Computational Modelling of Atherosclerosis

Atherosclerosis is one of the principle pathologies of cardiovascular disease with blood cholesterol a significant risk factor. The World Health Organisation estimates that approximately 2.5 million deaths occur annually due to the risk from elevated cholesterol with 39% of adults worldwide at future risk. Atherosclerosis emerges from the combination of many dynamical factors, including haemodynamics, endothelial damage, innate immunity and sterol biochemistry. Despite its significance to public health, the dynamics that drive atherosclerosis remain poorly understood. As a disease that depends on multiple factors operating on different length scales, the natural framework to apply to atherosclerosis is mathematical and computational modelling. A computational model provides an integrated description of the disease and serves as an in silico experimental system from which we can learn about the disease and develop therapeutic hypotheses. Although the work completed in this area to-date has been limited, there are clear signs that interest is growing and that a nascent field is establishing itself. This paper discusses the current state of modelling in this area, bringing together many recent results for the first time. We review the work that has been done, discuss its scope and highlight the gaps in our understanding that could yield future opportunities.

💡 Research Summary

Atherosclerosis remains one of the leading contributors to cardiovascular mortality worldwide, with elevated blood cholesterol identified as a major modifiable risk factor. The World Health Organization estimates that roughly 2.5 million deaths each year are linked to high cholesterol levels, and that about 39 % of adults worldwide face an increased future risk. Despite its public‑health importance, the disease is driven by a complex interplay of processes that operate across vastly different spatial and temporal scales: hemodynamic forces acting on the arterial wall, endothelial injury, innate immune responses, and sterol biochemistry. Because a single‑scale perspective cannot capture this multi‑factorial nature, the authors argue that mathematical and computational modeling provides the most natural framework for integrating these mechanisms into a coherent, testable representation of disease progression.

The paper begins with a concise overview of the biological background, emphasizing how low shear stress regions in arteries promote endothelial dysfunction, increase permeability to low‑density lipoprotein (LDL), and facilitate oxidative modification of LDL (OxLDL). OxLDL, in turn, triggers recruitment of monocytes, differentiation into macrophages, and formation of foam cells—key events that seed the atherosclerotic plaque. The authors then categorize existing computational approaches into three primary layers.

1. Hemodynamic Modeling (Computational Fluid Dynamics, CFD).
   High‑resolution patient‑specific vascular geometries derived from MRI or CT are used to solve the Navier‑Stokes equations under pulsatile boundary conditions that mimic the cardiac cycle. CFD quantifies wall shear stress (WSS) and identifies low‑WSS zones that correlate with plaque initiation in both animal experiments and clinical imaging studies. Some models also incorporate fluid‑structure interaction to capture how plaque growth alters arterial compliance, creating a feedback loop between flow and wall mechanics.

2. Cellular / Immune Modeling (Agent‑Based Models, ABM, and PDEs).
   Immune cells such as monocytes, macrophages, dendritic cells, and T‑lymphocytes are represented as discrete agents that sense local cues—WSS, OxLDL concentration, chemokine gradients—and execute rule‑based behaviors (migration, activation, cytokine secretion). ABM simulations reproduce the spatial heterogeneity observed in early lesions, including the clustering of foam cells and the formation of necrotic cores. Coupled reaction‑diffusion equations describe the spread of cytokines (e.g., IL‑1β, MCP‑1) and the degradation of extracellular matrix by matrix metalloproteinases.

3. Lipid Metabolism Modeling (ODEs and Reaction‑Diffusion).
   Systemic cholesterol dynamics are captured by ordinary differential equations that track plasma LDL, high‑density lipoprotein (HDL), hepatic synthesis, intestinal absorption, and receptor‑mediated uptake. Parameters are calibrated against clinical lipid panels and, where available, genetic knock‑out mouse data. Some studies extend these equations to the arterial wall, linking local LDL influx (modulated by endothelial permeability) to intracellular cholesterol ester accumulation.

The authors highlight several pioneering multiscale integration efforts that couple these layers. For example, CFD‑derived WSS maps feed into the ABM to modulate endothelial injury rates; the resulting increase in endothelial permeability raises LDL influx in the lipid metabolism module, which then accelerates foam‑cell formation in the cellular model. Conversely, plaque growth alters arterial geometry, feeding back into the CFD simulation and reshaping the WSS landscape. These bidirectional couplings generate nonlinear feedback loops that can reproduce the rapid acceleration of plaque development observed clinically after a threshold of lipid accumulation is reached.

Validation strategies are discussed in depth. In‑vitro flow chambers and ex‑vivo arterial segments provide experimental data on shear‑dependent endothelial responses, while animal models (e.g., ApoE‑/‑ mice) offer longitudinal measurements of plaque size, composition, and inflammatory markers. Human cohort studies contribute population‑level correlations between LDL trajectories and cardiovascular events, which are used to benchmark model predictions. However, the review notes that most existing models have been validated only at a single scale, and that comprehensive multiscale validation remains scarce due to limited availability of high‑resolution, longitudinal multimodal datasets.

A substantial portion of the paper is devoted to identifying gaps and future opportunities. The authors argue that the field suffers from a lack of standardization—different groups use disparate nomenclature, parameter sets, and numerical solvers, making cross‑study comparison difficult. Uncertainty quantification is rarely performed; Bayesian inference or Markov‑chain Monte Carlo methods could systematically assess parameter sensitivity and propagate measurement error through the model hierarchy. The authors also stress the need for data integration pipelines that combine imaging, genomics, proteomics, and metabolomics to enable truly personalized simulations. High‑performance computing platforms, possibly coupled with machine‑learning‑driven surrogate models, would allow exploration of large therapeutic parameter spaces (e.g., timing and dosage of statins, anti‑inflammatory agents, PCSK9 inhibitors) in silico before clinical trials.

In conclusion, the review positions computational modeling of atherosclerosis as an emerging, interdisciplinary field that has already yielded valuable mechanistic insights at isolated scales. By advancing multiscale coupling, rigorous validation, and uncertainty analysis, future models could serve as virtual laboratories for hypothesis testing, drug development, and individualized risk assessment. The authors are optimistic that, with coordinated efforts in data sharing, methodological standardization, and computational infrastructure, the nascent field will mature rapidly and become an indispensable tool in the fight against cardiovascular disease.