Lesion-aware cross-phase attention network for renal tumor subtype classification on multi-phase CT scans.

Multi-phase computed tomography (CT) has been widely used for the preoperative diagnosis of kidney cancer due to its non-invasive nature and ability to characterize renal lesions. However, since enhancement patterns of renal lesions across CT phases are different even for the same lesion type, the visual assessment by radiologists suffers from inter-observer variability in clinical practice. Although deep learning-based approaches have been recently explored for differential diagnosis of kidney cancer, they do not explicitly model the relationships between CT phases in the network design, limiting the diagnostic performance. In this paper, we propose a novel lesion-aware cross-phase attention network (LACPANet) that can effectively capture temporal dependencies of renal lesions across CT phases to accurately classify the lesions into five major pathological subtypes from time-series multi-phase CT images. We introduce a 3D inter-phase lesion-aware attention mechanism to learn effective 3D lesion features that are used to estimate attention weights describing the inter-phase relations of the enhancement patterns. We also present a multi-scale attention scheme to capture and aggregate temporal patterns of lesion features at different spatial scales for further improvement. Extensive experiments on multi-phase CT scans of kidney cancer patients from the collected dataset demonstrate that our LACPANet outperforms state-of-the-art approaches in diagnostic accuracy.

Cross-domain Denoising for Low-dose Multi-frame Spiral Computed Tomography.

Computed tomography (CT) has been used worldwide as a non-invasive test to assist in diagnosis. However, the ionizing nature of X-ray exposure raises concerns about potential health risks such as cancer. The desire for lower radiation doses has driven researchers to improve reconstruction quality. Although previous studies on low-dose computed tomography (LDCT) denoising have demonstrated the effectiveness of learning-based methods, most were developed on the simulated data. However, the real-world scenario differs significantly from the simulation domain, especially when using the multi-slice spiral scanner geometry. This paper proposes a two-stage method for the commercially available multi-slice spiral CT scanners that better exploits the complete reconstruction pipeline for LDCT denoising across different domains. Our approach makes good use of the high redundancy of multi-slice projections and the volumetric reconstructions while leveraging the over-smoothing issue in conventional cascaded frameworks caused by aggressive denoising. The dedicated design also provides a more explicit interpretation of the data flow. Extensive experiments on various datasets showed that the proposed method could remove up to 70% of noise without compromised spatial resolution, while subjective evaluations by two experienced radiologists further supported its superior performance against state-of-the-art methods in clinical practice. Code is available at https://github.com/YCL92/TMD-LDCT.

RefQSR: Reference-Based Quantization for Image Super-Resolution Networks.

Single image super-resolution (SISR) aims to reconstruct a high-resolution image from its low-resolution observation. Recent deep learning-based SISR models show high performance at the expense of increased computational costs, limiting their use in resource-constrained environments. As a promising solution for computationally efficient network design, network quantization has been extensively studied. However, existing quantization methods developed for SISR have yet to effectively exploit image self-similarity, which is a new direction for exploration in this study. We introduce a novel method called reference-based quantization for image super-resolution (RefQSR) that applies high-bit quantization to several representative patches and uses them as references for low-bit quantization of the rest of the patches in an image. To this end, we design dedicated patch clustering and reference-based quantization modules and integrate them into existing SISR network quantization methods. The experimental results demonstrate the effectiveness of RefQSR on various SISR networks and quantization methods.

Stabilizing the Catalyst Layer for Durable and High Performance Alkaline Membrane Fuel Cells and Water Electrolyzers.

Anion exchange membrane (AEM) fuel cells (AEMFCs) and water electrolyzers (AEMWEs) suffer from insufficient performance and durability compared with commercialized energy conversion systems. Great efforts have been devoted to designing high-quality AEMs and catalysts. However, the significance of the stability of the catalyst layer has been largely disregarded. Here, an in situ cross-linking strategy was developed to promote the interactions within the catalyst layer and the interactions between catalyst layer and AEM. The adhesion strength of the catalyst layer after cross-linking was improved 7 times compared with the uncross-linked catalyst layer due to the formation of covalent bonds between the catalyst layer and AEM. The AEMFC can be operated under 0.6 A cm-2 for 1000 h with a voltage decay rate of 20 µV h-1. The related AEMWE achieved an unprecedented current density of 15.17 A cm-2 at 2.0 V and was operated at 0.5, 1.0, and 1.5 A cm-2 for 1000 h.

Ecological Interaction between Bacteriophages and Bacteria in Sub-Arctic Kongsfjorden Bay, Svalbard, Norway.

Marine virus diversity and their relationships with their hosts in the marine environment remain unclear. This study investigated the co-occurrence of marine DNA bacteriophages (phages) and bacteria in the sub-Arctic area of Kongsfjorden Bay in Svalbard (Norway) in April and June 2018 using metagenomics tools. Of the marine viruses identified, 48-81% were bacteriophages of the families Myoviridae, Siphoviridae, and Podoviridae. Puniceispirillum phage HMO-2011 was dominant (7.61%) in April, and Puniceispirillum phage HMO-2011 (3.32%) and Pelagibacter phage HTVC008M (3.28%) were dominant in June. Gammaproteobacteria (58%), including Eionea flava (14.3%) and Pseudomonas sabulinigri (12.2%), were dominant in April, whereas Alphaproteobacteria (87%), including Sulfitobacter profundi (51.5%) and Loktanella acticola (32.4%), were dominant in June. The alpha diversity of the bacteriophages and bacterial communities exhibited opposite patterns. The diversity of the bacterial community was higher in April and lower in June. Changes in water temperature and light can influence the relationship between bacteria and bacteriophages.

Global, Regional, and National Epidemiology of Alopecia Areata: A Systematic Review and Modelling Study.

BACKGROUND: Alopecia areata (AA) is a chronic autoimmune disease that leads to a high psychiatric, economic, and systemic disease burden. A comprehensive understanding of AA epidemiology is essential for evaluating healthcare source utilization; however, there is a lack of systematic approach for summarizing epidemiologic data on AA. OBJECTIVES: To systematically investigate the global, regional, and national incidence and prevalence of AA. METHODS: A structured search was conducted using the Ovid MEDLINE, EMBASE, Cochrane Library, Web of Science, SciELO, and Korean journal databases from their inception date to October 4, 2023. Studies that reported the prevalence or incidence of AA were included. We used a Bayesian hierarchical linear mixed model to analyse the prevalence estimates. The primary outcomes of our study were the global, regional, and national prevalence of physician-diagnosed AA for overall population, adults, and children. The incidence data were summarised descriptively. RESULTS: In total, 88 studies from 28 countries were included in the analysis. The reported incidence of alopecia areata tended to be higher in adults aged 19-50 years, and this trend was consistent with its estimated prevalence. The reported prevalence in overall population tended to be higher in men compared to in women. The estimated lifetime prevalence of AA was 0.10% (95% credible intervals, 0.03%-0.39%) in the general population worldwide, 0.12% (95% credible intervals, 0.02%-0.52%) in adults, and 0.03% (95% credible intervals, 0.01%-0.12%) in children. The estimated prevalence was highest in the Asian region and lowest in the African region. CONCLUSIONS: In this study, 48% of the total Global Burden of Disease regions had insufficient data reporting the prevalence or incidence of AA. Further studies are needed to provide epidemiological information on middle- and low-income countries. Our study can serve as a crucial reference in terms of healthcare policy decisions.

Triptycene Branched Poly(aryl-co-aryl piperidinium) Electrolytes for Alkaline Anion Exchange Membrane Fuel Cells and Water Electrolyzers.

Alkaline polymer electrolytes (APEs) are essential materials for alkaline energy conversion devices such as anion exchange membrane fuel cells (AEMFCs) and water electrolyzers (AEMWEs). Here, we report a series of branched poly(aryl-co-aryl piperidinium) with different branching agents (triptycene: highly-rigid, three-dimensional structure; triphenylbenzene: planar, two-dimensional structure) for high-performance APEs. Among them, triptycene branched APEs showed excellent hydroxide conductivity (193.5 mS cm-1 @80 °C), alkaline stability, mechanical properties, and dimensional stability due to the formation of branched network structures, and increased free volume. AEMFCs based on triptycene-branched APEs reached promising peak power densities of 2.503 and 1.705 W cm-2 at 75/100 % and 30/30 % (anode/cathode) relative humidity, respectively. In addition, the fuel cells can run stably at a current density of 0.6 A cm-2 for 500 h with a low voltage decay rate of 46 µV h-1 . Importantly, the related AEMWE achieved unprecedented current densities of 16 A cm-2 and 14.17 A cm-2 (@2 V, 80 °C, 1 M NaOH) using precious and non-precious metal catalysts, respectively. Moreover, the AEMWE can be stably operated under 1.5 A cm-2 at 60 °C for 2000 h. The excellent results suggest that the triptycene-branched APEs are promising candidates for future AEMFC and AEMWE applications.

High Free Volume Polyelectrolytes for Anion Exchange Membrane Water Electrolyzers with a Current Density of 13.39 A cm-2 and a Durability of 1000 h.

The rational design of the current anion exchange polyelectrolytes (AEPs) is challenging to meet the requirements of both high performance and durability in anion exchange membrane water electrolyzers (AEMWEs). Herein, highly-rigid-twisted spirobisindane monomer is incorporated in poly(aryl-co-aryl piperidinium) backbone to construct continuous ionic channels and to maintain dimensional stability as promising materials for AEPs. The morphologies, physical, and electrochemical properties of the AEPs are investigated based on experimental data and molecular dynamics simulations. The present AEPs possess high free volumes, excellent dimensional stability, hydroxide conductivity (208.1 mS cm-1 at 80 °C), and mechanical properties. The AEMWE of the present AEPs achieves a new current density record of 13.39 and 10.7 A cm-2 at 80 °C by applying IrO2 and nonprecious anode catalyst, respectively, along with outstanding in situ durability under 1 A cm-2 for 1000 h with a low voltage decay rate of 53 µV h-1 . Moreover, the AEPs can be applied in fuel cells and reach a power density of 2.02 W cm-2 at 80 °C under fully humidified conditions, and 1.65 W cm-2 at 100 °C, 30% relative humidity. This study provides insights into the design of high-performance AEPs for energy conversion devices.

Reversed domain adaptation for nuclei segmentation-based pathological image classification.

Despite the fact that digital pathology has provided a new paradigm for modern medicine, the insufficiency of annotations for training remains a significant challenge. Due to the weak generalization abilities of deep-learning models, their performance is notably constrained in domains without sufficient annotations. Our research aims to enhance the model's generalization ability through domain adaptation, increasing the prediction ability for the target domain data while only using the source domain labels for training. To further enhance classification performance, we introduce nuclei segmentation to provide the classifier with more diagnostically valuable nuclei information. In contrast to the general domain adaptation that generates source-like results in the target domain, we propose a reversed domain adaptation strategy that generates target-like results in the source domain, enabling the classification model to be more robust to inaccurate segmentation results. The proposed reversed unsupervised domain adaptation can effectively reduce the disparities in nuclei segmentation between the source and target domains without any target domain labels, leading to improved image classification performance in the target domain. The whole framework is designed in a unified manner so that the segmentation and classification modules can be trained jointly. Extensive experiments demonstrate that the proposed method significantly improves the classification performance in the target domain and outperforms existing general domain adaptation methods.

Marine Bacterioplankton Community Dynamics and Potentially Pathogenic Bacteria in Seawater around Jeju Island, South Korea, via Metabarcoding.

Understanding marine bacterioplankton composition and distribution is necessary for improving predictions of ecosystem responses to environmental change. Here, we used 16S rRNA metabarcoding to investigate marine bacterioplankton diversity and identify potential pathogenic bacteria in seawater samples collected in March, May, September, and December 2013 from two sites near Jeju Island, South Korea. We identified 1343 operational taxonomic units (OTUs) and observed that community diversity varied between months. Alpha- and Gamma-proteobacteria were the most abundant classes, and in all months, the predominant genera were Candidatus Pelagibacter, Leisingera, and Citromicrobium. The highest number of OTUs was observed in September, and Vibrio (7.80%), Pseudoalteromonas (6.53%), and Citromicrobium (6.16%) showed higher relative abundances or were detected only in this month. Water temperature and salinity significantly affected bacterial distribution, and these conditions, characteristic of September, were adverse for Aestuariibacter but favored Citromicrobium. Potentially pathogenic bacteria, among which Vibrio (28 OTUs) and Pseudoalteromonas (six OTUs) were the most abundant in September, were detected in 49 OTUs, and their abundances were significantly correlated with water temperature, increasing rapidly in September, the warmest month. These findings suggest that monthly temperature and salinity variations affect marine bacterioplankton diversity and potential pathogen abundance.

Deep Learning-Based Evaluation of Ultrasound Images for Benign Skin Tumors.

In this study, a combined convolutional neural network for the diagnosis of three benign skin tumors was designed, and its effectiveness was verified through quantitative and statistical analysis. To this end, 698 sonographic images were taken and diagnosed at the Department of Dermatology at Severance Hospital in Seoul, Korea, between 10 November 2017 and 17 January 2020. Through an empirical process, a convolutional neural network combining two structures, which consist of a residual structure and an attention-gated structure, was designed. Five-fold cross-validation was applied, and the train set for each fold was augmented by the Fast AutoAugment technique. As a result of training, for three benign skin tumors, an average accuracy of 95.87%, an average sensitivity of 90.10%, and an average specificity of 96.23% were derived. Also, through statistical analysis using a class activation map and physicians' findings, it was found that the judgment criteria of physicians and the trained combined convolutional neural network were similar. This study suggests that the model designed and trained in this study can be a diagnostic aid to assist physicians and enable more efficient and accurate diagnoses.

Selective attachment of prokaryotes and emergence of potentially pathogenic prokaryotes on four plastic surfaces: Adhesion study in a natural marine environment.

This study employed 16S rRNA metabarcoding to establish the diversity of prokaryotic communities and specific characteristics of potentially pathogenic prokaryotic primary colonizers of four plastic materials (EPS, expanded polystyrene; PE, polyethylene; PP, polypropylene; and PET, polyethylene terephthalate). Bacteria inhabiting plastic and seawater differ; thus, distinct changes in the attached prokaryotic community were observed over an exposure time of 21 days, specifically on Days 3, 6, 9, and 12-21. Frist colonizers were Gammaproteobacteria and Alphaproteobacteria; Bacilli and Clostridia represented secondary colonizers. On Day 3, Pseudoalteromonas had a relative abundance >80 %; whereas, the prevalence of Vibrio spp. (potentially pathogenic prokaryotes) increased rapidly on Days 6 and 9. However, after Day 12, the prevalence of other potential pathogens, namely, Clostridium spp., steadily increased. Despite the diversity of the plastic surfaces, attached prokaryotes changed over time instead of showing similar adherent diversity in all plastic materials.

Metavirome Profiling and Dynamics of the DNA Viral Community in Seawater in Chuuk State, Federated States of Micronesia.

Despite their abundance and ecological importance, little is known about the diversity of marine viruses, in part because most cannot be cultured in the laboratory. Here, we used high-throughput viral metagenomics of uncultivated viruses to investigate the dynamics of DNA viruses in tropical seawater sampled from Chuuk State, Federated States of Micronesia, in March, June, and December 2014. Among the identified viruses, 71-79% were bacteriophages belonging to the families Myoviridae, Siphoviridae, and Podoviridae (Caudoviriales), listed in order of abundance at all sampling times. Although the measured environmental factors (temperature, salinity, and pH) remained unchanged in the seawater over time, viral dynamics changed. The proportion of cyanophages (34.7%) was highest in June, whereas the proportion of mimiviruses, phycodnaviruses, and other nucleo-cytoplasmic large DNA viruses (NCLDVs) was higher in March and December. Although host species were not analysed, the dramatic viral community change observed in June was likely due to changes in the abundance of cyanophage-infected cyanobacteria, whereas that in NCLDVs was likely due to the abundance of potential eukaryote-infected hosts. These results serve as a basis for comparative analyses of other marine viral communities, and guide policy-making when considering marine life care in Chuuk State.

Co-variance between free-living bacteria and Cochlodinium polykrikoides (Dinophyta) harmful algal blooms, South Korea.

To understand the co-variance between common free-living bacteria and Cochlodinium polykrikoides harmful algal blooms (HABs) and their metabolic functions, we investigated 110 sampling sites in the Southern Sea of South Korea. These sampling sites were divided into three groups based on environmental factors and phytoplankton data with a similarity of 85% using non-metric multidimensional scaling. One group represented high-severity C. polykrikoides blooms, while the other two represented low-severity or no blooms. In high-severity HABs, inorganic phosphorous and dissolved organic carbon concentrations were strongly correlated with C. polykrikoides density (p < 0.01). This may reflect the changes in biochemical cycling due to inorganic and organic substrates released by HAB cells (or by cell destruction). Furthermore, 88 common bacterial operational taxonomic units (OTUs, with mean relative abundance > 1%) were identified. These included Gammaproteobacteria (36 OTUs), Flavobacteriia (24), Alphaproteobacteria (18), and other taxa (11). When C. polykrikoides blooms intensified, the relative abundances of Gammaproteobacteria also increased. OTU #030 (Flavicella sp., Flavobacteria, 96%) was positively correlated with C. polykrikoides abundance (r = 0.77, p < 0.001). Functional analysis based on the dominant bacterial OTUs revealed that chemoheterotrophy-related functions were more common in high-severity sites of HABs than in other groups. Therefore, the occurrence of HABs highlighted their interactions with bacteria and affected the bacterial community structure and metabolic functions.

Monocular depth estimation network with single-pixel depth guidance.

Due to the scale ambiguity problem, the performance of monocular depth estimation (MDE) is inherently restricted. Multi-camera systems, especially those equipped with active depth cameras, have addressed this problem at the expense of increased hardware costs and space. In this Letter, we adopt a similar but cost-effective solution using only single-pixel depth guidance with a single-photon avalanche diode. To this end, we design a single-pixel guidance module (SPGM) that combines the global information from the single-pixel depth guidance with the spatial information from the image at the feature level. By integrating SPGMs into an MDE network, we introduce PhoMoNet, the first, to the best of our knowledge, end-to-end MDE network with single-pixel depth guidance. Experimental results show the effectiveness and superiority of PhoMoNet over state-of-the-art MDE networks on synthetic and real-world datasets.

Covariance of Marine Nucleocytoplasmic Large DNA Viruses with Eukaryotic Plankton Communities in the Sub-Arctic Kongsfjorden Ecosystem: A Metagenomic Analysis of Marine Microbial Ecosystems.

Nucleocytoplasmic large DNA viruses (NCLDVs) infect various marine eukaryotes. However, little is known about NCLDV diversity and their relationships with eukaryotic hosts in marine environments, the elucidation of which will advance the current understanding of marine ecosystems. This study characterizes the interplay between NCLDVs and the eukaryotic plankton community (EPC) in the sub-Arctic area using metagenomics and metabarcoding to investigate NCLDVs and EPC, respectively, in the Kongsfjorden ecosystem of Svalbard (Norway) in April and June 2018. Gyrodinium helveticum (Dinophyceae) is the most prevalent eukaryotic taxon in the EPC in April, during which time Mimiviridae (31.8%), Poxviridae (25.1%), Phycodnaviridae (14.7%) and Pandoraviridae (13.1%) predominate. However, in June, the predominant taxon is Aureococcus anophagefferens (Pelagophyceae), and the NCLDVs, Poxviridae (32.9%), Mimiviridae (29.1%), and Phycodnaviridae (18.5%) appear in higher proportions with an increase in Pelagophyceae, Bacillariophyceae, and Chlorophyta groups. Thus, differences in NCLDVs may be caused by changes in EPC composition in response to environmental changes, such as increases in water temperature and light intensity. Taken together, these findings are particularly relevant considering the anticipated impact of NCLDV-induced EPC control mechanisms on polar regions and, therefore, improve the understanding of the Sub-Arctic Kongsfjorden ecosystem.

Risk-aware survival time prediction from whole slide pathological images.

Deep-learning-based survival prediction can assist doctors by providing additional information for diagnosis by estimating the risk or time of death. The former focuses on ranking deaths among patients based on the Cox model, whereas the latter directly predicts the survival time of each patient. However, it is observed that survival time prediction for the patients, particularly with close observation times, possibly has incorrect orders, leading to low prediction accuracy. Therefore, in this paper, we present a whole slide image (WSI)-based survival time prediction method that takes advantage of both the risk as well as time prediction. Specifically, we propose to combine these two approaches by extracting the risk prediction features and using them as guides for the survival time prediction. Considering the high resolution of WSIs, we extract tumor patches from WSIs using a pre-trained tumor classifier and apply the graph convolutional network to aggregate information across these patches effectively. Extensive experiments demonstrate that the proposed method significantly improves the time prediction accuracy when compared with direct prediction of the survival times without guidance and outperforms existing methods.