Longitudinal Hemodynamic Mapping Framework Is a Breakthrough in Cardiovascular Medicine

To truly understand and treat progressive conditions like cardiovascular disease, we need more than a snapshot — we need the full picture. The longer we can track a patient’s physiological data, the clearer and more actionable their cardiovascular profile becomes.

A quick look at the way blood flows inside the heart and blood vessels (hemodynamics) tells us some things, but continual monitoring and data collection over time reveal much more.

The pioneering Longitudinal Hemodynamic Mapping Framework (LHMF) — developed in my lab — gives us that longer, more detailed look.

The strength of LHMF lies in two key advances: the real-time integration of physiological data from wearable biosensors and the development of a novel algorithmic framework that enables long-term blood flow simulation at scale. Together, they allow us to move beyond isolated measurements and toward a complete, time-resolved understanding of cardiovascular health.

A critical part of the Duke Center for Computational and Digital Health Innovation’s mission to find, track, and treat human disease, the framework integrates — in real time — data generated by a person’s wearable with his or her digital twin.

This novel approach to cardiovascular medicine enables us to collect highly detailed, individualized information for weeks or months at a time, far beyond anything previously possible.

The clinical value (and challenges) of getting a complete hemodynamic picture

The way blood flows through our cardiovascular system (hemodynamics) is constantly changing, adapting to evolving physiological states and needs.

Factors that influence hemodynamics include: 

• Heart function
• Blood pressure
• Physical activity

Whether these factors are normal (like those caused by aging) or abnormal (like those caused by conditions like cardiovascular disease), their effects on blood flow develop and become detectable over time.

That’s why long-term data collection and monitoring are essential. The longer we can gather, integrate, and analyze patients’ real-time physiological data, the more we can learn about their personal cardiovascular function, risk factors, and, when applicable, disease progression — and the more targeted and effective their treatment can become

Ongoing monitoring can also uncover subtle patterns and trends that single-heartbeat or point-in-time measurements may miss — helping clinicians make more informed, proactive, and data-driven therapeutic decisions.

Historically, however, cardiovascular monitoring has offered only brief, time-limited snapshots of heart and vessel function. Prior to LHMF, even the most advanced 3D blood flow simulations covered under 50 heartbeats. This limited temporal scope has made it nearly impossible to monitor how a particular patient’s hemodynamics evolve over time — and likely caused early indicators of disease or its progression to go undetected.   

Adding to the challenge, most conventional monitoring is done in clinical settings, making it difficult to assess how blood flow responds during daily life.

LHMF overcomes these constraints by providing longitudinal data with a level of detail and personalization that was previously out of reach. It enables patient-specific, remote monitoring based on comprehensive physiological insight. 

Another obstacle to analyzing blood flow models over time has been the enormous amount of computational power necessary to create long-term 3D hemodynamic simulations. Not to mention the literal months and years it takes; by the time all of the data can be captured and examined — literally months or years — it may be too late.

A smarter platform for cardiovascular care

By removing key challenges to long-term blood flow analysis, LHMF represents a sea change in cardiovascular medicine. One of LHMF’s most significant breakthroughs is the introduction of a parallel-in-time simulation approach based on discrete hemodynamic units (HUs), meaning millions of heartbeats can be modeled and analyzed simultaneously — not sequentially. What used to take months or years to compute can now be processed in a fraction of the time, without losing clinical fidelity.

The integration of this algorithmic framework with real-time wearable data is what elevates LHMF from a personalized model to a dynamic digital twin — one that continually adapts to a person’s changing physiological status and enables ongoing, individualized care.

These wearable technologies and biosensors — from fitness trackers, smartwatches, and rings to implanted sensors and biometric clothing — monitor and transmit physiological metrics such as heart rate, cardiac electrical signals, and blood pressure. Their data feeds directly into the framework, enabling long-term tracking and simulation of cardiovascular health outside of clinical settings.

By combining real-world, real-time data with high performance and cloud-based computing, LHMF empowers care teams to proactively monitor, diagnose, and treat cardiovascular disease —  making earlier interventions and better outcomes possible.

Our framework is the first 3D blood flow simulation capable of capturing more than six weeks of data — over 4.5 million heartbeats — far surpassing prior models limited to just a few beats. This extended, high-resolution view is continuously informed by data from wearable biosensors, allowing each patient’s digital twin to reflect their current physiological state in real time.

Because these exact patient replicas can detect even very subtle changes in blood flow over time, they give clinicians highly refined insights for making clinical decisions — and may help them detect cardiovascular disease and related complications before they cause symptoms.

Expanding on our work with HARVEY (software that creates patient-specific blood flow simulations), we enabled the framework to perform parallel HU simulation, greatly improving the efficiency of computations.

By uniting wearable technologies with state-of-the-art computational methods, we are leading the charge in developing a new, futuristic type of patient-specific care.

Our Longitudinal Hemodynamic Mapping Framework is a transformational advance in cardiovascular monitoring and a forward-thinking approach to healthcare today. Learn more about collaborating with us to drive progress in computational health innovation.

Amanda Randles, Ph.D., is Director of the Duke Center for Computational and Digital Health Innovation and Alfred Winborne Mordecai and Victoria Stover Mordecai Associate Professor of Biomedical Engineering at the Pratt School of Engineering.