Key points The capillary module, consisting of parallel capillaries from arteriole to venule, is classically considered as the building block of complex capillary networks. In skeletal muscle, this structure fails to address how blood flow is regulated along the entire length of the synchronously contracting muscle fibres. Using intravital video microscopy of resting extensor digitorum longus muscle in rats, we demonstrated the capillary fascicle as a series of interconnected modules forming continuous columns that align naturally with the dimensions of the muscle fascicle. We observed structural heterogeneity for module topology, and functional heterogeneity in space and time for capillary‐red blood cell (RBC) haemodynamics within a module and between modules. We found that module RBC haemodynamics were independent of module resistance, providing direct evidence for microvascular flow regulation at the level of the capillary module. The capillary fascicle is an updated paradigm for characterizing blood flow and RBC distribution in skeletal muscle capillary networks. Abstract Capillary networks are the fundamental site of oxygen exchange in the microcirculation. The capillary module (CM), consisting of parallel capillaries from terminal arteriole (TA) to post‐capillary venule (PCV), is classically considered as the building block of complex capillary networks. In skeletal muscle, this structure fails to address how blood flow is regulated along the entire length of the synchronously contracting muscle fibres, requiring co‐ordination from numerous modules. It has previously been recognized that TAs and PCVs interact with multiple CMs, creating interconnected networks. Using label‐free intravital video microscopy of resting extensor digitorum longus muscle in rats, we found that these networks form continuous columns of linked CMs spanning thousands of microns, herein denoted as the capillary fascicle (CF); this structure aligns naturally with the dimensions of the muscle fascicle. We measured capillary‐red blood cell (RBC) haemodynamics and module topology (n = 9 networks, 327 modules, 1491 capillary segments). The average module had length 481 μm, width 157 μm and 9.51 parallel capillaries. We observed structural heterogeneity for CM topology, and functional heterogeneity in space and time for capillary‐RBC haemodynamics within a module and between modules. There was no correlation between capillary RBC velocity and lineal density. A passive inverse relationship between module length and haemodynamics was remarkably absent, providing direct evidence for microvascular flow regulation at the level of the CM. In summary, the CF is an updated paradigm for characterizing RBC distribution in skeletal muscle, and strengthens the theory of capillary networks as major contributors to the signal that regulates capillary perfusion.

Fractures of the cervical spine are a medical emergency and may lead to permanent paralysis and even death. Accurate diagnosis in patients with suspected fractures by computed tomography (CT) is critical to patient management. In this paper, we propose a deep convolutional neural network (DCNN) with a bidirectional long-short term memory (BLSTM) layer for the automated detection of cervical spine fractures in CT axial images. We used an annotated dataset of 3,666 CT scans (729 positive and 2,937 negative cases) to train and validate the model. The validation results show a classification accuracy of 70.92% and 79.18% on the balanced (104 positive and 104 negative cases) and imbalanced (104 positive and 419 negative cases) test datasets, respectively.

The capillary module (CM), consisting of parallel capillaries from terminal arteriole to post‐capillary venule, is classically considered to be the building block of complex capillary networks. In skeletal muscle, CMs form interconnected columns spanning thousands of microns, which we recently described as the capillary fascicle. However, detailed evaluation of CM haemodynamics has not been described, and may provide insight into mechanisms of blood flow regulation in the microcirculation. We used intravital videomicroscopy from resting extensor digitorum longus muscle in rats (n = 9 networks, 112 capillary modules), as well as dual‐phase computational modelling of blood flow in simulated CM geometries. We found that the mean driving pressure across CMs was 3.236 ± 1.833 mmHg. Red blood cell (RBC) flow was independent of CM resistance, and the ratio of blood flow in adjacent modules was not correlated with their ratio of resistances. In simulated CM geometries, increases to driving pressure produced a direct linear increase to RBC and plasma flow, with no changes to RBC distribution; increases to arteriolar inflow haematocrit resulted in increased RBC flow, but with viscosity‐dependent increases to CM resistance. CM RBC flow heterogeneity was higher than plasma flow heterogeneity in experimental data, in contrast to simulated geometries, suggesting that time‐dependent flow variability may have important consequences for RBC distribution. In summary, these findings suggest that CMs are active participants in microvascular flow regulation, likely achieved through adjustments to CM driving pressure using pre‐ and post‐capillary loci of flow control. Increases to CM viscosity may be important during the regulation of functional hyperaemia. Key points The capillary module (CM), consisting of parallel capillaries from the arteriole to venule, is classically considered to be the building block of capillary networks in skeletal muscle. A detailed evaluation of module haemodynamics may provide insight into mechanisms of blood flow regulation in the microcirculation. Using experimental data from resting skeletal muscle in rats, as well as dual‐phase computational models of blood flow, we analysed haemodynamic relationships and the impact of variations to boundary conditions on red blood cell and plasma distribution. We showed that driving pressure across CMs is low, and that simulated increases to inflow haematocrit have important viscosity‐dependent effects on module resistance. We found that red blood cell flow was independent from module resistance, which strongly suggests the regulation of driving pressure at the level of the capillary module using pre‐ and post‐capillary loci of flow control. These findings place CMs as central participants in microvascular flow regulation, with important consequences for disease and functional hyperaemia.

No abstract

Purpose Vision impairment affects 2.2 billion people worldwide, half of which is preventable with early detection and treatment. Currently, automatic screening of ocular pathologies using convolutional neural networks (CNNs) on retinal fundus photographs is limited to a few pathologies. Simultaneous detection of multiple ophthalmic pathologies would increase clinical usability and uptake. Methods Two thousand five hundred sixty images were used from the Retinal Fundus Multi-Disease Image Dataset (RFMiD). Models were trained ( n = 1920) and validated ( n = 640). Five selected CNN architectures were trained to predict the presence of any pathology and categorize the 28 pathologies. All models were trained to minimize asymmetric loss, a modified form of binary cross-entropy. Individual model predictions were averaged to obtain a final ensembled model and assessed for mean area under the receiver-operator characteristic curve (AUROC) for disease screening (healthy versus pathologic image) and classification (AUROC for each class). Results The ensemble network achieved a disease screening (healthy versus pathologic) AUROC score of 0.9613. The highest single network score was 0.9586 using the SE-ResNeXt architecture. For individual disease classification, the average AUROC score for each class was 0.9295. Conclusions Retinal fundus images analyzed by an ensemble of CNNs trained to minimize asymmetric loss were effective in detection and classification of ocular pathologies than individual models. External validation is needed to translate machine learning models to diverse clinical contexts. Translational Relevance This study demonstrates the potential benefit of ensemble-based deep learning methods on improving automatic screening and diagnosis of multiple ocular pathologies from fundoscopy imaging.