Background Stent thrombosis is a lethal complication of endovascular intervention. Concern has been raised for the inherent risk associated with specific stent designs and drug-eluting coatings, yet clinical and animal support are equivocal. Methods and Results We examined whether drug-eluting coatings are inherently thrombogenic and if the response to these materials was determined to a greater degree by stent design and deployment using custom-built stents. Drug/polymer coatings uniformly reduce rather than increase thrombogenicity relative to matched bare-metal counterparts (0.65-fold, p=0.011). Thick-strutted (162 μm) stents were 1.5-fold more thrombogenic than otherwise identical thin-strutted (81 μm) devices in ex vivo flow loops (p<0.001), commensurate with 1.6-fold greater thrombus coverage three days after implantation in porcine coronary arteries (p=0.004). When bare-metal stents were deployed in malapposed or overlapping configurations, thrombogenicity increased compared to apposed, length-matched controls (1.58-fold, p=0.001 and 2.32-fold, p<0.001). The thrombogenicity of polymer-coated stents with thin struts was lowest in all configurations and remained insensitive to incomplete deployment. Computational modeling-based predictions of stent-induced flow derangements correlated with spatial distribution of formed clots. Conclusions Contrary to popular conception drug/polymer coatings do not inherently increase acute stent clotting – they reduce thrombosis. However, strut dimensions and positioning relative to the vessel wall are critical factors in modulating stent thrombogenicity. Optimal stent geometries and surfaces, as demonstrated with thin stent struts, help reduce the potential for thrombosis despite complex stent configurations and variability in deployment.

Artificial intelligence (AI) driven by machine learning (ML) algorithms is a branch in computer science that is rapidly gaining popularity within the healthcare sector. Recent regulatory approvals of AI-driven companion diagnostics and other products are glimmers of a future in which these tools could play a key role by defining the way medicine will be practiced. Educating the next generation of medical professionals with the right ML techniques will enable them to become part of this emerging data science revolution.

Individuals with CKD are particularly predisposed to thrombosis after vascular injury. Using mouse models, we recently described indoxyl sulfate, a tryptophan metabolite retained in CKD and an activator of tissue factor (TF) through aryl hydrocarbon receptor (AHR) signaling, as an inducer of thrombosis across the CKD spectrum. However, the translation of findings from animal models to humans is often challenging. Here, we investigated the uremic solute-AHR-TF thrombosis axis in two human cohorts, using a targeted metabolomics approach to probe a set of tryptophan products and high-throughput assays to measure AHR and TF activity. Analysis of baseline serum samples was performed from 473 participants with advanced CKD from the Dialysis Access Consortium Clopidogrel Prevention of Early AV Fistula Thrombosis trial. Participants with subsequent arteriovenous thrombosis had significantly higher levels of indoxyl sulfate and kynurenine, another uremic solute, and greater activity of AHR and TF, than those without thrombosis. Pattern recognition analysis using the components of the thrombosis axis facilitated clustering of the thrombotic and nonthrombotic groups. We further validated these findings using 377 baseline samples from participants in the Thrombolysis in Myocardial Infarction II trial, many of whom had CKD stage 2-3. Mechanistic probing revealed that kynurenine enhances thrombosis after vascular injury in an animal model and regulates thrombosis in an AHR-dependent manner. This human validation of the solute-AHR-TF axis supports further studies probing its utility in risk stratification of patients with CKD and exploring its role in other diseases with heightened risk of thrombosis.

Microfluidic fabrication technologies are emerging as viable platforms for extracorporeal lung assist devices and oxygenators for cardiac surgical support and critical care medicine, based in part on their ability to more closely mimic the architecture of the human vasculature than existing technologies. In comparison with current hollow fiber oxygenator technologies, microfluidic systems have more physiologically-representative blood flow paths, smaller cross section blood conduits and thinner gas transfer membranes. These features can enable smaller device sizes and a reduced blood volume in the oxygenator, enhanced gas transfer efficiencies, and may also reduce the tendency for clotting in the system. Several critical issues need to be addressed in order to advance this technology from its current state and implement it in an organ-scale device for clinical use. Here we report on the design, fabrication and characterization of multilayer microfluidic oxygenators, investigating scaling effects associated with fluid mechanical resistance, oxygen transfer efficiencies, and other parameters in multilayer devices. Important parameters such as the fluidic resistance of interconnects are shown to become more predominant as devices are scaled towards many layers, while other effects such as membrane distensibility become less significant. The present study also probes the relationship between blood channel depth and membrane thickness on oxygen transfer, as well as the rate of oxygen transfer on the number of layers in the device. These results contribute to our understanding of the complexity involved in designing three-dimensional microfluidic oxygenators for clinical applications.

Endovascular stents reside in a dynamic flow environment and yet the impact of flow on arterial drug deposition after stent-based delivery is only now emerging. We employed computational fluid dynamic modeling tools to investigate the influence of luminal flow patterns on arterial drug deposition and distribution. Flow imposes recirculation zones distal and proximal to the stent strut that extend the coverage of tissue absorption of eluted drug and induce asymmetry in tissue drug distribution. Our analysis now explains how the disparity in sizes of the two recirculation zones and the asymmetry in drug distribution are determined by a complex interplay of local flow and strut geometry. When temporal periodicity was introduced as a model of pulsatile flow, the net luminal flow served as an index of flow-mediated spatio-temporal tissue drug uptake. Dynamically changing luminal flow patterns are intrinsic to the coronary arterial tree. Coronary drug eluting stents should be appropriately considered where luminal flow, strut design and pulsatility have direct effects on tissue drug uptake after local delivery.

IntroductionChronic kidney damage is routinely assessed semiquantitatively by scoring the amount of fibrosis and tubular atrophy in a renal biopsy sample. Although image digitization and morphometric techniques can better quantify the extent of histologic damage, we need more widely applicable ways to stratify kidney disease severity.MethodsWe leveraged a deep learning architecture to better associate patient-specific histologic images with clinical phenotypes (training classes) including chronic kidney disease (CKD) stage, serum creatinine, and nephrotic-range proteinuria at the time of biopsy, and 1-, 3-, and 5-year renal survival. Trichrome-stained images processed from renal biopsy samples were collected on 171 patients treated at the Boston Medical Center from 2009 to 2012. Six convolutional neural network (CNN) models were trained using these images as inputs and the training classes as outputs, respectively. For comparison, we also trained separate classifiers using the pathologist-estimated fibrosis score (PEFS) as input and the training classes as outputs, respectively.ResultsCNN models outperformed PEFS across the classification tasks. Specifically, the CNN model predicted the CKD stage more accurately than the PEFS model (κ = 0.519 vs. 0.051). For creatinine models, the area under curve (AUC) was 0.912 (CNN) versus 0.840 (PEFS). For proteinuria models, AUC was 0.867 (CNN) versus 0.702 (PEFS). AUC values for the CNN models for 1-, 3-, and 5-year renal survival were 0.878, 0.875, and 0.904, respectively, whereas the AUC values for PEFS model were 0.811, 0.800, and 0.786, respectively.ConclusionThe study demonstrates a proof of principle that deep learning can be applied to routine renal biopsy images.

Chronic kidney disease (CKD/uremia) remains vexing because it increases the risk of atherothrombosis and is also associated with bleeding complications on standard antithrombotic/antiplatelet therapies. Although the associations of indolic uremic solutes and vascular wall proteins [such as tissue factor (TF) and aryl hydrocarbon receptor (AHR)] are being defined, the specific mechanisms that drive the thrombotic and bleeding risks are not fully understood. We now present an indolic solute–specific animal model, which focuses on solute-protein interactions and shows that indolic solutes mediate the hyperthrombotic phenotype across all CKD stages in an AHR- and TF-dependent manner. We further demonstrate that AHR regulates TF through STIP1 homology and U-box–containing protein 1 (STUB1). As a ubiquitin ligase, STUB1 dynamically interacts with and degrades TF through ubiquitination in the uremic milieu. TF regulation by STUB1 is supported in humans by an inverse relationship of STUB1 and TF expression and reduced STUB1-TF interaction in uremic vessels. Genetic or pharmacological manipulation of STUB1 in vascular smooth muscle cells inhibited thrombosis in flow loops. STUB1 perturbations reverted the uremic hyperthrombotic phenotype without prolonging the bleeding time, in contrast to heparin, the standard-of-care antithrombotic in CKD patients. Our work refines the thrombosis axis (STUB1 is a mediator of indolic solute–AHR-TF axis) and expands the understanding of the interconnected relationships driving the fragile thrombotic state in CKD. It also establishes a means of minimizing the uremic hyperthrombotic phenotype without altering the hemostatic balance, a long-sought-after combination in CKD patients.

IntroductionThe number of glomeruli and glomerulosclerosis evaluated on kidney biopsy slides constitute as standard components of a renal pathology report. Prevailing methods for glomerular assessment remain manual, labor intensive and non-standardized. We developed a deep learning framework to accurately identify and segment glomeruli from digitized images of human kidney biopsies.MethodsTrichrome-stained images (n=275) from renal biopsies of 171 chronic kidney disease patients treated at the Boston Medical Center from 2009-12 were analyzed. A sliding window operation was defined to crop each original image to smaller images. Each cropped image was then evaluated by three experts into three categories: (a) No glomerulus, (b) Normal or partially sclerosed glomerulus and (c) Globally sclerosed glomerulus. This led to identification of 751 unique images representing non­glomerular regions, 611 images with either normal or partially sclerosed (NPS) glomeruli and 134 images with globally sclerosed (GS) glomeruli. A convolutional neural network (CNN) was trained with cropped images as inputs and corresponding labels as output. Using this model, an image processing routine was developed to scan the test data images to segment the GS glomeruli.ResultsThe CNN model was able to accurately discriminate non-glomerular images from NPS and GS images (Performance on test data - Accuracy: 92.67±2.02% and Kappa: 0.8681±0.0392). The segmentation model that was based on the CNN multi-label classifier accurately marked the GS glomeruli on the test data (Matthews correlation coefficient = 0.628).ConclusionThis work demonstrates the power of deep learning for assessing complex histologic structures from digitized human kidney biopsies.