The third MeDiTATe project newsletter is out! It contains main news, latest events and much more of the activities carried out in the second trimester of the 2023.
The newsletter can be downloaded at the following link: MeDiTATe Newsletter #3 July 2023.
The third MeDiTATe project newsletter is out! It contains main news, latest events and much more of the activities carried out in the second trimester of the 2023.
The newsletter can be downloaded at the following link: MeDiTATe Newsletter #3 July 2023.
A new paper “Prediction of guidewire-induced aortic deformations during EVAR: a finite element and in vitro study” has been published in the Frontiers in Physiology Journal. The paper was written by Monica Emendi ESR 11 of the MeDiTATe project.
The abstract is as follows:
Pre-operative planning and intra-operative navigation of endovascular aneurysm repair (EVAR) are currently challenged by guidewire-induced deformations. In this work we aim to predict the above-mentioned deformations by means of a validated finite element approach. Hence, three physical patient-specific models of increasing anatomical complexity, i.e., including the intraluminal thrombus (ILT) and calcifications, were manufactured. The numerical and experimental results were compared. A good agreement was found between the experimental and the computational studies, characterized by an increasing trend in deformations from a simplified model to a complete model (with ILT and calcifications). The presented computational approach, used in combination with tracking technology, can potentially display a continuously updated 3D anatomical map to the operator during EVAR, limiting the use of X-ray and contrast agent.
The full paper is available at this link.
The MeDiTATe project was present at the 28th Congress of the European Society of Biomechanics (ESBiomech23). The event took place in Maastricht, The Netherlands from 9 to 12 July 2023.
In this occasion six of the Early-Stage Researchers (ESRs) presented the results of their activities with both oral and poster presentations: Leonardo Geronzi – ESR 02, Beatrice Bisighini – ESR 03, Federica Galbiati – ESR 07, Rahul Vellaparambil – ESR08, Martino Andrea Scarpolini – ESR 09, Francesco Bardi – ESR 10.
A new paper “Computer-aided shape features extraction and regression models for predicting the ascending aortic aneurysm growth rate” has been published in the Computers in Biology and Medicine Journal. The paper was written by Leonardo Geronzi (ESR 02) as well as by other including Marco Evangelos Biancolini, Principal Investigator of the MeDiTATe project.
The abstract is as follows:
Ascending aortic aneurysm growth prediction is still challenging in clinics. In this work, a method to exploit local and global shape features for the prediction of the ascending aortic aneurysm growth rate is presented. By using a framework based on geometric decomposition, mesh morphing, statistical shape analysis and regression methods, we were able to extract local and global shape features potentially valuable for improving the prediction of the aneurysm growth. The results of this study show that aneurysms close to the root with a large initial diameter report faster growth.
The full paper is available at this link.
A new paper “Patient-specific computational modelling of endovascular treatment for intracranial aneurysms” has been published in the Brain Multiphysics Journal. The paper was written by Beatrice Bisighini (ESR 03) as well as by other authors ìincluding Stéphane Avril, Research Coordinator of MeDiTATe project and PhD Supervisor of Beatrice Bisighini.
Here some highlights from the paper:
The full paper is available at this link.
The MeDiTATe project will be present at the 28th Congress of the European Society of Biomechanics (ESBiomech23). The event will be held in Maastricht, The Netherlands July 9 -12, 2023.
In this occasion six of the Early-Stage Researchers (ESRs) will be presenting the results of their activities with both oral and poster presentations:
The evaluation of a pathological ascending aorta is challenging due to its morphological variabilities, which cannot be adequately assessed by measuring the diameter alone. Additional measurements derived from a 3D model are necessary to accurately evaluate the condition. Computational biomarkers have already demonstrated their significance in diagnosing cardiovascular diseases. In this study, we used longitudinal data to observe the progression of the aneurysm over time. By decomposing the geometry of the ascending tract of the aorta in 3D Slicer, we extracted a set of local and global shape features. Machine learning classification and regression models are then employed and compared to predict the risk associated with the disease.
The approach of using in-silico models based on computer-aided engineering (CAE) to study diseases, suggest treatment strategies and predict surgical outcomes proved to be crucial in supporting the clinical staff. However, the high numerical complexity and computational cost still represent a challenge with respect to the exploitation of such results. To overcome these limitations, new methods based on medical Digital Twins (DTs) are under development. DTs are virtual replicas of physical systems able to digitally replicate their behaviors providing a connection between the physical entities and the corresponding digital models. In this work, we describe the techniques we use to build medical DTs and their application to two clinical cases: the drug-delivery simulation of the airway system and the determination of the effect of an exertion activity on the ascending aortic aneurysm.
This work presents a machine learning-based reduced-order model scheme, trained on finite element simulations, to compute the deployed configuration of flow diverters within patient-specific models in real-time. Flow diverters are very dense braided stents used in the endovascular treatment of cerebral aneurysms. Computational tools could be useful in assisting surgeons in the selection of the best device for a patient, especially in complex cases. However, due to the large amount of degrees of freedom and the necessity to solve the contact with the wall, the computational time required by traditional techniques alone, such as finite element modelling, is excessively high. Machine learning-based reduced order modelling can enable real-time prediction while retaining the mechanical realism and predictability of the stent deployed configuration.
The presence of a bicuspid aortic valve (BAV), with consequent eccentricity of aortic inlet flow, gives rise to altered aortic hemodynamics. This seems to play a role in the highest prevalence of ascending thoracic aortic aneurysm (aTAA) in BAV patients, compared to subjects with tricuspid aortic valve (TAV). Characteristics of aortic hemodynamics in presence of eccentric inlet flow in both healthy and aneurysmatic aorta have been widely addressed, while perspective studies that investigate the evolution of hemodynamic patterns concurrently with aTAA progression are still limited. Therefore, the main goal of this study is to assess the role of both TAV and eccentric BAV inflow on hemodynamic results at different stages of aTAA, coupled with the development of a parametrized time-space-varying inflow model to reproduce patient-specific TAV / BAV aortic inlet condition.
Computational fluid dynamics assessment of patient-specific hemodynamics using full order models is a viable tool to gain insights for pathology evaluation, treatment, and prevention. However, its use in clinical practice is hindered by its high computational cost in terms of infrastructures and timing. Moreover, these limitations are reflected in the difficulties to be directly translated to clinicians. In this context, nonintrusive data-driven Reduced Order Models (ROMs) represent a promising tool for facing these limitations, allowing high-fidelity and fast hemodynamic evaluation in a user-friendly setup. In this study, a workflow to create a ROM for the evaluation of aortic hemodynamics is presented, with the specific aim to investigate the effect of inlet conditions with a focus on the effects of an eccentric inflow caused by aortic valve morphology and pathologies.
Angulated iliac aneurysms are known to influence kinks and stent-migration on Stent-grafts (SGs) used in endovascular aortic repair (EVAR) after deployment. A recent work from our group has demonstrated the potential of SGs derived from auxetic unit cells in withstanding high bending angulations of 180°, which have been noted in highly tortuous iliac aneurysms. In this study, we intend to investigate the deployment of SGs derived from auxetic unit cells in an idealized curved iliac aneurysm using finite element analysis and quantitatively evaluate their mechanical performance using luminal reduction and apposition to aortic wall. All SGs derived from auxetic unit cells showed a maximum of 32% reduction in stent-graft cross sectional area, with SGs comprising of re-entrant (RE) and chiral-re-entrant (CRE) unit cells demonstrating no mal-apposition to the aortic wall. Thus, SGs derived from auxetic unit cells display promising results for EVAR applications despite increased tortuosity.
The integration of in silico and in vivo data is crucial to the development of high-fidelity digital twins of the cardiovascular (CV) system, but is a challenging task that requires specific imaging techniques and in silico setups. Advancements in imaging technologies are making it possible to gather a large amount of patient information. In parallel, in silico models are becoming a useful tool to simulate patient-specific conditions, treatments and therapies. However, the latter models require a large amount of physical parameters to be known, which are often very difficult to measure in vivo. In this study, we used data-assimilation techniques to merge high-resolution temporal CT scans with fluid structure interaction (FSI) simulations, resulting in the creation of a high-fidelity digital twin of the left ventricle (LV) and aorta system of a patient.
Cardiovascular diseases (CVDs) are a leading cause of death in Europe, accounting for 45% of all deaths. The adoption of computational power and numerical tools has advanced the development of digital twins for CVDs, but their high computational cost and time have limited their implementation in clinical settings. In this work, we propose a workflow that combines deep learning, reduced order modelling (ROM) and computational fluid dynamics (CFD) to automatically build a digital twin of the patient’s thoracic aorta.
The development and the progression of abdominal aortic aneurysms (AAA) are greatly influenced by the blood flow behavior. Important hemodynamic variables, such as the wall shear stress and the oscillatory index can be used to better understand this phenomenon. However, current imaging approaches cannot accurately evaluate these indicators in-vivo, numerical simulations are used instead. In order to trust the results of numerical simulation it is fundamental to properly validate them in-vitro. In this work a cost-effective LED illuminate PIV setup and a novel Hybrid Mock Circulatory Loop were used to validate in-vitro the results of patient specific numerical simulation. Despite a slight underestimation of the velocity magnitude, the main complex flow features occurring in the aneurysmatic sac were correctly predicted. These results confirms that FSI simulations are a reliable tool to study the fluid dynamic of AAAs.
