June 17, 2026 | Randles Lab
Amanda Randles delivered the Midweek Keynote at ISC High Performance 2026, highlighting how advances in HPC and AI are enabling vascular digital twins to move toward longitudinal, proactive healthcare.
Amanda Randles delivered the Midweek Keynote at ISC High Performance 2026, presenting a vision for how high performance computing can enable vascular digital twins capable of transforming how cardiovascular health is modeled, monitored, and ultimately managed.
Digital twins offer the potential to move healthcare beyond isolated clinical snapshots toward continuously evolving computational representations of individual patients. Randles discussed how patient-specific models of the cardiovascular system could help detect changes earlier, track disease progression over time, and create opportunities for intervention before clinical deterioration occurs.
Realizing this vision, however, requires major advances in computational scale. Cardiovascular physiology spans spatial scales from individual blood cells to entire vascular networks and temporal scales from fractions of a heartbeat to months or years of disease progression. Randles highlighted work from the lab aimed at overcoming these challenges through scalable simulation, longitudinal hemodynamic modeling, and approaches such as adaptive physics refinement that deploy computational detail where and when it is needed.
The keynote also explored the increasingly complementary roles of AI and HPC. Machine learning can accelerate components of patient-specific modeling and help interpret increasingly large volumes of longitudinal data, while physics-based simulation provides mechanistic information and an important foundation for validating predictions. Together, these capabilities can support digital twins that are both computationally tractable and grounded in physiology.
The talk highlighted the Randles Lab’s broader goal of moving cardiovascular modeling from simulations of individual moments toward predictive models of physiological trajectories—using HPC, AI, and mechanistic modeling to help shift healthcare from reactive treatment toward proactive, personalized care.