📄 medicine

A 4D Ultrasound Strain Imaging Framework for In Vivo Identification of Anisotropic Properties in Abdominal Aortic Aneurysms: Preliminary Validation against Mechanical Testing of Porcine and Human Tissue

This paper presents and preliminarily validates a computationally efficient, non-invasive 4D ultrasound framework that identifies patient-specific anisotropic mechanical properties of abdominal aortic aneurysm walls by solving an inverse finite element problem, demonstrating strong agreement with mechanical testing in both porcine and human tissues.

Original authors: Achim Hegner, Andreas Wittek, Wojciech Derwich, Manuel Schönborn, Kyriakos Oikonomou, Armin Huß, Antonio J. Gámez, Christopher Blase

Published 2026-10-19

📖 3 min read☕ Coffee break read

CC BY 4.0

Original authors: Achim Hegner, Andreas Wittek, Wojciech Derwich, Manuel Schönborn, Kyriakos Oikonomou, Armin Huß, Antonio J. Gámez, Christopher Blase

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The human body is a complex machine, and the abdominal aorta, the main artery carrying blood to the lower half of the body, is one of its most critical components. Sometimes, the wall of this artery weakens and bulges outward, forming a condition known as an abdominal aortic aneurysm. This bulge is dangerous because it can burst, leading to a life-threatening emergency. Currently, doctors decide whether to repair an aneurysm based largely on its size, measuring how wide the bulge has become. However, size alone does not tell the whole story; two aneurysms of the same size can have very different risks of bursting because the material strength of the artery wall varies from person to person. To predict a rupture accurately, doctors need to understand the specific mechanical properties of the patient's own artery wall—how stiff it is, how it stretches, and how it handles pressure. The challenge is that these properties are difficult to measure without invasive surgery, and the artery is often covered by a thick, jelly-like clot called an intraluminal thrombus, which obscures the wall and makes standard measurements unreliable.

A team of researchers has developed a new, non-invasive method to solve this problem by using advanced ultrasound imaging to "feel" the stiffness of the artery wall from the outside. Instead of relying on a single measurement of size, their approach uses a four-dimensional ultrasound scan to capture how the artery wall moves and stretches during a heartbeat. By tracking these tiny movements, they can work backward to calculate the unique material properties of the artery wall for that specific patient. This process involves creating a detailed computer model of the patient's artery, including the clot, and then adjusting the model's material settings until the simulated movement matches the movement seen in the ultrasound scan. The researchers tested this new framework first on a pig aorta in a controlled laboratory setting, where they could compare their ultrasound results against direct mechanical tests. They found that the ultrasound method could predict the tissue's behavior with remarkable accuracy, matching the results of physical stretching tests almost perfectly.

The team then moved to human patients, applying their method to two individuals undergoing surgery for aneurysms. Before the surgery, they scanned the patients' arteries with the ultrasound device. After the surgery, they took small samples of the actual artery wall and tested them in a lab to see how they reacted to stretching. When they compared the properties predicted by the ultrasound scan with the actual properties measured from the tissue samples, the results were encouraging. For one patient, the match was very good; for the other, it was moderate. The researchers discovered that the accuracy of their predictions depended heavily on two factors: knowing the exact thickness of the artery wall and ensuring the ultrasound captured enough of the artery's natural stiffening behavior during the heartbeat. While the study involved only a small number of patients and is considered a preliminary proof of concept, it demonstrates that it is possible to map the mechanical strength of a living human aneurysm without cutting into the body. This capability represents a significant step toward personalized medicine, where treatment decisions could be based on the actual strength of a patient's artery rather than just its size, potentially saving lives by identifying high-risk aneurysms that might otherwise go unnoticed.

Technical Summary: A 4D Ultrasound Strain Imaging Framework for In Vivo Identification of Anisotropic Properties in Abdominal Aortic Aneurysms

Problem Statement
Abdominal aortic aneurysms (AAAs) pose a significant risk of rupture, yet current clinical risk stratification relies heavily on population-based diameter thresholds, which may lead to overtreatment or missed interventions. Accurate patient-specific biomechanical assessment requires knowledge of the wall's material properties, specifically its hyperelastic and anisotropic behavior. However, these properties are typically unknown in vivo. While inverse finite element analysis (FEA) offers a path to identify these parameters from clinical imaging, existing methods face critical limitations: they are often computationally expensive, exclude the intraluminal thrombus (ILT)—a feature present in ~75% of AAAs—or lack rigorous experimental validation against mechanical testing of harvested tissue. Furthermore, validating in vivo identification methods against ex vivo mechanical tests on the same patient's tissue remains a scarce and underexplored area.

Methodology
The authors present a computationally efficient framework for identifying patient-specific, anisotropic material properties of AAA walls using non-invasive 4D ultrasound (4D-US) strain imaging. The core methodology involves:

Data Acquisition & Processing: 4D-US data is acquired using a real-time 3D echocardiography system. The luminal border is segmented, and speckle tracking is performed to generate a field of material points representing wall displacements over the cardiac cycle. Local in-plane Biot strain tensors are computed from these displacements relative to the end-diastolic configuration.
Constitutive Modeling: The vessel wall is modeled using the anisotropic, hyperelastic Gasser-Ogden-Holzapfel (GOH) material model, defined by five parameters ( $\mu, k_1, k_2, \kappa, \phi$ ). The ILT is modeled as a single-layer, isotropic, nearly incompressible neo-Hookean material, with its Poisson's ratio estimated individually from volume changes.
Inverse Optimization: The identification is formulated as an inverse FEA problem. A patient-specific FE model is constructed from the US-measured geometry. An optimization algorithm (self-adaptive JADE, a Differential Evolution variant) iteratively adjusts the GOH parameters to minimize the squared error between the US-measured cyclic strains and the strains computed by the FE model under physiological blood pressure loading (diastole and systole).
Validation Strategy: The framework undergoes a two-part preliminary validation:
- In Vitro: A porcine aorta is subjected to controlled inflation-extension tests. Material parameters are identified independently from (a) optical camera data/mechanical force measurements and (b) 4D-US data. These two sets of parameters are then compared.
- In Vivo: Two AAA patients undergo pre-operative 4D-US imaging. Following surgical repair, excised wall tissue samples ( $n=6$ ) are subjected to uniaxial tensile tests. Parameters identified from the in vivo US data are compared against those derived from the ex vivo mechanical tests of the same patients' tissues.
Sensitivity Analysis: The robustness of the identification is assessed by varying wall thickness, axial force, and the deformation range (diastole-to-systole) to determine their influence on the resulting parameters.

Key Contributions

Integrated ILT Modeling: Unlike previous 4D-US approaches, this framework explicitly incorporates the ILT geometry and mechanics, making it applicable to the majority of clinical AAA cases.
Computational Efficiency: The method reduces computation time by approximately 86% (7-fold) compared to previous iterations, enabling application to a broader spectrum of clinical cases.
Experimental Validation: The study provides a rare direct comparison of in vivo parameter identification against ex vivo mechanical testing on harvested tissue from the same patients, addressing a critical gap in the literature.
Robustness Analysis: The study identifies wall thickness and the extent of the nonlinear stiffening range captured during the cardiac cycle as the most influential factors on identification accuracy.

Results

In Vitro (Porcine Aorta): The stress-stretch responses derived from 4D-US and mechanical testing agreed closely, with coefficients of determination ( $R^2$ ) ranging from 0.96 to 0.99. The identified GOH parameters from both modalities were nearly identical, confirming the framework's accuracy under controlled conditions.
In Vivo (Human Patients):
- Patient 1: Showed good agreement between in vivo (US) and ex vivo (tensile test) material behaviors ( $R^2$ between 0.319–0.885 across load paths).
- Patient 2: Showed moderate agreement (Mean Absolute Percentage Error ranging from 0.51% to 17.83%).
- The authors note that while the quantitative parameter values differed between methods, the qualitative mechanical response (stress-stretch curves) was well-matched. Discrepancies were attributed to the heterogeneity of the excised samples versus the averaged nature of the US-based identification, as well as the non-uniqueness of GOH parameters when fitted to limited load cases.
Sensitivity: A 10% variation in wall thickness resulted in stress deviations of up to 6.9% in the physiological range, highlighting the need for accurate thickness estimation. Conversely, variations in axial force had a negligible impact. The identification required capturing at least 70% of the healthy pulse pressure range (or ~50% of the nonlinear stiffening region) to achieve sufficient accuracy.

Significance and Claims
The paper positions these results as a "preliminary validation" and a "proof-of-concept." The authors claim that the framework demonstrates the feasibility of non-invasive, patient-specific identification of anisotropic AAA wall properties using only two load cases (diastole/systole).

The significance lies in the ability to incorporate the ILT and reduce computational costs, moving the technology closer to clinical applicability. The authors explicitly state that while the current in vivo cohort is small ( $N=2$ ) and results are modest, the work represents a "major step toward personalized biomechanical modeling and improved clinical risk stratification." They do not claim immediate clinical deployment but rather establish the methodological foundation for future studies with larger cohorts and the assessment of predictive value for rupture risk. The study acknowledges that wall thickness estimation remains a significant challenge for in vivo application and suggests future integration with CT or advanced ultrasound for thickness mapping.

More like this