Cardiovascular Engineering and Technology最新文献

Purpose: Transcatheter aortic valve replacement (TAVR) is the standard treatment for patients with aortic diseases at high surgical risk. Transcatheter heart valve prostheses (THV) are inserted into the aortic valve, creating a new area between the native and artificial leaflets. This area, known as neo-sinus, increases the thrombogenicity of THVs. But there is a lack of testing methods that evaluate thrombogenicity in vitro.

Methods: To analyze the flow field within the native sinus and the neo-sinus, Particle Image Velocimetry (PIV) was performed with a thrombosis tester. Additionally, a comparative study was conducted with porcine blood on two polycarbonate urethane valves, with and without neo-sinus, respectively. Blood samples collected every hour were analyzed for platelet count, coagulation via ROTEM parameters, and plasma-free hemoglobin. Thrombus formation was detected optically.

Results: The PIV measurements yield a physiological flow field in the aortic root that were consistent with those reported in literature. The analyzed blood parameters reveal no obvious difference between the valve with neo-sinus and the valve without. A higher amount of thrombus material for the valve with neo-sinus was found.

Conclusion: The visualized flow field shows low velocities and stagnation zones due to the presence of native leaflets. Clot formation at the heart valve prostheses are in accordance with in-vivo findings. The benchmark of the two valves indicates an increased thrombogenic potential due to the neo-sinus. The thrombosis tester simulates the natural environment after TAVR. Thereby, newly developed THVs can be evaluated in vitro and consequently optimized regarding their thrombogenicity.

Background: The electrocardiogram (ECG) is an almost universally accessible diagnostic tool for heart disease. An ECG is measured by using an electrocardiograph, and today's electrocardiographs use built-in software to interpret the ECGs automatically after they are recorded. However, these algorithms exhibit limited performance, and therefore, clinicians usually have to manually interpret the ECG, regardless of whether an algorithm has interpreted it or not. Manual interpretation of the ECG can be time-consuming and requires specific skills. Therefore, better algorithms are clearly needed to make correct ECG interpretations more accessible and time-efficient. Algorithms based on artificial intelligence (AI) have demonstrated promising performance in various fields, including ECG interpretation, over the past few years and may represent an alternative to manual ECG interpretation by doctors.

Results: We trained and validated a convolutional neural network with an Inception architecture on a dataset with 88253 12-lead ECGs, and classified 30 of the most frequent annotated cardiac conditions in the dataset. We assessed two different loss functions and different ECG sampling rates and the best-performing model used double soft F1-loss and ECGs downsampled to 75Hz. This model achieved an F1-score of $0.420\pm 0.017$ , accuracy $=0.954\pm 0.002$ , and an AUROC score of $0.832\pm 0.019$ . An aggregated saliency map, showing the global importance of all 12 ECG leads for the 30 cardiac conditions, was generated using Local Interpretable Model-Agnostic Explanations (LIME). The global saliency map showed that the Inception model paid the most attention to the limb leads and the augmented leads and less importance to the precordial leads.

Conclusions: One of the more significant contributions that emerge from this study is the use of aggregated saliency maps to obtain global ECG lead importance for different cardiac conditions. In addition, we emphasized the relevance of evaluating different loss functions, and in this specific case, we found double soft F1-loss to be slightly better than binary cross-entropy (BCE). Finally, we found it somewhat surprising that drastic downsampling of the ECG led to higher performance than higher sampling frequencies, such as 500Hz. These findings contribute in several ways to our understanding of the artificial intelligence-based interpretation of ECGs, but further studies should be carried out to validate these findings in other datasets from other patient cohorts.

Purpose: Aortic dissection (AD) is a rare condition with a high mortality rate, necessitating accurate and rapid diagnosis. This study develops an automated deep learning pipeline for identifying, segmenting, and Stanford subtyping AD using computed tomography angiography (CTA) images.

Methods: This pipeline consists of four interconnected modules: aorta segmentation, AD identification, true lumen (TL) and false lumen (FL) segmentation, and Stanford subtyping. In the aorta segmentation module, a 3D full-resolution nnU-Net is trained. The segmented aorta's boundary is extracted using morphological operations and projected from multiple views in the AD identification module. AD identification is then performed using the multi-view projection data. For AD cases, a 3D nnU-Net is further trained for TL/FL segmentation based on the segmented aorta. Finally, a network is trained for Stanford subtyping using multi-view maximum density projections of the segmented TL/FL. A total of 386 CTA scans were collected for training, validation, and testing of the pipeline.

Results: For AD identification, the method achieved an accuracy of 0.979. The TL/FL segmentation for TypeA-AD and Type-B-AD achieved average Dice coefficient of 0.968 for TL and 0.971 for FL. For Stanford subtyping, the multi-view method achieved an accuracy of 0.990.

Conclusion: The automated pipeline enables rapid and accurate identification, segmentation, and Stanford subtyping of AD using CTA images, potentially accelerating the diagnosis and treatment. The segmented aorta and TL/FL can also serve as references for physicians. The code, models, and pipeline are publicly available at https://github.com/zhuangCJ/A-pipeline-of-AD.git .

Purpose: In order to explore the correlation between the initial morphology of the valve and hemodynamic and valve dynamic performance, this study is based on the fact that polymeric prostheses are more convenient to manufacture, and have the possibility of preparing complex geometric shapes and directly obtaining the initial morphologies of different valves, aims to research the effect of different initial opening morphologies of polymeric valves on hemodynamic performance.

Method: Valve models with different opening shapes were established. Polyurethane materials were used to manufacture the valve samples by dip-coating molding. The stress distribution of three different initial opening shapes was compared by finite element simulation. The hemodynamics and the leaflets dynamic performance of the three polymeric valves were analyzed by in vitro pulsatile flow experiments and particle image velocity measurement experiments.

Results: The valve morphology at 0.025s, 0.053s, and 0.079s was selected as the initial shape and was recorded as PHV1, PHV2, and PHV3. Finite element analysis found that during the systolic phase, the stress concentration area of PHV1 was the highest among the three types of valves, while during the diastolic phase, the stress concentration area of PHV1 was the lowest. Similarly, the maximum principal strain of PHV1, PHV2, and PHV3 decreased in turn at the time of peak systole but increased in turn at the time of peak diastole. In vitro testing results showed that valves with smaller opening areas had smaller regurgitant volume, while valves with larger opening areas had larger EOA, as well as smaller vorticity and viscous shear stress.

Conclusion: Valves with a smaller initial opening area have a better effect in preventing regurgitation, whereas valve with a larger initial opening area has a larger opening area and a lower risk of thrombosis. Therefore, comprehensive considerations are needed when designing the initial morphology of the polymeric artificial heart valve.

Purpose: An accurate mathematical model of intracranial aneurysm (IA) mechanics is of great value for its potential utility in assessing probability of IA rupture. Such a model for spherical IAs has been developed which predicts a wall-thickness-to-IA-radius ratio (WTR) of 6.1 × 10^-3 at which IAs rupture. To our knowledge, no further work has been done to validate this model with clinical data. We aim to assess the accuracy and utility of this model of spherical IA rupture mechanics.

Methods: Aneurysm height, width, neck diameter, and vessel radius were measured on radiologic images of IAs of the basilar terminus, anterior communicating, and posterior communicating arteries. Geometric modeling was used to approximate IA wall thickness. Calculations were performed with and without accounting for changes in IA morphology which have been shown to occur post-rupture. Receiver operating characteristic (ROC) curves and positive likelihood ratios (LR) were produced for WTR, aspect ratio (AR), bottleneck factor (BF), and size ratio (SR). Logistic regression analysis was performed to determine the WTR where there is a 50% chance of presentation as a ruptured aneurysm in our cohort.

Results: 52 unruptured and 28 ruptured spherical IAs were included. ROC curve analysis revealed similar areas under the curve for WTR, AR, BF, and SR, ranging from 0.636 to 0.773 with overlapping confidence intervals. LRs ranged from 1.34 (95% CI 1.09-1.65) for AR calculated with post-rupture dimensional adjustments to 2.14 (95% CI 1.45-3.14) for WTR and BF calculated without post-rupture adjustments. Logistic regression revealed a strong association between decreased WTR and rupture status. The point at which there is a 50% chance of presentation as ruptured was found to be WTR = 7.9 × 10^-3 when calculated without post-rupture adjustments and WTR' = 6.2 × 10^-3 when calculated with post-rupture adjustments, from which the proposed 6.1 × 10^-3 differs by 23% and 1.4%, respectively.

Conclusions: The model for IA rupture mechanics assessed in this study agrees reasonably well with clinical data and could serve as a foundation for further investigation. It additionally performs well in discriminating between ruptured and unruptured aneurysms, though its performance in this dataset is similar to more conventional, simpler parameters. Most importantly, this study demonstrates that biomathematical models can provide valuable insight into the nature of aneurysmal lesions despite their simplifying assumptions.

Objective: Assessing circulatory hemodynamics in-vitro is crucial for cardiovascular device design before in-vivo testing. Current mock circulation loops (MCLs) rely on simplified, lumped-parameter hydraulic representations of human circulation. There is a need for a more sophisticated MCL that can accurately represent the human circulatory physiology and allow for critical assessment of device-supported hemodynamics.

Methods: An anatomy-mimicking MCL design guided by one-dimensional flow models has been developed, using tree-like arterial casts to create a complex system. The MCL comprises cardiac simulators, systemic circulatory subsystems consisting of 46 connected arterial casts, and lumped venous and pulmonary components. A parameter tuning process was also developed to ensure that the simulated MCL baselines are consistent with targeted healthy or heart failure scenarios.

Results: Blood pressure and flow waveforms in the thoracic aorta, upper and lower limb arteries and abdominal organs (kidney, liver, spleen, etc.) were reproduced and validated against published data. Complex afferent and efferent flows in cerebral circulation and phasic coronary flow subjected to myocardial compression effect were replicated with precision. Pulse wave behavior was authentically generated and compared favorably to the published in-vivo and in-silico results.

Conclusion: This wave transport-preserving MCL is able to simulate pulsatile human circulatory hemodynamics with sufficient detail and accuracy. Complex cardiovascular device-intervened hemodynamics in large arteries and end organs can be systematically assessed using this new MCL, potentially contributing to a rapid and accurate in-vitro simulation to help advance device design and functional optimization.

Purpose: Type II endoleaks (T2EL) are a common complication after endovascular aneurysm repair. AneuFix is a newly designed elastic polymer for T2EL. AneuFix contains tantalum for visualization during fluoroscopy, which is crucial for monitoring the polymer in the side branches. The purpose of this study was to find the lowest concentration tantalum that is sufficient for safe injection in the aneurysmal sac.

Methods: AneuFix polymer with tantalum concentrations between 0 and 30% was injected into endoleak phantoms, connected to a pulsatile flow setup and with a realistic background for fluoroscopy. Furthermore, the radiopacity was investigated on fluoroscopic systems from three different vendors, using static phantoms. Results from both the dynamic and static phantoms were qualitatively evaluated by 10 clinical experts.

Results: Concentrations of ≥ 20% tantalum were consistently detected within the first 5 mm after entering the side branch, with a corresponding contrast-to-noise ratio of 2.23 ± 0.21. Furthermore, sufficient detectability scores (of at least 3 out of 5) were given to ≥ 15% tantalum. Significant differences were found in detectability scores on different fluoroscopic systems, using the default lowest-radiation-dose scan protocol for each system.

Conclusions: This study showed that tantalum concentrations ≥ 20% are consistently detected on fluoroscopy in the specified region. Compared to the original 30%, this would reduce imaging artifacts from high attenuation and scattering on follow-up imaging, while retaining sufficient detectability during injection. However, because of differences in fluoroscopic systems and scan protocols between hospitals, the combination of tantalum concentration and scan protocol should be optimized for each clinical setting.

Purpose: Catheter-based total renal denervation (TRDN) has recently gained FDA approval to lower blood pressure in patients with treatment-resistant hypertension. Current TRDN technologies indiscriminately destroy efferent (sympathetic) and afferent (sensory) renal nerves. However, preclinical studies suggest that the beneficial effects of TRDN may be due to ablation of afferent, rather than efferent, renal nerves. We developed a novel method for chemical ablation of afferent renal nerves by periaxonal application of capsaicin which has been employed in mouse and rat models of hypertension. In certain rodent models afferent-specific renal denervation (ARDN) has been shown to lower arterial pressure to the same degree as TRDN. The objective of the present study was to develop a catheter-based method for ARDN in a large animal model with the long-term goal of translating this treatment to humans. We tested the feasibility of using the Peregrine™ catheter infusion system, currently used to perform TRDN in humans by injection of ethanol, to perform catheter-based afferent renal denervation in sheep by periaxonal application of capsaicin.

Methods: Castrated, adult, male, Friesen sheep underwent Sham RDN (saline, n = 2), TRDN (ethanol, n = 4), or ARDN (capsaicin, n = 4) with the Peregrine™ catheter before termination > 2 weeks after the procedure. Denervation of renal efferents was verified by measurement of renal cortical norepinephrine (NE) content and anti-tyrosine hydroxylase (TH) staining; denervation of renal afferents was verified with anti-calcitonin gene-related peptide (CGRP) staining.

Results: There was a significant decrease in TH + and CGRP + fibers in TRDN kidneys and in CGRP + but not TH + fibers in ARDN kidneys. TRDN significantly reduced renal cortical norepinephrine (NE) content by 89% while ARDN had similar NE content to Sham RDN kidneys.

Conclusions: This study establishes the feasibility of performing catheter-based afferent renal denervation in a large animal model. Furthermore, this study provides a translational model to evaluate catheter-based ARDN as a potential treatment for hypertension.

Purpose: This study aims to investigate the effects of transcatheter edge-to-edge repair (TEER) on left ventricular hemodynamics and its potential implications for patient health.

Methods: An in vitro experimental platform was designed to replicate the anatomical and functional characteristics of the left ventricle (LV). This platform integrates native porcine mitral and aortic valves with a patient-specific 3D-printed silicone LV. The LV hemodynamics after TEER is assessed using echocardiography and particle image velocimetry, focusing on critical indices such as vorticity, Reynolds shear stress (RSS), viscous shear stress (VSS), and energy dissipation rate (ε).

Results: TEER effectively reduces the degree of mitral regurgitation (MR); however, it significantly increases RSS, VSS, and ε due to the formation of numerous small-scale vortices in the LV.

Conclusion: These hemodynamic changes may lead to adverse left ventricular remodeling, red blood cell damage, and reduced cardiac pumping efficiency, which have to be taken into consideration to optimize the TEER procedure and improve patient outcomes.

Purpose: To assess the impact of enhancements to the Atrioventricular Synchrony (AVS) algorithms of a next generation Micra leadless pacemaker (Micra AV2).

Methods: Accelerometer data were extracted from the AccelAV clinical study and were used to create virtual patients. A series of Monte Carlo simulations were run for each virtual patient to compare an enhanced Atrial Sensing Setup algorithm and Auto + A3 Threshold algorithm vs. original algorithms. A real-world survey was also conducted to observe clinical time savings from AVS programming burden reduction.

Results: The enhanced Atrial Sensing Setup in Micra AV2 devices demonstrated > 70% AVS in 27 of 30 (90%) patients while 13 of 30 (43%) Micra AV patients had > 70% AVS (p < 0.001) with no manual programming. The Micra AV2 Auto + A3 Threshold without additional manual programming demonstrated improved overall ambulatory AVS in the 80-100 bpm range (84.1%). Based upon survey results, the enhanced Atrial Sensing Setup algorithm accounted for an estimated reduction in median device check time of 13 min per patient.

Conclusions: Simulation-based analyses of the Micra AV2 leadless pacemaker projected significant improvements in automatic AVS at high sinus rates and an increase in the number of patients with > 70% AVS without clinician programming. Real-world survey results reported a reduction in device check time with the improvements.

Significance: Improvements in the AVS algorithms in Micra AV2 allow for better automatic AVS at higher heart rates and reduced clinic utilization burden.