High-throughput cancer hypothesis testing with an integrated PhysiCell-EMEWS workflow

Abstract

BackgroundCancer is a complex, multiscale dynamical system, with interactions between tumor cells and non-cancerous host systems. Therapies act on this combined cancer-host system, sometimes with unexpected results. Systematic investigation of mechanistic computational models can augment traditional laboratory and clinical studies, helping identify the factors driving a treatment’s success or failure. However, given the uncertainties regarding the underlying biology, these multiscale computational models can take many potential forms, in addition to encompassing high-dimensional parameter spaces. Therefore, the exploration of these models is computationally challenging. We propose that integrating two existing technologies—one to aid the construction of multiscale agent-based models, the other developed to enhance model exploration and optimization—can provide a computational means for high-throughput hypothesis testing, and eventually, optimization.ResultsIn this paper, we introduce a high throughput computing (HTC) framework that integrates a mechanistic 3-D multicellular simulator (PhysiCell) with an extreme-scale model exploration platform (EMEWS) to investigate high-dimensional parameter spaces. We show early results in applying PhysiCell-EMEWS to 3-D cancer immunotherapy and show insights on therapeutic failure. We describe a generalized PhysiCell-EMEWS workflow for high-throughput cancer hypothesis testing, where hundreds or thousands of mechanistic simulations are compared against data-driven error metrics to perform hypothesis optimization.ConclusionsWhile key notational and computational challenges remain, mechanistic agent-based models and high-throughput model exploration environments can be combined to systematically and rapidly explore key problems in cancer. These high-throughput computational experiments can improve our understanding of the underlying biology, drive future experiments, and ultimately inform clinical practice.

Optimizing Radiation Therapy Treatments by Exploring Tumour Ecosystem Dynamics in – silico

“…Population dynamics of tumours have been further studied by Gonzales-Garcia et al (2002) and Sole (2003) who concluded through agent-based modelling that spatial genetic heterogeneity observed in tumours naturally follows from simple ecological competitor dynamics. Since then, many multi-agent models of tumour growth with increasing physical accuracy have been proposed (Abbott et al 2006, Zhang et al 2009, Bentley 2013, Ozik et al 2018. Yet, few of these approaches take into account the dynamics of and interaction with the tumour environment, particularly host tissue, immune system, and boundary conditions imposed by anti-cancer therapies.…”

Supporting Computational Apprenticeship through educational and software infrastructure. A case study in a mathematical oncology research lab

“…These approaches, however, neglect the complex interactions in the evolving multi-level networks of normal and diseased tissues [35]. Targeted interventions do not only affect just single cell types; biochemical and biophysical feedbacks-combined with intercellular heterogeneity and natural selection [32] and amplified by physical constraints [36]-can cause secondary effects such as therapeutic resistance (e.g., by selecting for resistant cancer clones), worsened drug delivery, and treatment toxicity [6,32,37]. Thus, next-generation therapies must not just treat single cell types, but rather steer the multicellular systems towards balance.…”

“…In particular, computational modeling approaches, which include both mathematical modeling of complex cell and molecular systems and statistical modeling of large datasets, are a powerful vehicle for synthesizing disparate and sometimes conflicting data into an integrated biological understanding. Ideally, computational modeling approaches work recursively with experimental workflows; mathematical models quantitate and formalize the largely qualitative, observation-driven "mental models", and the iterative comparison of model outputs to experimental data informs model refinement and also suggests new experimental directions [6][7][8][9].The use of mathematical models clarifies the biological conditions or parameters under which the "mental model" can explain the experimental and simulation data. Increasingly, statistical modeling approaches including machine learning and bioinformatics are used to complement mathematical modeling of cell and molecular biological systems [7].…”

Towards an evolvable cancer treatment simulator

“…This began the design of cell-cell interaction rules to create a multicellular cargo delivery system that actively delivers a cancer therapeutic beyond regular drug transport limits to hypoxic cancer regions. These model rules were manually tuned to achieve this (as yet unoptimised) design objective, requiring weeks of people-hours to configure, code, test, visualise, and evaluate (Ozik et al, 2018). Ghaffarizadeh et al (2018) also presented 3-D simulations of cancer immunotherapy.…”

“…The results provided insights into therapeutic failure, thus demonstrating the potential of high-throughput computing to investigate high dimensional cancer simulator parameter spaces. High-throughput model investigation and hypothesis testing provides a new paradigm for solving complex problems, gaining new insights, and improving cancer treatment strategies (Ozik et al, 2018).…”