By Erin Baudendistel-Happ
What if doctors could simulate your cardiovascular system before deciding on a treatment?
This is no longer a distant vision. At ISC 2026, Dr. Amanda Randles will demonstrate how high performance computing (HPC) is making this possible through vascular digital twins – one of the most compelling and rapidly advancing applications of HPC, where simulation, data and medicine converge.
As Director of the Duke Center for Computational and Digital Health, Randles works at the intersection of HPC, physics, and medicine. Her research is driven by a powerful idea: with sufficiently accurate simulations of the human body, we can better understand and even anticipate disease before it manifests.
At the center of her research is HARVEY, a specialized HPC platform that Randles and her team developed to simulate blood flow across the entire human vascular system. Running on advanced supercomputers, HARVEY combines massively parallel computing and sophisticated fluid-dynamics models to capture both large-scale flow in arteries and veins and microscopic interactions between individual blood cells and vessel walls. Modeling billions of interacting particles over time, it pushes the limits of biomedical simulation and reflects the true complexity of blood as a multiscale system.
Building on this foundation, the Randles Lab develops image-based digital twins of the human vascular system. These models integrate physics-based simulation, artificial intelligence, and clinical imaging data to replicate blood flow through an individual patient’s body and the evolution of cardiovascular disease over time.
A key shift in this work is the move from static models to continuously evolving simulations. As Randles explains: “We’re trying to capture what’s happening as you move through daily life.”
This shift is enabled by innovations such as Longitudinal Hemodynamic Mapping, a framework that simulates millions of heartbeats over weeks or months. By incorporating continuous data from wearable devices, these digital twins reflect a patient’s changing physiological state, bringing simulation closer to clinical use.
While vascular digital twins are advancing rapidly, significant challenges remain before doctors can use them. Randles’ work highlights the need for extreme-scale computing resources, as well as reliable, high-quality data streams, specifically from wearable devices, to ensure accuracy. Early studies show that these models can match invasive diagnostic methods such as fractional flow reserve measurements and correlate well with ultrasound and pressure data, suggesting strong potential for clinical use. However, broader adoption will require large-scale validation, improved computational efficiency and standardized frameworks. At the same time, important ethical considerations, including data privacy, model transparency, and potential bias, must be addressed as digital twins move closer to real-world healthcare applications.
The broader vision represents a fundamental transition from reacting to disease to anticipating it. Instead of waiting for symptoms to appear, doctors could identify risks earlier, test interventions in simulation, and tailor treatments to the individual patient. For the HPC community, this marks an important evolution, particularly extending HPC into healthcare with direct human impact.
In her keynote, “HPC for Vascular Digital Twins,” Randles will explore how these technologies are coming together and what challenges remain in bringing digital twins into the healthcare system.
📍 Wednesday, June 24, 2026, in Hamburg, Germany
