Tumor Ensemble-Based Modeling and Visualization of Emergent Angiogenic Heterogeneity in Breast Cancer

Spyros K Stamatelos,Eugene Kim,Arvind P Pathak,Akanksha Bhargava,Aleksander S Popel

doi:10.1038/s41598-019-40888-w

Abstract

There is a critical need for new tools to investigate the spatio-temporal heterogeneity and phenotypic alterations that arise in the tumor microenvironment. However, computational investigations of emergent inter- and intra-tumor angiogenic heterogeneity necessitate 3D microvascular data from ‘whole-tumors’ as well as “ensembles” of tumors. Until recently, technical limitations such as 3D imaging capabilities, computational power and cost precluded the incorporation of whole-tumor microvascular data in computational models. Here, we describe a novel computational approach based on multimodality, 3D whole-tumor imaging data acquired from eight orthotopic breast tumor xenografts (i.e. a tumor ‘ensemble’). We assessed the heterogeneous angiogenic landscape from the microvascular to tumor ensemble scale in terms of vascular morphology, emergent hemodynamics and intravascular oxygenation. We demonstrate how the abnormal organization and hemodynamics of the tumor microvasculature give rise to unique microvascular niches within the tumor and contribute to inter- and intra-tumor heterogeneity. These tumor ensemble-based simulations together with unique data visualization approaches establish the foundation of a novel ‘cancer atlas’ for investigators to develop their own in silico systems biology applications. We expect this hybrid image-based modeling framework to be adaptable for the study of other tissues (e.g. brain, heart) and other vasculature-dependent diseases (e.g. stroke, myocardial infarction).

Highlights

The origins of microvascular heterogeneity in different organs have been investigated in numerous studies using experimental and computational approaches[6,7,8,9,10,11]
We characterized the emergent phenotype arising from these different tumor microenvironments based on their vascular morphology and hemodynamics for the entire tumor ensemble and discovered that hemodynamic parameters alone were insufficient for classifying these different tumor microenvironments
The soft tissue contrast acquired from magnetic resonance microscopy (MRM) imaging (~40 μm resolution) is shown in grey and the 3D microvascular network acquired from micro-CT imaging (8 μm resolution) is shown in a red color map scaled by the mean vessel diameter

Summary

Introduction

The origins of microvascular heterogeneity in different organs have been investigated in numerous studies using experimental and computational approaches[6,7,8,9,10,11]. Combinations of imaging and mathematical modeling permit detailed vessel-by-vessel simulations of microvascular blood flow and oxygen transport[15,17,18] These advances have been accompanied by the development of new computational approaches for modeling hemodynamics, the transport of oxygen and other molecules such as growth factors and nanoparticle-mediated drug transport in realistic tumor vasculatures[19]. Traditional image-based computational frameworks are not well suited to modeling the emergent phenotypic heterogeneity that arises from the whole-tumor microvasculature because they are limited by the sparse spatial coverage of the underlying tumor microvascular network (usually of the order of

Objectives

Methods

Results

Conclusion