arXiv subjects

Lyes Kadem

Publications and source records attributed to Lyes Kadem.

14 recordsLinked to original sources

arXiv2026

Hemodynamic Performance and Blood Damage of the Intra-Aortic Pumps: A CFD-Based Investigation

Three intra-aortic blood pumps were evaluated and compared using CFD simulations. The aim of this study is to evaluate the hemodynamic performance and calculate the hemolytic potential of the pumps. The flow fields generated by the pumps were simulated using CFD with a wall-modeled large eddy simulation (WMLES) approach. The simulations produced pressure-flow curves, hydraulic efficiency, shear stress distributions, and hemolysis predictions. A grid study was conducted using the Grid Convergence Index (GCI) method, and a new dimensionless parameter, the Hemolytic Number (HN), was introduced as a standardized metric to compare hemolysis and universal pump performance. The impeller-driven pump had the highest-pressure head (~800 Pa at 4 L/min) and the best hydraulic efficiency (~6% at 14 L/min), outperforming both the single (maximum 2.7%) and triplet (maximum 2.2%) pumps. The NIH values were also lowest for the impeller pump (NIH = 0.0035 g/100L), indicating high hemocompatibility. Both the single and triplet pumps showed regions of recirculation, particularly at lower flow rates. A smaller HN indicates better hemocompatibility; for the impeller-driven pump, HN remains below 1 across the investigated flow rates. Overall, the impeller-driven pump outperformed the other designs in terms of both hemodynamic and hemolytic measurements. The findings provide valuable insights for the future development of intra-aortic pumps and the personalized selection of devices for individual patients.

physics.med-ph↗

arXiv2025

End-to-End Framework Integrating Generative AI and Deep Reinforcement Learning for Autonomous Ultrasound Scanning

Cardiac ultrasound (US) is among the most widely used diagnostic tools in cardiology for assessing heart health, but its effectiveness is limited by operator dependence, time constraints, and human error. The shortage of trained professionals, especially in remote areas, further restricts access. These issues underscore the need for automated solutions that can ensure consistent, and accessible cardiac imaging regardless of operator skill or location. Recent progress in artificial intelligence (AI), especially in deep reinforcement learning (DRL), has gained attention for enabling autonomous decision-making. However, existing DRL-based approaches to cardiac US scanning lack reproducibility, rely on proprietary data, and use simplified models. Motivated by these gaps, we present the first end-to-end framework that integrates generative AI and DRL to enable autonomous and reproducible cardiac US scanning. The framework comprises two components: (i) a conditional generative simulator combining Generative Adversarial Networks (GANs) with Variational Autoencoders (VAEs), that models the cardiac US environment producing realistic action-conditioned images; and (ii) a DRL module that leverages this simulator to learn autonomous, accurate scanning policies. The proposed framework delivers AI-driven guidance through expert-validated models that classify image type and assess quality, supports conditional generation of realistic US images, and establishes a reproducible foundation extendable to other organs. To ensure reproducibility, a publicly available dataset of real cardiac US scans is released. The solution is validated through several experiments. The VAE-GAN is benchmarked against existing GAN variants, with performance assessed using qualitative and quantitative approaches, while the DRL-based scanning system is evaluated under varying configurations to demonstrate effectiveness.

cs.CV↗

arXiv2025

Comprehensive Review of Reinforcement Learning for Medical Ultrasound Imaging

Medical Ultrasound (US) imaging has seen increasing demands over the past years, becoming one of the most preferred imaging modalities in clinical practice due to its affordability, portability, and real-time capabilities. However, it faces several challenges that limit its applicability, such as operator dependency, variability in interpretation, and limited resolution, which are amplified by the low availability of trained experts. This calls for the need of autonomous systems that are capable of reducing the dependency on humans for increased efficiency and throughput. Reinforcement Learning (RL) comes as a rapidly advancing field under Artificial Intelligence (AI) that allows the development of autonomous and intelligent agents that are capable of executing complex tasks through rewarded interactions with their environments. Existing surveys on advancements in the US scanning domain predominantly focus on partially autonomous solutions leveraging AI for scanning guidance, organ identification, plane recognition, and diagnosis. However, none of these surveys explore the intersection between the stages of the US process and the recent advancements in RL solutions. To bridge this gap, this review proposes a comprehensive taxonomy that integrates the stages of the US process with the RL development pipeline. This taxonomy not only highlights recent RL advancements in the US domain but also identifies unresolved challenges crucial for achieving fully autonomous US systems. This work aims to offer a thorough review of current research efforts, highlighting the potential of RL in building autonomous US solutions while identifying limitations and opportunities for further advancements in this field.

eess.IV↗

arXiv2025

CACTUS: An Open Dataset and Framework for Automated Cardiac Assessment and Classification of Ultrasound Images Using Deep Transfer Learning

Cardiac ultrasound (US) scanning is a commonly used techniques in cardiology to diagnose the health of the heart and its proper functioning. Therefore, it is necessary to consider ways to automate these tasks and assist medical professionals in classifying and assessing cardiac US images. Machine learning (ML) techniques are regarded as a prominent solution due to their success in numerous applications aimed at enhancing the medical field, including addressing the shortage of echography technicians. However, the limited availability of medical data presents a significant barrier to applying ML in cardiology, particularly regarding US images of the heart. This paper addresses this challenge by introducing the first open graded dataset for Cardiac Assessment and ClassificaTion of UltraSound (CACTUS), which is available online. This dataset contains images obtained from scanning a CAE Blue Phantom and representing various heart views and different quality levels, exceeding the conventional cardiac views typically found in the literature. Additionally, the paper introduces a Deep Learning (DL) framework consisting of two main components. The first component classifies cardiac US images based on the heart view using a Convolutional Neural Network (CNN). The second component uses Transfer Learning (TL) to fine-tune the knowledge from the first component and create a model for grading and assessing cardiac images. The framework demonstrates high performance in both classification and grading, achieving up to 99.43% accuracy and as low as 0.3067 error, respectively. To showcase its robustness, the framework is further fine-tuned using new images representing additional cardiac views and compared to several other state-of-the-art architectures. The framework's outcomes and performance in handling real-time scans were also assessed using a questionnaire answered by cardiac experts.

cs.CV↗

arXiv2025

Numerical investigation of the flow induced by a transcatheter intra-aortic entrainment pump

This study evaluates the fluid dynamics inside and outside transcatheter blood pump positioned in the aorta. We focus on the pump's impact on blood component damage and arterial wall stress. CFD simulations were performed for rotational speeds ranging from 6000 to 15000 rpm, with a blood flow rate of 1.6 L/min. Results show that significant blood damage may occur at speeds as low as 12000 rpm, and the pump's outflow jet induces elevated wall shear stress, potentially leading to arterial aneurysms. These findings suggest the need for further design improvements to reduce risks when used in prolonged or transplant-related applications.

physics.flu-dyn↗

arXiv2024

Three-Dimensional Lagrangian Coherent Structures in Patients with Aortic Regurgitation

Understanding blood transport in cardiovascular flows is important for managing patients with cardiovascular disease. In this study, three-dimensional Lagrangian coherent structures have been extracted for the first time in both healthy patients and patients with aortic regurgitation. To achieve this, a computationally efficient approach based on Lagrangian descriptors was employed with four dimensional (4D) magnetic resonance imaging velocity fields. In healthy subjects, Lagrangian coherent structures (LCSs) analysis revealed well-defined mitral jet structures during early filling, directing flow toward ejection during systole. For patients with aortic regurgitation, complex flow structures included interactions between the mitral and regurgitant jets, indicating altered blood transport mechanisms. This study highlights the ability of Lagrnagian descriptors to extract coherent structures from patient-specific 4D flow MRI data in a computationally efficient way. It also underscores the importance of extracting three-dimensional Lagrangian coherent structures to gain a better understanding of the complex interaction between the mitral inflow and the regurgitant jet.

physics.med-ph↗

arXiv2023

Evaluating Uncertainties in CFD Simulations of Patient-Specific Aorta Models using Grid Convergence Index Method

Cardiovascular diseases are among the most important causes of global mortality. Computational Fluid Dynamics (CFD) is a powerful research tool that analyzes the hemodynamics of artery and blood flow patterns. In this study, CFD simulations are performed to assess the patient-specific healthy aorta, fusiform, and saccular aneurysm with various mesh types, including tetrahedral, polyhedral, and poly-hexacore. The aim of this study is to explore how different mesh types and grid densities impact the hemodynamic properties of physiological flows, with the goal of identifying the most cost-effective meshing approach. A mesh independence study is carried out to ensure the precision of the results, considering the wall shear stress distribution. For this, five different mesh resolutions are generated for each geometry. The uncertainties of the simulations associated with the discretization techniques and solutions are evaluated using the Grid Convergence Index (GCI) method. The findings showed that increasing the mesh density provides smaller uncertainty. GCI values for the wall shear stress are in the range of convergence, indicating that the results are reliable and accurate. Mesh type selection affects the accuracy and computational cost of our simulations. The polyhedral and poly-hexacore meshes lead to a good compromise between precision and computational cost, while the tetrahedral mesh style gives the most precise results with fluctuation. This work provides a systematic approach based on the Grid Convergence Index method in order to select the most appropriate mesh type for evaluating uncertainties in CFD simulations of patient-specific healthy aortas and aortas with abdominal aneurysms. According to the findings and GCI analysis, the polyhedral mesh type was chosen for all patient-specific aorta models. The study clearly demonstrated its superiority over other mesh types, ...

physics.flu-dyn↗

arXiv2021

Experimental Investigation of Aortic Pressure Variations Following a Simulated Thoracic Impact

Blunt traumatic aortic rupture is a heart injury that can occur in falls, automobile accidents, and sporting injuries involving impact to the thorax. Despite its severity and high morbidity rate, the research still does not provide a consistent description of the mechanism of rupture. In this study, a crash testing dummy with an in vitro pumping heart, 3D printed ribcage, and ballistic gel damping layer was developed to reproduce a realistic response to thoracic impact. Testing was performed using a standardized pendulum used for calibration of crash test dummies, with the location of impact being the middle of the sternum. Different impact severities were tested by adjusting the kinetic energy at impact with the initial height of the pendulum. Measurements of the dummy include instantaneous aortic pressure waveforms prior, during and following the impact. The results of this experiment show that aortic pressure experiences significant changes in magnitude during simulated impact. This work could help contribute towards a better understanding of the mechanisms leading to blunt traumatic aortic rupture and the development of preventative measures.

physics.med-ph↗

arXiv2019

Experimental investigation of the flow downstream of a dysfunctional bileaflet mechanical aortic valve

Mechanical heart valve replacement is the preferred alternative in younger patients with severe symptomatic aortic valve disease. However, thrombus and pannus formations are common complications associated with bileaflet mechanical heart valves. This leads to risks of valve leaflet dysfunction, a life threatening event. In this experimental study, we investigate, using time resolved planar particle image velocimetry, the flow characteristics in the ascending aorta in the presence of a dysfunctional bileaflet mechanical heart valve. Several configurations of leaflet dysfunction are investigated and the induced flow disturbances in terms of velocity fields, viscous energy dissipation, wall shear stress, and accumulation of viscous shear stresses are evaluated. We also explore the ability of a new set of parameters, solely based on the analysis of the normalized axial velocity profiles in the ascending aorta, to detect bileaflet mechanical heart valve dysfunction and differentiate between the different configurations tested in this study. Our results show that a bileaflet mechanical heart valve dysfunction leads to a complex spectrum of flow disturbances with each flow characteristic evaluated having its own worst case scenario in terms of dysfunction configuration. We also show that the suggested approach based on the analysis of the normalized axial velocity profiles in the ascending aorta has the potential to clearly discriminate not only between normal and dysfunctional bilealfet heart valves but also between the different leaflet dysfunction configurations. This approach could be easily implemented using phase contrast MRI to follow up patients with bileaflet mechanical heart valves.

physics.med-ph↗

arXiv2019

Reduced-order modeling of left ventricular flow subject to aortic valve regurgitation

The present focus of heart flow studies is largely based on flow within the left ventricle and how this flow changes when subject to disease. However, despite recent advancements, a simple tractable model of even healthy left ventricular flow has not been produced and made available. Reduced-order modeling techniques, such as proper orthogonal decomposition (POD) and dynamic mode decomposition (DMD), offer an effective means of expressing the large datasets obtained from experiments or numerical simulations using low-dimensional models. While POD and DMD are often used to identify coherent structures in fluid dynamics, their use as a modeling tool has not found much merit in the cardiovascular flow community. In this work, we use POD and DMD to construct reduced-order models for a healthy left ventricular flow as well as for that under the influence of a particular disease shown to exhibit rich and unique intraventricular fluid dynamics, namely, aortic regurgitation (a leaking aortic valve). The performance of the two methods in reconstructing the intraventricular flows and derived quantities is evaluated, and the selected reduced-order models are made available.

physics.flu-dyn↗

arXiv2018

Jet collisions and vortex reversal in the human left ventricle

Unnatural dynamics of the notorious vortex in the left ventricle is often associated with cardiac disease. Understanding how different cardiac diseases alter the flow physics in the left ventricle may therefore provide a powerful tool for disease detection. In this work, the fluid dynamics in the left ventricle subject to different severities of aortic regurgitation is experimentally investigated by performing time-resolved particle image velocimetry in a left heart duplicator. Diastolic vortex reversal was observed in the left ventricle accompanied by an increase in viscous energy dissipation. Vortex dynamics and energy dissipation may provide useful insights on sub-optimal flow patterns in the left ventricle.

physics.med-ph↗

arXiv2018

Numerical simulation of flows in a circular pipe transversely subjected to a localized impulsive body force with applications to blunt traumatic aortic rupture

Much debate surrounds the mechanisms responsible for the occurrence of blunt traumatic aortic rupture in car accidents, particularly on the role of the inertial body force experienced by the blood due to the abrupt deceleration. The isolated influence of such body forces acting on even simple fluid flows is a fundamental problem in fluid dynamics that has not been thoroughly investigated. This study numerically investigates the fundamental physical problem, where the pulsatile flow in a straight circular pipe is subjected to a transverse body force on a localized volume of fluid. The body force is applied as a brief rectangular impulse in three distinct cases, namely during the accelerating, peak, and decelerating phases of the pulsatile flow. A dimensionless number, termed the degree of influence of the body force (Ψ), is devised to quantify the relative strength of the body force over the flow inertia. The impact induces counter-rotating cross-stream vortices at the boundaries of the forced section accompanied by complex secondary flow structures. This secondary flow is found to develop slowest for an impact occurring during an accelerating flow and fastest during a decelerating flow. The peak skewness of the velocity field, however, occurred at successively later times for the three respective cases. After the impact, these secondary flows act to restore the unforced state and such dominant spatial structures are revealed by proper orthogonal decomposition of the velocity field. This work presents a new class of problems that requires further theoretical and experimental investigation.

physics.flu-dyn↗

arXiv2018

Material transport in the left ventricle with aortic valve regurgitation

This experimental in vitro work investigates material transport properties in a model left ventricle in the case of aortic regurgitation, a valvular disease characterized by a leaking aortic valve and consequently double-jet filling within the elastic left ventricular geometry. This study suggests that material transport phenomena are strongly determined by the motion of the counterrotating vortices driven by the regurgitant aortic and mitral jets. The overall particle residence time appears to be significantly longer with moderate aortic regurgitation, attributed to the dynamics resulting from the timing between the onset of the two jets. Increasing regurgitation severity is shown to be associated with higher frequencies in the time-frequency spectra of the velocity signals at various points in the flow, suggesting nonlaminar mixing past moderate regurgitation. Additionally, a large part of the regurgitant inflow is retained for at least one cardiac cycle. Such an increase in particle residence time accompanied by the occurrence and persistence of a number of attracting Lagrangian coherent structures presents favorable conditions and locations for activated platelets to agglomerate within the left ventricle, potentially posing an additional risk factor for patients with aortic regurgitation.

physics.flu-dyn↗

arXiv2009

Mixing Within Patterned Vortex Core

The video shows the flow dynamics within inner and outer regions of a vortex core. The observed phenomena mimic a transport process occurring within the Antarctic vortex. The video shows two distinct regions: a strongly mixed core and broad ring of weakly mixed region extending out the vortex core boundaries. The two regions are separated by a thin layer that isolates the weakly and strongly mixed regions; this thin layer behaves as barrier to the mixing of the two regions. The video shows that the barriers deplete when a swirl of the vortex core increases and the vortex core espouses a triangular pattern.

physics.flu-dyn↗