← Latest papers
🏥 surgery

Learned Human Aortic Morphodynamics: A Dynamical-Systems Model of Post-EVAR Sac Remodeling

This study demonstrates that sparse, noisy clinical imaging data from post-EVAR patients can be used with the Z-SINDy method to infer distinct dynamical-system models of aortic sac remodeling, revealing that successful and failed outcomes follow different morphological trajectories and timescales, thereby enabling earlier and more robust clinical predictions than static measurements alone.

Original authors: Pugar, J., Kim, J., Mansour, M., Davis, C., Nguyen, N., Lee, C. J., Babrowski, T., Verhagen, H., Milner, R., Klishin, A., Pocivavsek, L.

Published 2026-09-18 (v3)
📖 5 min read🧠 Deep dive

Original authors: Pugar, J., Kim, J., Mansour, M., Davis, C., Nguyen, N., Lee, C. J., Babrowski, T., Verhagen, H., Milner, R., Klishin, A., Pocivavsek, L.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

The human body is not a static machine; it is a living system that constantly reshapes itself in response to its environment. When a blood vessel is stressed or damaged, the tissue does not simply sit there; it reorganizes, thickens, or shrinks in an attempt to find a new balance. This process, known as remodeling, is governed by physical laws, but observing those laws in a living human is incredibly difficult. Unlike a laboratory experiment where scientists can take measurements every second, doctors rely on medical scans taken only once every year or two. These snapshots are few, far between, and often noisy, making it nearly impossible to see the smooth, continuous story of how an organ changes over time. For decades, this lack of continuous data has kept researchers from writing down the actual rules that dictate how organs heal or fail to heal after surgery.

A team of researchers at the University of Chicago and other institutions set out to solve this puzzle by looking at the abdominal aorta, the main artery carrying blood to the lower body. When this artery weakens and bulges into an aneurysm, surgeons often perform a procedure called endovascular aneurysm repair, or EVAR. They insert a stent graft inside the vessel to seal off the bulge, effectively taking the pressure off the weakened wall. The goal is for the aneurysm sac to shrink and heal, a process called regression. However, sometimes the sac remains stable and does not shrink, which can lead to worse long-term outcomes. Doctors currently watch these patients by measuring the size of the sac on CT scans, but a single measurement cannot tell them if the sac is actively healing or just sitting still. The researchers wanted to know if they could discover the hidden mathematical rules of this healing process using only the sparse, irregular scans available in real-world clinics.

To do this, the team gathered data from 100 patients across two medical centers in the United States and the Netherlands, totaling 220 CT scans. Instead of just measuring the diameter of the artery, they reduced the complex 3D shape of the aneurysm into two simple numbers: its overall size and a measure of its surface shape, specifically how much the surface curves and twists. They normalized these numbers against healthy aortas, creating a simple map where every patient's condition could be plotted as a single point. By connecting the dots from each patient's scans over time, they could see the path, or trajectory, that each aneurysm took as it changed.

The researchers then used a sophisticated computer method to find the simplest possible set of rules that could explain how these points moved across the map. They discovered that the healing process follows a predictable pattern, much like a ball rolling down a hill toward a resting spot. For patients whose aneurysms were healing successfully, the rules showed a clear path toward a specific destination where the shape of the artery had returned to near-normal, even though the size remained slightly larger than a healthy artery. This healing process happened relatively quickly, with the shape correcting itself in about 1.3 years and the size settling over roughly 4.6 years. In contrast, patients whose aneurysms did not heal were governed by a different set of rules. Their path led to a completely different destination, one where both the size and shape remained enlarged, and the journey toward this resting point was agonizingly slow, taking over a decade to complete.

The most striking finding was that the difference between a healing and a non-healing aneurysm is not just about where the sac is, but how it is moving. A single snapshot of an aneurysm cannot tell a doctor if it is healing, because a healing sac and a stable sac can look identical at any given moment. However, by looking at the direction and speed of the change—the trajectory—the researchers found they could predict the outcome much earlier. When they built a computer model to simulate these paths, they discovered that a method tracking the motion of the aneurysm could correctly classify a patient's outcome in less than a year. A method that only looked at the position, or the current size and shape, failed to make a confident prediction even after five years. This suggests that the way an organ moves through its healing process carries more information than its current state.

The study also addressed a major problem in medical data: patients often drop out of follow-up or miss scans. The researchers showed that even with this missing data, their model could still learn the underlying rules of healing. By using the few scans they had to infer the smooth path between them, they could generate a "digital twin" for each patient—a virtual simulation of how that specific aneurysm would likely behave in the future. These simulations were not just guesses; they came with a built-in measure of uncertainty, showing a range of possible futures rather than a single prediction. This approach allows doctors to see not just where a patient is now, but where they are likely to go, providing a foundation for better clinical decisions based on the sparse records that routine imaging actually produces.

Ultimately, this work demonstrates that even with limited and messy data, it is possible to uncover the fundamental laws governing how human organs remodel. The researchers did not just describe what happens; they recovered the equations that drive it. They found that successful healing is a fast, directed journey toward a specific shape, while failure is a slow drift toward a different, enlarged state. By focusing on the motion rather than the static position, the study offers a new way to understand organ recovery, turning a series of disconnected medical images into a continuous, understandable story of healing. This approach could eventually be applied to other organs and diseases, helping doctors predict outcomes and tailor treatments based on the unique dynamical path of each patient's body.

Drowning in papers in your field?

Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.

Try Digest →