BackgroundThe impact of different classes of microbial pathogens on mortality in severe community-acquired pneumonia is not well elucidated. Previous studies have shown significant variation in the incidence of viral, bacterial and mixed infections, with conflicting risk associations for mortality. We aimed to determine the risk association of microbial aetiologies with hospital mortality in severe CAP, utilising a diagnostic strategy incorporating molecular testing. Our primary hypothesis was that respiratory viruses were important causative pathogens in severe CAP and was associated with increased mortality when present with bacterial pathogens in mixed viral-bacterial co-infections.MethodsA retrospective cohort study from January 2014 to July 2015 was conducted in a tertiary hospital medical intensive care unit in eastern Singapore, which has a tropical climate. All patients diagnosed with severe community-acquired pneumonia were included.ResultsA total of 117 patients were in the study. Microbial pathogens were identified in 84 (71.8%) patients. Mixed viral-bacterial co-infections occurred in 18 (15.4%) of patients. Isolated viral infections were present in 32 patients (27.4%); isolated bacterial infections were detected in 34 patients (29.1%). Hospital mortality occurred in 16 (13.7%) patients. The most common bacteria isolated was Streptococcus pneumoniae and the most common virus isolated was Influenza A. Univariate and multivariate logistic regression showed that serum procalcitonin, APACHE II severity score and mixed viral-bacterial infection were associated with increased risk of hospital mortality. Mixed viral-bacterial co-infections were associated with an adjusted odds ratio of 13.99 (95% CI 1.30–151.05, p = 0.03) for hospital mortality.ConclusionsRespiratory viruses are common organisms isolated in severe community-acquired pneumonia. Mixed viral-bacterial infections may be associated with an increased risk of mortality.

With the rapid advances in remote-sensing technologies and the larger number of satellite images, fast and effective object detection plays an important role in understanding and analyzing image information, which could be further applied to civilian and military fields. Recently object detection methods with region-based convolutional neural network have shown excellent performance. However, these two-stage methods contain region proposal generation and object detection procedures, resulting in low computation speed. Because of the expensive manual costs, the quantity of well-annotated aerial images is scarce, which also limits the progress of geospatial object detection in remote sensing. In this paper, on the one hand, we construct and release a large-scale remote-sensing dataset for geospatial object detection (RSD-GOD) that consists of 5 different categories with 18,187 annotated images and 40,990 instances. On the other hand, we design a single shot detection framework with multi-scale feature fusion. The feature maps from different layers are fused together through the up-sampling and concatenation blocks to predict the detection results. High-level features with semantic information and low-level features with fine details are fully explored for detection tasks, especially for small objects. Meanwhile, a soft non-maximum suppression strategy is put into practice to select the final detection results. Extensive experiments have been conducted on two datasets to evaluate the designed network. Results show that the proposed approach achieves a good detection performance and obtains the mean average precision value of 89.0% on a newly constructed RSD-GOD dataset and 83.8% on the Northwestern Polytechnical University very high spatial resolution-10 (NWPU VHR-10) dataset at 18 frames per second (FPS) on a NVIDIA GTX-1080Ti GPU.

Inertial microfluidics has been proven to be a powerful tool for high-throughput, size-based cell sorting in diverse biomedical applications. In the case of Candida-related sepsis, Candida species and major blood cells (i.e., red blood cells and white blood cells) have a size distribution of 3−5 and 6−30 μm, respectively. To effectively retrieve a majority of Candida species and remove most of the interfering blood cells for accurate molecular analysis, inertial sorting of micron-sized biological particles with submicron size difference is highly desired, but far unexplored till now. In this work, we present a new channel design for an inertial microfluidic sorting device by embedding microsquares to construct periodic contractions along a series of repeating curved units. This unique channel design allows us to enhance inertial lift force at the microsquare zone and produce localized secondary Dean flow drag force in addition to global Dean flow drag force. This inertial sorting device has successfully separated 5.5 μm particles from 6.0 μm particles with a recovery ratio higher than 80% and a purity higher than 92%, demonstrating a size-based inertial sorting at submicron resolution (i.e., 0.5 μm). We further applied this inertial sorting device to purify Candida species from whole blood sample for enhanced molecular diagnosis of bloodstream Candida infection and especially compared it with the commonly used lysis-centrifugation-based purification method (STEM method) by recovering two species of Candida (Cornus glabrata and Candida albicans) from Candida-spiked blood samples. Through quantitative polymerase chain reaction (qPCR) analysis, we found that our inertial sorting approach has nearly 3-fold improvement on the pathogen recovery than the STEM method at pathogen abundances of 10 3 cfu/mL and 10 2 cfu/mL. The present inertial sorting at submicron resolution provides a simple, rapid, and efficient pathogen purification method for significantly improved molecular diagnosis of bloodstream Candida infection.

No abstract