Arteriovenous shunt stenosis assessment based on empirical mode decomposition and 1D-convolutional neural network: Clinical trial stage

Effect of blood viscosity on the hemodynamics of arteriovenous fistulae based on numerical investigation

Arteriovenous fistula (AVF) is the most commonly used vascular access for hemodialysis in patients with end-stage renal disease. Vascular diseases such as atherosclerosis and thrombosis, triggered by altered hemodynamic conditions, are the main causes of access failure. Changes in blood viscosity accelerate access dysfunction by affecting local velocities and wall shear stress (WSS) distribution in the circulation. Numerical simulation was employed to analyze and compare the hemodynamic behavior of AVF under different blood viscosities (0.001-0.012 Pa∙s). An idealized three-dimensional model with end-to-side anastomosis was established. Transient simulations were conducted using pulsatile inlet velocity and outflow as boundary conditions. The simulation results reveal the blood flow state of AVF under different viscosity physiological conditions and derive the rule of change. When blood viscosity increases, the local velocity in the disturbed region slows down and the stagnation time becomes longer, resulting in increased deposition of substances. As blood viscosity increases, the level of shear stress on the entire wall of the fistula increases accordingly. WSS values at high viscosities above 0.007 Pa∙s showed significantly larger low-shear regions near the anastomosis and increased chances of inducing atheromatous plaques. This research has revealed the correlation between blood dynamic viscosity and the hemodynamic behavior of AVF. Elevated whole blood viscosity increases the incidence of access obstruction and vascular disease leading to fistula failure. The study provides a basis for optimizing the distribution of hemodynamic parameters in the fistula for hemodialysis patients.

Structural design and mechanical performance of composite vascular grafts

This study reviews the state of the art in structural design and the corresponding mechanical behaviours of composite vascular grafts. We critically analyse surface and matrix designs composed of layered, embedded, and hybrid structures along the radial and longitudinal directions; materials and manufacturing techniques, such as tissue engineering and the use of textiles or their combinations; and the corresponding mechanical behaviours of composite vascular grafts in terms of their physical–mechanical properties, especially their stress–strain relationships and elastic recovery. The role of computational studies is discussed with respect to optimizing the geometrics designs and the corresponding mechanical behaviours to satisfy specialized applications, such as those for the aorta and its subparts. Natural and synthetic endothelial materials yield improvements in the mechanical and biological compliance of composite graft surfaces with host arteries. Moreover, the diameter, wall thickness, stiffness, compliance, tensile strength, elasticity, and burst strength of the graft matrix are determined depending on the application and the patient. For composite vascular grafts, hybrid architectures are recommended featuring multiple layers, dimensions, and materials to achieve the desired optimal flexibility and function for complying with user-specific requirements. Rapidly emerging artificial intelligence and big data techniques for diagnostics and the three-dimensional (3D) manufacturing of vascular grafts will likely yield highly compliant, subject-specific, long-lasting, and economical vascular grafts in the near-future.

Graphic abstract

Breast Lesions Screening of Mammographic Images with 2D Spatial and 1D Convolutional Neural Network-Based Classifier

Mammography is a first-line imaging examination that employs low-dose X-rays to rapidly screen breast tumors, cysts, and calcifications. This study proposes a two-dimensional (2D) spatial and one-dimensional (1D) convolutional neural network (CNN) to early detect possible breast lesions (tumors) to reduce patients’ mortality rates and to develop a classifier for use in mammographic images on regions of interest where breast lesions (tumors) may likely occur. The 2D spatial fractional-order convolutional processes are used to strengthen and sharpen the lesions’ features, denoise, and improve the feature extraction processes. Then, an automatic extraction task is performed using a specific bounding box to sequentially pick out feature patterns from each mammographic image. The multi-round 1D kernel convolutional processes can also strengthen and denoise 1D feature signals and assist in the identification of the differentiation levels of normality and abnormality signals. In the classification layer, a gray relational analysis-based classifier is used to screen the possible lesions, including normal (Nor), benign (B), and malignant (M) classes. The classifier development for clinical applications can reduce classifier’s training time, computational complexity level, computational time, and achieve a more accurate rate for meeting clinical/medical purpose. Mammographic images were selected from the mammographic image analysis society image database for experimental tests on breast lesions screening and K-fold cross-validations were performed. The experimental results showed promising performance in quantifying the classifier’s outcome for medical purpose evaluation in terms of recall (%), precision (%), accuracy (%), and F1 score.

Posteroanterior Chest X-ray Image Classification with a Multilayer 1D Convolutional Neural Network-Based Classifier for Cardiomegaly Level Screening

Palpitations, chest tightness, and shortness of breath are early indications of cardiomegaly, which is an asymptomatic disease. Their causes and treatment strategies are different due to differing indications. Hence, early screening of cardiomegaly levels can be used to make a strategy for administering drugs and surgical treatments. In this study, we will establish a multilayer one-dimensional (1D) convolutional neural network (CNN)-based classifier for automatic cardiomegaly level screening based on chest X-ray (CXR) image classification in frontal posteroanterior view. Using two-round 1D convolutional processes in the convolutional pooling layer, two-dimensional (2D) feature maps can be converted into feature signals, which can enhance their characteristics for identifying normal condition and cardiomegaly levels. In the classification layer, a classifier based on gray relational analysis, which has a straightforward mathematical operation, is used to screen the cardiomegaly levels. Based on the collected datasets from the National Institutes of Health CXR image database, the proposed multilayer 1D CNN-based classifier with K-fold cross-validation has promising results for the intended medical purpose, with precision of 97.80%, recall of 98.20%, accuracy of 98.00%, and F1 score of 0.9799.

Photoplethysmography Analysis with Duffing–Holmes Self-Synchronization Dynamic Errors and 1D CNN-Based Classifier for Upper Extremity Vascular Disease Screening

Common upper limb peripheral artery diseases (PADs) are atherosclerosis, embolic diseases, and systemic diseases, which are often asymptomatic, and the narrowed arteries (stenosis) will gradually reduce blood flow in the right or left upper limbs. Upper extremity vascular disease (UEVD) and atherosclerosis are high-risk PADs for patients with Type 2 diabetes or with both diabetes and end-stage renal disease. For early UEVD detection, a fingertip-based, toe-based, or wrist-based photoplethysmography (PPG) tool is a simple and noninvasive measurement system for vital sign monitoring and healthcare applications. Based on time-domain PPG analysis, a Duffing–Holmes system with a master system and a slave system is used to extract self-synchronization dynamic errors, which can track the differences in PPG morphology (in amplitudes (systolic peak) and time delay (systolic peak to diastolic peak)) between healthy subjects and PAD patients. In the preliminary analysis, the self-synchronization dynamic errors can be used to evaluate risk levels based on the reflection index (RI), which includes normal condition, lower PAD, and higher PAD. Then, a one-dimensional convolutional neural network is established as a multilayer classifier for automatic UEVD screening. The experimental results indicated that the self-synchronization dynamic errors have a positive correlation with the RI (R2 = 0.6694). The K-fold cross-validation is used to verify the performance of the proposed classifier with recall (%), precision (%), accuracy (%), and F1 score.