Lead PI:
Center Researchers:
The Challenge
Traditional methods of cardiac health monitoring provide a limited snapshot of an individual’s cardiovascular condition. While early digital twins have captured 3D flow profiles, they have been restricted to short timescales—previously, the longest published 3D simulation covered just 30 heartbeats. This limitation presents a major challenge in understanding how cardiovascular function evolves over weeks, months, or even years.
As a result, subtle but critical physiological changes that could indicate early disease progression often go undetected. Without continuous, long-term insights, treatment strategies may be less personalized and reactive rather than proactive. Additionally, conventional monitoring is typically confined to clinical settings, limiting the ability to track cardiovascular health during real-world activities and across different physiological states.
Our Solution
Through the Duke Center for Digital and Computational Health Innovation team, the Randles Lab has developed Longitudinal Hemodynamic Maps. This groundbreaking project represents a significant advancement in cardiac health monitoring by simulating millions of heartbeats, providing a comprehensive view of cardiac function over extended periods—from weeks to months. Ours was the first model of 3D flow over six weeks.
The longitudinal approach allows for a more detailed and accurate assessment of cardiovascular conditions, facilitating early intervention and the development of personalized treatment strategies.
A key advancement of this work is the direct integration of data from wearable devices, which continuously stream physiological signals to drive the simulation and generation of hemodynamic units (HUs). Unlike traditional static modeling approaches, LHMs are dynamically informed by real-world activity data, ensuring that simulations reflect the evolving physiological state of each patient over time.
This framework has already led to major computational breakthroughs:
- First-ever 3D hemodynamic simulation spanning six weeks, far exceeding prior state-of-the-art 30-heartbeat models.
- Seamless integration of real-time wearable biosensor data to continuously update digital twin models.
- High-throughput simulation of HUs in parallel, dramatically increasing computational efficiency.
Building on their work on HARVEY, a specialized software package that simulates patient-specific blood flow models, the Randles Lab enabled parallel simulation of hemodynamic units (HUs), making the computational process much more efficient.
By integrating Longitudinal Hemodynamic Maps with wearable technology, the Duke Center for Digital and Computational Health Innovation is at the forefront of digital health. This project is setting critical foundation to transform the way we understand, monitor, and address cardiac health.