Segmenting medical images with limited data.

While computer vision has proven valuable for medical image segmentation, its application faces challenges such as limited dataset sizes and the complexity of effectively leveraging unlabeled images. To address these challenges, we present a novel semi-supervised, consistency-based approach termed the data-efficient medical segmenter (DEMS). The DEMS features an encoder-decoder architecture and incorporates the developed online automatic augmenter (OAA) and residual robustness enhancement (RRE) blocks. The OAA augments input data with various image transformations, thereby diversifying the dataset to improve the generalization ability. The RRE enriches feature diversity and introduces perturbations to create varied inputs for different decoders, thereby providing enhanced variability. Moreover, we introduce a sensitive loss to further enhance consistency across different decoders and stabilize the training process. Extensive experimental results on both our own and three public datasets affirm the effectiveness of DEMS. Under extreme data shortage scenarios, our DEMS achieves 16.85% and 10.37% improvement in dice score compared with the U-Net and top-performed state-of-the-art method, respectively. Given its superior data efficiency, DEMS could present significant advancements in medical segmentation under small data regimes. The project homepage can be accessed at https://github.com/NUS-Tim/DEMS.

CECT: Controllable ensemble CNN and transformer for COVID-19 image classification.

The COVID-19 pandemic has resulted in hundreds of million cases and numerous deaths worldwide. Here, we develop a novel classification network CECT by controllable ensemble convolutional neural network and transformer to provide a timely and accurate COVID-19 diagnosis. The CECT is composed of a parallel convolutional encoder block, an aggregate transposed-convolutional decoder block, and a windowed attention classification block. Each block captures features at different scales from 28 × 28 to 224 × 224 from the input, composing enriched and comprehensive information. Different from existing methods, our CECT can capture features at both multi-local and global scales without any sophisticated module design. Moreover, the contribution of local features at different scales can be controlled with the proposed ensemble coefficients. We evaluate CECT on two public COVID-19 datasets and it reaches the highest accuracy of 98.1% in the intra-dataset evaluation, outperforming existing state-of-the-art methods. Moreover, the developed CECT achieves an accuracy of 90.9% on the unseen dataset in the inter-dataset evaluation, showing extraordinary generalization ability. With remarkable feature capture ability and generalization ability, we believe CECT can be extended to other medical scenarios as a powerful diagnosis tool. Code is available at https://github.com/NUS-Tim/CECT.

GSDA: Generative adversarial network-based semi-supervised data augmentation for ultrasound image classification.

Medical Ultrasound (US) is one of the most widely used imaging modalities in clinical practice, but its usage presents unique challenges such as variable imaging quality. Deep Learning (DL) models can serve as advanced medical US image analysis tools, but their performance is greatly limited by the scarcity of large datasets. To solve the common data shortage, we develop GSDA, a Generative Adversarial Network (GAN)-based semi-supervised data augmentation method. GSDA consists of the GAN and Convolutional Neural Network (CNN). The GAN synthesizes and pseudo-labels high-resolution, high-quality US images, and both real and synthesized images are then leveraged to train the CNN. To address the training challenges of both GAN and CNN with limited data, we employ transfer learning techniques during their training. We also introduce a novel evaluation standard that balances classification accuracy with computational time. We evaluate our method on the BUSI dataset and GSDA outperforms existing state-of-the-art methods. With the high-resolution and high-quality images synthesized, GSDA achieves a 97.9% accuracy using merely 780 images. Given these promising results, we believe that GSDA holds potential as an auxiliary tool for medical US analysis.

Recent progress in transformer-based medical image analysis.

The transformer is primarily used in the field of natural language processing. Recently, it has been adopted and shows promise in the computer vision (CV) field. Medical image analysis (MIA), as a critical branch of CV, also greatly benefits from this state-of-the-art technique. In this review, we first recap the core component of the transformer, the attention mechanism, and the detailed structures of the transformer. After that, we depict the recent progress of the transformer in the field of MIA. We organize the applications in a sequence of different tasks, including classification, segmentation, captioning, registration, detection, enhancement, localization, and synthesis. The mainstream classification and segmentation tasks are further divided into eleven medical image modalities. A large number of experiments studied in this review illustrate that the transformer-based method outperforms existing methods through comparisons with multiple evaluation metrics. Finally, we discuss the open challenges and future opportunities in this field. This task-modality review with the latest contents, detailed information, and comprehensive comparison may greatly benefit the broad MIA community.

Tumor necroptosis-mediated shedding of cell surface proteins promotes metastasis of breast cancer by suppressing anti-tumor immunity.

Necroptosis is a form of regulated necrosis and is executed by MLKL when MLKL is engaged in triggering the rupture of cell plasma membrane. MLKL activation also leads to the protease, ADAMs-mediated ectodomain shedding of cell surface proteins of necroptotic cells. Tumor necroptosis often happens in advanced solid tumors, and blocking necroptosis by MLKL deletion in breast cancer dramatically reduces tumor metastasis. It has been suggested that tumor necroptosis affects tumor progression through modulating the tumor microenvironment. However, the exact mechanism by which tumor necroptosis promotes tumor metastasis remains elusive. Here, we report that the ectodomain shedding of cell surface proteins of necroptotic cells is critical for the promoting effect of tumor necroptosis in tumor metastasis through inhibiting the anti-tumor activity of T cells. We found that blocking tumor necroptosis by MLKL deletion led to the dramatic reduction of tumor metastasis and significantly elevated anti-tumor activity of tumor-infiltrating and peripheral blood T cells. Importantly, the increased anti-tumor activity of T cells is a key cause for the reduced metastasis as the depletion of CD8+ T cells completely restored the level of metastasis in the Mlkl KO mice. Interestingly, the levels of some soluble cell surface proteins including sE-cadherin that are known to promote metastasis are also dramatically reduced in MLKL null tumors/mice. Administration of ADAMs pan inhibitor reduces the levels of soluble cell surface proteins in WT tumors/mice and leads to the dramatic decrease in metastasis. Finally, we showed the sE-cadherin/KLRG1 inhibitory receptor is the major pathway for necroptosis-mediated suppression of the anti-tumor activity of T cells and the promotion of metastasis. Hence, our study reveals a novel mechanism of tumor necroptosis-mediated promotion of metastasis and suggests that tumor necroptosis and necroptosis-activated ADAMs are potential targets for controlling metastasis.

Investigation on an outbreak of cutaneous anthrax in a county of Shandong Province, China, 2021.

BACKGROUND: In order to verify the existence of an anthrax outbreak, determine its scope, grasp the epidemiological characteristics and find out the cause of the outbreak and recommend preventive and control measures. METHODS: Etiological hypothesis was developed through descriptive epidemiological methods. Hypotheses were tested by analyzing epidemiological methods by comparing the differences in the incidence of different exposure types. Nucleic acid detection and bacterial isolation and culture in the BSL-2 laboratories. SPSS 21 was used to conduct statistical analysis. RESULTS: A total of 126 family, workshop, shop environment samples and meat samples were collected, and 6 samples were collected from skin lesions of suspected cutaneous anthrax cases. 41 samples were positive by rPCR and 8 strains of Bacillus anthracis were cultivated. Participated in slaughtering, cutting beef of sick cattles was significantly associated with cutaneous anthrax (RR 3.75, 95% CI 1.08-13.07), this behavior is extremely dangerous. CONCLUSIONS: Comprehensive analysis of laboratory results and epidemiological survey results and environmental assessments, we judge this epidemic to be an outbreak of cutaneous anthrax, associated with slaughtering and other processes from infected cattle imported from other province.

The Prediction of Enterprise Stock Change Trend by Deep Neural Network Model.

This study aims to accurately predict the changing trend of stocks in stock trading so that company investors can obtain higher returns. In building a financial forecasting model, historical data and learned parameters are used to predict future stock prices. Firstly, the relevant theories of stock forecasting are discussed, and problems in stock forecasting are raised. Secondly, the inadequacies of deep neural network (DNN) models are discussed. A prediction trend model of enterprise stock is established based on long short-term memory (LSTM). The uniqueness and innovation lie in using the stock returns of Bank of China securities in 2022 as the training data set. LSTM prediction models are used to perform error analysis on company data training. The 20-day change trend of the company's stock returns under different models is predicted and analyzed. The results show that as the number of iterations increases, the loss rate of the LSTM training curve keeps decreasing until 0. The average return price of the LSTM prediction model is 14.01. This figure is closest to the average real return price of 13.89. Through the forecast trend analysis under different models, LSTM predicts that the stock change trend of the enterprise model is closest to the changing trend of the actual earnings price. The prediction accuracy is better than other prediction models. In addition, this study explores the characteristics of high noise and complexity of corporate stock time series, designs a DNN prediction model, and verifies the feasibility of the LSTM model to predict corporate stock changes with high accuracy.

ZBP1 not RIPK1 mediates tumor necroptosis in breast cancer.

Tumor necrosis happens commonly in advanced solid tumors. We reported that necroptosis plays a major role in tumor necrosis. Although several key necroptosis regulators including receptor interacting protein kinase 1 (RIPK1) have been identified, the regulation of tumor necroptosis during tumor development remains elusive. Here, we report that Z-DNA-binding protein 1 (ZBP1), not RIPK1, mediates tumor necroptosis during tumor development in preclinical cancer models. We found that ZBP1 expression is dramatically elevated in necrotic tumors. Importantly, ZBP1, not RIPK1, deletion blocks tumor necroptosis during tumor development and inhibits metastasis. We showed that glucose deprivation triggers ZBP1-depedent necroptosis in tumor cells. Glucose deprivation causes mitochondrial DNA (mtDNA) release to the cytoplasm and the binding of mtDNA to ZBP1 to activate MLKL in a BCL-2 family protein, NOXA-dependent manner. Therefore, our study reveals ZBP1 as the key regulator of tumor necroptosis and provides a potential drug target for controlling tumor metastasis.

G3BP1 Inhibition Alleviates Intracellular Nucleic Acid-Induced Autoimmune Responses.

The detection of intracellular nucleic acids is a fundamental mechanism of host defense against infections. The dysregulated nucleic acid sensing, however, is a major cause for a number of autoimmune diseases. In this study, we report that GTPase-activating protein SH3 domain-binding protein 1 (G3BP1) is critical for both intracellular DNA- and RNA-induced immune responses. We found that in both human and mouse cells, the deletion of G3BP1 led to the dampened cGAS activation by DNA and the insufficient binding of RNA by RIG-I. We further found that resveratrol (RSVL), a natural compound found in grape skin, suppressed both intracellular DNA- and RNA-induced type I IFN production through inhibiting G3BP1. Importantly, using experimental mouse models for Aicardi-Goutières syndrome, an autoimmune disorder found in humans, we demonstrated that RSVL effectively alleviated intracellular nucleic acid-stimulated autoimmune responses. Thus, our study demonstrated a broader role of G3BP1 in sensing different kinds of intracellular nucleic acids and presented RSVL as a potential treatment for autoimmune conditions caused by dysregulated nucleic acid sensing.

Acetylation Blocks cGAS Activity and Inhibits Self-DNA-Induced Autoimmunity.

The presence of DNA in the cytoplasm is normally a sign of microbial infections and is quickly detected by cyclic GMP-AMP synthase (cGAS) to elicit anti-infection immune responses. However, chronic activation of cGAS by self-DNA leads to severe autoimmune diseases for which no effective treatment is available yet. Here we report that acetylation inhibits cGAS activation and that the enforced acetylation of cGAS by aspirin robustly suppresses self-DNA-induced autoimmunity. We find that cGAS acetylation on either Lys384, Lys394, or Lys414 contributes to keeping cGAS inactive. cGAS is deacetylated in response to DNA challenges. Importantly, we show that aspirin can directly acetylate cGAS and efficiently inhibit cGAS-mediated immune responses. Finally, we demonstrate that aspirin can effectively suppress self-DNA-induced autoimmunity in Aicardi-Goutières syndrome (AGS) patient cells and in an AGS mouse model. Thus, our study reveals that acetylation contributes to cGAS activity regulation and provides a potential therapy for treating DNA-mediated autoimmune diseases.

Porcine reproductive and respiratory syndrome virus NADC30-like strain accelerates Streptococcus suis serotype 2 infection in vivo and in vitro.

Porcine reproductive and respiratory syndrome (PRRS), an economically significant pandemic disease, commonly results in increased impact of bacterial infections, including those by Streptococcus suis (S. suis). In recent years, PRRS virus (PRRSV) NADC30-like strain has emerged in different regions of China, and coinfected with S. suis and PRRSV has also gradually increased in clinical performance. However, the mechanisms involved in host innate responses towards S. suis and their implications of coinfection with NADC30-like strain remain unknown. Therefore, the pathogenicity of NADC30-like strain and S. suis serotype 2 (SS2) coinfection in vivo and in vitro was investigated in this study. The results showed that NADC30-like increased the invasion and proliferation of SS2 in blood and tissues, resulting in more severe pneumonia, myocarditis, and peritonitisas well as higher mortality rate in pigs. In vitro, NADC30-like strain increased the invasion and survival of SS2 in porcine alveolar macrophages (PAM) cells, causing more drastic expression of inflammatory cytokines and activation of NF-ĸB signalling. These results pave the way for understanding the interaction of S. suis with the swine immune system and their modulation in a viral coinfection.

G3BP1 promotes DNA binding and activation of cGAS.

Cyclic GMP-AMP synthase (cGAS) is a key sensor responsible for cytosolic DNA detection. Here we report that GTPase-activating protein SH3 domain-binding protein 1 (G3BP1) is critical for DNA sensing and efficient activation of cGAS. G3BP1 enhanced DNA binding of cGAS by promoting the formation of large cGAS complexes. G3BP1 deficiency led to inefficient DNA binding by cGAS and inhibited cGAS-dependent interferon (IFN) production. The G3BP1 inhibitor epigallocatechin gallate (EGCG) disrupted existing G3BP1-cGAS complexes and inhibited DNA-triggered cGAS activation, thereby blocking DNA-induced IFN production both in vivo and in vitro. EGCG administration blunted self DNA-induced autoinflammatory responses in an Aicardi-Goutières syndrome (AGS) mouse model and reduced IFN-stimulated gene expression in cells from a patient with AGS. Thus, our study reveals that G3BP1 physically interacts with and primes cGAS for efficient activation. Furthermore, EGCG-mediated inhibition of G3BP1 provides a potential treatment for cGAS-related autoimmune diseases.

RINCK-mediated monoubiquitination of cGAS promotes antiviral innate immune responses.

BACKGROUND: As an important danger signal, the presence of DNA in cytoplasm triggers potent immune responses. Cyclic GMP-AMP synthase (cGAS) is a recently characterized key sensor for cytoplasmic DNA. The engagement of cGAS with DNA leads to the synthesis of a second messenger, cyclic GMP-AMP (cGAMP), which binds and activates the downstream adaptor protein STING to promote type I interferon production. Although cGAS has been shown to play a pivotal role in innate immunity, the exact regulation of cGAS activation is not fully understood. RESULTS: We report that an E3 ubiquitin ligase, RING finger protein that interacts with C kinase (RINCK, also known as tripartite motif protein 41, TRIM41), is critical for cGAS activation by mediating the monoubiquitination of cGAS. Using CRISPR/Cas9, we generated RINCK-deletion cells and showed that the deficiency of RINCK resulted in dampened interferon production in response to cytosolic DNA. Consistently, the RINCK-deletion cells also exhibited insufficient interferon production upon herpes simplex virus 1, a DNA virus, infection. As a result, the viral load in RINCK-deficient cells was significantly higher than that in wild-type cells. We also found that RINCK deficiency inhibited the up-stream signaling of DNA-triggered interferon production pathway, which was reflected by the phosphorylation of the TANK-binding kinase 1 and the interferon regulatory factor 3. Interestingly, we found that RINCK binds to cGAS and promotes the monoubiquitination of cGAS, thereby positively regulating the cGAS-mediated cGAMP synthesis. CONCLUSIONS: Our study reveals that monoubiquitination is an important regulation for cGAS activation and uncovers a critical role of RINCK in the cGAS-mediated innate immunity.

NLRP3 Phosphorylation Is an Essential Priming Event for Inflammasome Activation.

Many infections and stress signals can rapidly activate the NLRP3 inflammasome to elicit robust inflammatory responses. This activation requires a priming step, which is thought to be mainly for upregulating NLRP3 transcription. However, recent studies report that the NLRP3 inflammasome can be activated independently of transcription, suggesting that the priming process has unknown essential regulatory steps. Here, we report that JNK1-mediated NLRP3 phosphorylation at S194 is a critical priming event and is essential for NLRP3 inflammasome activation. We show that NLRP3 inflammasome activation is disrupted in NLRP3-S194A knockin mice. JNK1-mediated NLRP3 S194 phosphorylation is critical for NLRP3 deubiquitination and facilitates its self-association and the subsequent inflammasome assembly. Importantly, we demonstrate that blocking S194 phosphorylation prevents NLRP3 inflammasome activation in cryopyrin-associated periodic syndromes (CAPS). Thus, our study reveals a key priming molecular event that is a prerequisite for NLRP3 inflammasome activation. Inhibiting NLRP3 phosphorylation could be an effective treatment for NLRP3-related diseases.

SUMOylated NKAP is essential for chromosome alignment by anchoring CENP-E to kinetochores.

Chromosome alignment is required for accurate chromosome segregation. Chromosome misalignment can result in genomic instability and tumorigenesis. Here, we show that NF-κB activating protein (NKAP) is critical for chromosome alignment through anchoring CENP-E to kinetochores. NKAP knockdown causes chromosome misalignment and prometaphase arrest in human cells. NKAP dynamically localizes to kinetochores, and is required for CENP-E kinetochore localization. NKAP is SUMOylated predominantly in mitosis and the SUMOylation is needed for NKAP to bind CENP-E. A SUMOylation-deficient mutant of NKAP cannot support the localization of CENP-E on kinetochores or proper chromosome alignment. Moreover, Bub3 recruits NKAP to stabilize the binding of CENP-E to BubR1 at kinetochores. Importantly, loss of NKAP expression causes aneuploidy in cultured cells, and is observed in human soft tissue sarcomas. These findings indicate that NKAP is a novel and key regulator of mitosis, and its dysregulation might contribute to tumorigenesis by causing chromosomal instability.

Dysregulation of the miR-324-5p-CUEDC2 axis leads to macrophage dysfunction and is associated with colon cancer.

CUEDC2, a CUE-domain-containing protein, modulates inflammation, but its involvement in tumorigenesis is still poorly understood. Here, we report that CUEDC2 is a key regulator of macrophage function and critical for protection against colitis-associated tumorigenesis. CUEDC2 expression is dramatically upregulated during macrophage differentiation, and CUEDC2 deficiency results in excessive production of proinflammatory cytokines. The level of CUEDC2 in macrophages is modulated by miR- 324-5p. We find that Cuedc2 KO mice are more susceptible to dextran-sodium-sulfate-induced colitis, and macrophage transplantation results suggest that the increased susceptibility results from the dysfunction of macrophages lacking CUEDC2. Furthermore, we find that Cuedc2 KO mice are more prone to colitis-associated cancer. Importantly, CUEDC2 expression is almost undetectable in macrophages in human colon cancer, and this decreased CUEDC2 expression is associated with high levels of interleukin-4 and miR-324-5p. Thus, CUEDC2 plays a crucial role in modulating macrophage function and is associated with both colitis and colon tumorigenesis.

Glutathione peroxidase-1 is required for self-renewal of murine embryonic stem cells.

Embryonic stem (ES) cells are pluripotent cells that are capable of giving rise to any type of cells in the body and possess unlimited self-renewal potential. However, the exact regulatory mechanisms that govern the self-renewal ability of ES cells remain elusive. To understand the immediate early events during ES cell differentiation, we performed a proteomics study and analyzed the proteomic difference in murine ES cells before and after a 6-h spontaneous differentiation. We found that the expression level of glutathione peroxidase-1 (GPx-1), an antioxidant enzyme, is dramatically decreased upon the differentiation. Both knockdown of GPx-1 expression with shRNA and inhibiting GPx-1 activity by inhibitor led to the differentiation of ES cells. Furthermore, we showed that during early differentiation, the quick degradation of GPx-1 was mediated by proteasome. Thus, our data indicated that GPx-1 is a key regulator of self-renewal of murine embryonic stem cells.

Whole-Genome Sequencing of Lactobacillus shenzhenensis Strain LY-73T.

Lactobacillus shenzhenensis strain LY-73(T) is a novel species which was first isolated from fermented goods. Here, we report the draft genome sequence of Lactobacillus shenzhenensis LY-73(T).