This talk will explore how computational modeling—initially developed during my time as a DOE CSGF fellow—has evolved into digital twin technologies now aimed at transforming clinical care. I’ll begin by revisiting early vascular simulations built under CSGF support and show how these foundational efforts laid the groundwork for creating patient-specific models used today to predict and manage cardiovascular disease. We’ll examine the technical core of digital twins in healthcare: combining high-fidelity fluid dynamics, wearable sensor integration, and machine learning to simulate individual physiology in real time. These models enable noninvasive assessments of conditions like arterial stenosis, with implications for diagnostics, treatment planning, and long-term monitoring. I’ll also address computational and translational challenges, including data scale, real-time constraints, and the need for hybrid cloud-HPC platforms to enable deployment at scale. Along the way, I’ll reflect briefly on how key moments in the CSGF program—including mentorship, practicum experiences, and community connections—have shaped both the scientific questions I pursue and the broader vision for impact. This intersection of computational science and human health continues to push the boundaries of what’s possible—and CSGF has played a defining role in that journey.
Evolving Models: From Graduate Simulations to Scalable Digital Twins in Healthcare
Presenter:
Amanda
Randles
University:
Duke University
Program:
CSGF
Year:
2025