellular phenotypes are interconnected with the broader tissue-level architecture extending beyond the local TME context. To identify tissue-level determinants of metastasis and novel anti-metastatic targets we will leverage matched primary and metastatic tumor samples from breast cancer patients and from patient-derived xenograft models with different metastatic potential. We will create a multimodal, multiscale AI/ML model that predicts metastatic potential from cell-intrinsic, microenvironmental, and microanatomical features. This will enable us to discover signatures associated with metastatic potential, which we will validate in large-scale human cohorts of patients with known metastatic outcomes. Finally, we will test identified molecular predictors in vivo as potential therapeutic targets to prevent metastasis. This project combines the expertise of cancer- and computational biologists, and pathologists to empower a new understanding of the metastatic process across anatomical scales and develop biomarkers for breast cancer stratification and personalized therapy recommendation.