Search for: All records

The overall purpose of this paper is to demonstrate how data preprocessing, training size variation, and subsampling can dynamically change the performance metrics of imbalanced text classification. The methodology encompasses using two different supervised learning classification approaches of feature engineering and data preprocessing with the use of five machine learning classifiers, five imbalanced sampling techniques, specified intervals of training and subsampling sizes, statistical analysis using R and tidyverse on a dataset of 1000 portable document format files divided into five labels from the World Health Organization Coronavirus Research Downloadable Articles of COVID-19 papers and PubMed Central databases of non-COVID-19 papers for binary classification that affects the performance metrics of precision, recall, receiver operating characteristic area under the curve, and accuracy. One approach that involves labeling rows of sentences based on regular expressions significantly improved the performance of imbalanced sampling techniques verified by performing statistical analysis using a t-test documenting performance metrics of iterations versus another approach that automatically labels the sentences based on how the documents are organized into positive and negative classes. The study demonstrates the effectiveness of ML classifiers and sampling techniques in text classification datasets, with different performance levels and class imbalance issues observed in manual and automatic methods of data processing. 

Abstract Some textured silicone breast implants with high average surface roughness (‘macrotextured’) have been associated with a rare cancer of the immune system, Breast Implant-Associated Anaplastic Large Cell Lymphoma (BIA-ALCL). Silicone elastomer wear debris may lead to chronic inflammation, a key step in the development of this cancer. Here, we model the generation and release of silicone wear debris in the case of a folded implant-implant (‘shell-shell’) sliding interface for three different types of implants, characterized by their surface roughness. The ‘smooth’ implant shell with the lowest average surface roughness tested (R a = 2.7 ± 0.6 μ m) resulted in average friction coefficients of μ avg = 0.46 ± 0.11 across 1,000 mm of sliding distance and generated 1,304 particles with an average particle diameter of D avg = 8.3 ± 13.1 μ m. The ‘microtextured’ implant shell (R a = 32 ± 7.0 μ m) exhibited μ avg = 1.20 ± 0.10 and generated 2,730 particles with D avg = 4.7 ± 9.1 μ m. The ‘macrotextured’ implant shell (R a = 80 ± 10 μ m) exhibited the highest friction coefficients, μ avg = 2.82 ± 0.15 and the greatest number of wear debris particles, 11,699, with an average particle size of D avg = 5.3 ± 3.3 μ m. Our data may provide guidance for the design of silicone breast implants with lower surface roughness, lower friction, and smaller quantities of wear debris. 

Silicone elastomer medical implants are ubiquitous in medicine, particularly for breast augmentation. However, when these devices are placed within the body, disruption of the natural biological interfaces occurs, which significantly changes the native energy-dissipation mechanisms of living systems. These new interfaces can introduce non-physiological contact pressures and tribological conditions that provoke inflammation and soft tissue damage. Despite their significance, the biotribological properties of implant-tissue and implant-extracellular matrix (ECM) interfaces remain poorly understood. Here, we developed an in vitro model of soft tissue damage using a custom-built in situ biotribometer mounted onto a confocal microscope. Sections of commercially-available silicone breast implants with distinct and clinically relevant surface roughness ([Formula: see text]m, [Formula: see text]m, and [Formula: see text]m) were mounted to spherically-capped hydrogel probes and slid against collagen-coated hydrogel surfaces as well as healthy breast epithelial (MCF10A) cell monolayers to model implant-ECM and implant-tissue interfaces. In contrast to the “smooth” silicone implants ([Formula: see text]m), we demonstrate that the “microtextured” silicone implant ([Formula: see text]m) induced higher frictional shear stress ([Formula: see text]  Pa), which led to greater collagen removal and cell rupture/delamination. Our studies may provide insights into post-implantation tribological interactions between silicone breast implants and soft tissues. 

Using numerical simulations, we probe the fluid flow in an axisymmetric peristaltic vessel fitted with elastic bi-leaflet valves. In this biomimetic system that mimics the flow generated in lymphatic vessels, we investigate the effects of the valve and vessel properties on pumping performance of the valved peristaltic vessel. The results indicate that valves significantly increase pumping by reducing backflow. The presence of valves, however, increases the viscous resistance, therefore requiring greater work compared to valveless vessels. The benefit of the valves is the most significant when the fluid is pumped against an adverse pressure gradient and for low vessel contraction wave speeds. We identify the optimum vessel and valve parameters leading to the maximum pumping efficiency. We show that the optimum valve elasticity maximizes the pumping flow rate by allowing the valve to block the backflow more effectively while maintaining low resistance during the forward flow. We also examine the pumping in vessels where the vessel contraction amplitude is a function of the adverse pressure gradient, as found in lymphatic vessels. We find that, in this case, the flow is limited by the work generated by the contracting vessel, suggesting that the pumping in lymphatic vessels is constrained by the performance of the lymphatic muscle. Given the regional heterogeneity of valve morphology observed throughout the lymphatic vasculature, these results provide insight into how these variations might facilitate efficient lymphatic transport in the vessel's local physiologic context.