Breakthrough simulation sheds light on blood clot formation
- Babak Baghaei
- Feb 20
- 2 min read
Cardiovascular diseases remain one of the leading global health challenges, with uncontrolled thrombus formation at the core of many life-threatening events. Understanding how clots initiate, grow and interact with medical devices is essential — yet direct in-vivo measurement remains limited, costly and difficult.
A new paper published in Physics of Fluids (AIP Publishing), led by PhD researcher Sumanta Laha, and co-supervised by Dr Georgios Fourtakas, Prof Das, and Dr Amir Keshmiri (Founder and Technical Director of Mansim), introduces a major step forward: a novel GPU-optimised Smoothed Particle Hydrodynamics (SPH) model for simulating thrombus formation and growth under realistic haemodynamic conditions.
A new SPH methodology for thrombosis
The study presents a fully integrated SPH-based framework capable of modelling:
the coagulation cascade
platelet activation and aggregation
biochemical concentration fields (thrombin, prothrombin, fibrinogen, fibrin)
mechanical interactions with vessel walls
wall shear stress effects, a critical trigger for device-induced thrombosis
By implementing the model inside DualSPHysics — a widely used open-source SPH solver — the authors ensure reproducibility, accessibility and future extensibility across the cardiovascular-modelling community.
GPU acceleration for faster haemodynamic simulations
One of the main barriers in thrombosis modelling is computational cost. This work directly addresses that by optimising the SPH framework for GPU execution, reducing the time required for long-duration thrombus-growth simulations.
GPU acceleration makes it feasible to:
simulate longer physiological time scales
model complex geometries (medical devices, bifurcations, microchannels)
test parametric variations more efficiently
support real-time or near-real-time predictive tools in the future
Penalty vs. Dissipation growth strategies
The study evaluates two mechanistic approaches to model thrombus growth:
1. Penalty-based approach
Imposes a fibrin-linked velocity penalty, mimicking the mechanical restriction as clotting material accumulates.
2. Dissipation-based approach
Links Einstein’s viscosity equation with fibrin concentration, increasing local resistance as the clot evolves.
By comparing both methods across controlled test cases, the authors identify their respective strengths and suitability for different physiological or device-driven scenarios.
Validation in realistic flow environments
The model is validated using two widely studied configurations:
a backward-facing step, where shear gradients strongly influence coagulation
a microchannel, representing small-scale biomedical environments
Both cases demonstrate the framework’s capability to capture:
thrombus initiation and growth
interplay of shear stress and biochemical activation
spatial clot morphology
sensitivity to boundary conditions and flow regime
Why this research matters
The study strengthens computational modelling as a powerful alternative or complement to in-vivo and in-vitro approaches, providing:
deeper understanding of device-induced thrombosis
virtual testing for cardiovascular implants (valves, stents, grafts, LVADs)
predictive tools for improving medical-device safety
insight into clot–flow interactions not easily measurable experimentally
With cardiovascular disease continuing to place major demands on healthcare systems, advances like this help accelerate safer device design, personalised treatment strategies, and improved clinical outcomes.
Comments