Inference of Human Tumor Growth Dynamics Through Integrative Genomics and Spatial Mathematical Modeling

• Activity Number: 101
• Type: Invited
• Date/Time: Monday, August 4, 2014 : 8:30 AM to 10:20 AM
• Sponsor: WNAR
• Author(s): Christina Curtis*+
• Companies: University of Southern California Keck School of Medicine
• Keywords: cancer ; genomics ; tumor evolution ; heterogeneity ; modeling ; inference

Abstract

Although the direct observation of human tumor origins is impractical, the ancestral relationships between tumor cells are recorded in mutations acquired during somatic cell division. As such, the signature of the primordial tumor can be recovered because alterations in the first cell divisions may remain pervasive in the final neoplasm. We propose a Big Bang model, whereby a tumor grows predominantly as a single expansion, producing heterogeneous sub-clones. Multi-scale genomic profiling of 349 individual colorectal tumor glands indicated the absence of selective sweeps, uniformly high intra-tumor heterogeneity, and variegation in distant tumor regions, as postulated by the model. To test the predictions of this model we developed a statistical inference framework based on Approximate Bayesian Computing and 3-dimensional mathematical modeling. By analyzing genomic data within this framework, we demonstrate that most observable intra-tumor heterogeneity originates before the neoplasm is detectable, irrespective of microenvironmental effects or sub-clone fitness changes. Our findings provide new insights into tumor evolution with fundamental clinical implications.