DEK regulates B-cell proliferative capacity and is associated with aggressive disease in low-grade B-cell lymphomas.

This study sheds light on the pivotal role of the oncoprotein DEK in B-cell lymphoma. We reveal DEK expression correlates with increased tumor proliferation and inferior overall survival in cases diagnosed with low-grade B-cell lymphoma (LGBCL). We also found significant correlation between DEK expression and copy number alterations in LGBCL tumors, highlighting a novel mechanism of LGBCL pathogenesis that warrants additional exploration. To interrogate the mechanistic role of DEK in B-cell lymphoma, we generated a DEK knockout cell line model, which demonstrated DEK depletion caused reduced proliferation and altered expression of key cell cycle and apoptosis-related proteins, including Bcl-2, Bcl-xL, and p53. Notably, DEK depleted cells showed increased sensitivity to apoptosis-inducing agents, including venetoclax and staurosporine, which underscores the therapeutic potential of targeting DEK in B-cell lymphomas. Overall, our study contributes to a better understanding of DEK's role as an oncoprotein in B-cell lymphomas, highlighting its potential as both a promising therapeutic target and a novel biomarker for aggressive LGBCL. Further research elucidating the molecular mechanisms underlying DEK-mediated tumorigenesis could pave the way for improved treatment strategies and better clinical outcomes for patients with B-cell lymphoma.

Monocyte response to SARS-CoV-2 protein ORF8 is associated with severe COVID-19 infection in patients with chronic lymphocytic leukemia.

The open reading frame 8 (ORF8) protein, encoded by the SARS-CoV-2 virus after infection, stimulates monocytes/macrophages to produce pro-inflammatory cytokines. We hypothesized that a positive ex vivo monocyte response to ORF8 protein pre-COVID-19 would be associated with subsequent severe Coronavirus disease 2019 (COVID-19). We tested ORF8 ex vivo on peripheral blood mononuclear cells from 26 anonymous healthy blood donors and measured intracellular cytokine/ chemokine levels in monocytes by flow cytometry. The percentage of positive monocyte staining in the sample and change in mean fluorescence intensity (ΔMFI) after ORF8 were used to calculate the adjusted MFI for each cytokine. We then tested pre-COVID-19 peripheral blood mononuclear cell samples from 60 chronic lymphocytic leukemia (CLL) patients who subsequently developed COVID-19 infection. Severe COVID-19 was defined as hospitalization due to COVID-19. In the 26 normal donor samples, the adjusted MFI for interleukin (IL)-1ß, IL-6, IL-8, and CCL-2 were significantly different with ORF8 stimulation versus controls. We next analyzed monocytes from pre-COVID-19 PBMC samples from 60 CLL patients. The adjusted MFI to ORF8 stimulation of monocyte intracellular IL-1ß was associated with severe COVID-19 and a reactive ORF8 monocyte response was defined as an IL-1ß adjusted MFI ≥0.18 (sensitivity 67%, specificity 75%). The median time to hospitalization after infection in CLL patients with a reactive ORF8 response was 12 days versus not reached for patients with a non-reactive ORF8 response with a hazard ratio of 7.7 (95% confidence interval: 2.4-132; P=0.005). These results provide new insight on the monocyte inflammatory response to virus with implications in a broad range of disorders involving monocytes.

Long-term outcomes of endoscopic therapy versus surgical resection for 2-5 cm gastric gastrointestinal stromal tumors: A population-based comparative study.

BACKGROUND: Endoscopic therapy (ET) of gastrointestinal stromal tumors (GIST) has become a viable treatment. We intended to compare long-term outcomes of ET versus surgical resection for 2-5 cm GIST using the Surveillance, Epidemiology, and End Results (SEER) database. METHODS: A multicenter retrospective study was conducted to compare the long-term outcomes of patients treated with ET and surgical resection for GIST. The multivariate Cox proportional hazards models were used to identify predictors for patients survival. To balance the clinicopathologic characteristics, a 1:1 propensity score matching (PSM) was utilized. RESULTS: A total of 749 patients with 2-5 cm GIST were enrolled, of whom 113 accepted ET and 636 underwent surgical resection. Before PSM, there was no significant difference in long-term outcomes between ET and surgical resection (5-year overall survival (OS): 93.5% vs. 91.6%, P=0.374; 5-year cancer-specific survival (CSS): 99.1% vs. 96.5%, P=0.546; 10-year OS: 71.1% vs. 78.2%, P=0.374; 10-year CSS: 93.6% vs. 92.7%, P=0.546). After adjusting for the relevant variables using the multivariable Cox proportional hazards models, we observed that the ET and surgical resection groups were similar in OS (HR 0.726, 95%CI 0.457-1.153, P=0.175) and CSS (HR 1.286, 95%CI 0.474-3.488, P=0.621). After PSM, the long-term OS and CSS of patients with 2-5 cm GIST after ET and surgical resection were comparable. CONCLUSIONS: We found that the long-term survival of patients with 2-5 cm gastric GIST after ET and surgical resection were comparable. Further high-quality studies are needed to confirm the role of ET in 2-5 cm GIST.

Associations of the triglyceride-glucose index and remnant cholesterol with coronary artery disease: a retrospective study.

BACKGROUND: Remnant cholesterol (RC) represents a low-cost and readily measured lipid index that contributes significantly to residual cardiovascular disease risk. The triglyceride-glucose (TyG) index exhibits a significant correlation with cardiovascular disease occurrence. However, RC and the TyG index have rarely been examined for their potentials in predicting coronary artery disease (CAD). Accordingly, the study was designed to validate the correlations of these two biomarkers with CAD and to compare the forecasted values of these two biomarkers for newly diagnosed CAD. METHODS: Totally 570 subjects firstly administered coronary angiography were enrolled, including 431 newly diagnosed CAD cases and 139 individuals without CAD. The individuals were classified into two groups according to CAD diagnosis. RC was derived as total cholesterol content (mmol/L) - (high density lipoprotein cholesterol content + low density lipoprotein cholesterol content; both in mmol/L). The TyG index was determined as ln (fasting triglyceride level [mg/dL] × fasting plasma glucose level [mg/dL])/2. RESULTS: Baseline feature analysis revealed significant differences in RC and the TyG index between the CAD and non-CAD groups (both P < 0.001). RC and the TyG index were independent risk factors for CAD in accordance with logistic regression analysis (both P < 0.05). Moreover, spearman correlation analysis elucidated CAD had a more remarkable correlation with the TyG index compared with RC (both P < 0.001). Furthermore, according to receiver operating characteristic curve analysis, the TyG index was better than RC in predicting CAD. CONCLUSIONS: The TyG index and RC have significant associations with CAD. Compared with RC, the TyG index possesses a closer correlation with CAD and a higher predictive value for CAD.

Associations of the triglyceride-glucose index and atherogenic index of plasma with the severity of new-onset coronary artery disease in different glucose metabolic states.

BACKGROUND: The triglyceride-glucose (TyG) index is considered a dependable biomarker for gauging insulin resistance. The atherogenic index of plasma (AIP) represents a marker reflecting atherosclerosis. However, there is currently no study specifically exploring the associations of these two biomarkers with the severity of new-onset coronary artery disease (CAD) under different glucose metabolic states. Therefore, this study aims to evaluate the correlations of these two biomarkers with CAD severity in patients newly diagnosed with CAD under various glucose metabolism conditions. METHOD: Totally 570 subjects first administered coronary angiography were enrolled, including 431 first diagnosed CAD patients and 139 non-CAD patients. CAD severity  was gauged by the quantity of narrowed arteries (single-vessel and multi-vessel CAD). According to WHO diabetes guidelines, glucose metabolic states were divided into normal glucose regulation (NGR), pre-diabetes mellitus (Pre-DM), and diabetes mellitus (DM). The relationships of the TyG index and AIP with CAD severity were validated by logistic regression analysis, including adjustment for traditional cardiovascular risk elements and medical treatments. Their predictive efficacy for CAD was evaluated by receiver operating characteristic (ROC) curves. RESULT: The TyG index and AIP were independently correlated with CAD in accordance with logistic regression analysis (both P < 0.05). Regardless of the glucose metabolic states, there was no statistical correlation between the TyG index and CAD severity. However, AIP in NGR patients was significantly related to CAD severity (P < 0.05). The areas under the curve of the TyG index and AIP for predicting CAD were 0.682 and 0.642 (both P < 0.001), respectively, and their optimal cut-off values were 3.210 (Youden index: 0.305) and 0.095 (Youden index:0.246), respectively. CONCLUSION: The TyG index and AIP have significant associations with CAD. The TyG index had no association with CAD severity, regardless of glucose metabolic states. AIP exhibited a discernible link with CAD severity in NGR patients, but not in the pre-DM or DM populations. The TyG index and AIP have similar predictive values for new-onset CAD.

BST2 promotes gastric cancer metastasis under the regulation of HOXD9 and PABPC1.

Gastric cancer (GC) constitutes substantial cancer mortality worldwide. Several cancer types aberrantly express bone marrow stromal cell antigen 2 (BST2), yet its functional and underlying mechanisms in GC progression remain unknown. In our study, RNA sequencing data revealed that BST2 was transcriptionally activated by homeobox D9 (HOXD9). BST2 was significantly upregulated in GC tissues and promoted epithelial-mesenchymal transition and metastasis of GC. BST2 knockdown reversed HOXD9's oncogenic effect on GC metastasis. Moreover, BST2 messenger RNA stability could be enhanced by poly(A) binding protein cytoplasmic 1 (PABPC1) through the interaction between BST2 3'-UTR and PABPC1 in GC cells. PABPC1 promoted GC metastasis, which BST2 silencing attenuated in vitro and in vivo. In addition, positive correlations among HOXD9, BST2, and PABPC1 were established in clinical samples. Taken together, increased expression of BST2 induced by HOXD9 synergizing with PABPC1 promoted GC cell migration and invasion capacity.

Feedthrough effect in MEMS gyroscopes and fully differential feedthrough cancellation method.

In the operation of MEMS gyroscopes, the feedthrough signal is inevitably introduced through the feedthrough capacitance and significantly affects the performance of MEMS gyroscopes. The common feedthrough cancellation methods highly depend on the microfabrication process and the structural symmetry of MEMS gyroscopes, which cannot effectively eliminate the feedthrough signal. In order to solve this problem, a new feedthrough cancellation method based on a fully differential configuration is proposed in this paper. The influence of the feedthrough effect is analyzed by establishing an equivalent circuit model including the feedthrough capacitance and described by the admittance diagram. Different drive and detection configurations are compared for the problem of feedthrough mismatch on different paths. Based on the theoretical analysis, a fully differential feedthrough cancellation scheme combined with two inverse feedthrough cancellation circuits is proposed. The deviation of feedthrough signal caused by the asymmetry of two differential paths can be adjusted by the two inverse feedthrough cancellation circuits. The fully differential configuration can further increase the drive force and suppress the common mode errors. The experimental results indicate that the feedthrough signal is suppressed in a large degree by the designed feedthrough cancellation method. The feedthrough level is reduced by 50.53 dB, and the amplitude of the effective signal increases from 4.10 to 9.46 dB. The signal-to-noise ratio has an improvement of 212.48% than that before feedthrough cancellation. The proposed feedthrough cancellation method can significantly reduce the interference of feedthrough signals on the effective signal, effectively improving the signal quality of MEMS gyroscopes.

Salvage therapy expands highly cytotoxic and metabolically fit resilient CD8+ T cells via ME1 up-regulation.

Patients with advanced cancers who either do not experience initial response to or progress while on immune checkpoint inhibitors (ICIs) receive salvage radiotherapy to reduce tumor burden and tumor-related symptoms. Occasionally, some patients experience substantial global tumor regression with a rebound of cytotoxic CD8+ T cells. We have termed the rebound of cytotoxic CD8+ T cells in response to salvage therapy as T cell resilience and examined the underlying mechanisms of resilience. Resilient T cells are enriched for CX3CR1+ CD8+ T cells with low mitochondrial membrane potential, accumulate less reactive oxygen species (ROS), and express more malic enzyme 1 (ME1). ME1 overexpression enhanced the cytotoxicity and expansion of effector CD8+ T cells partially via the type I interferon pathway. ME1 also increased mitochondrial respiration while maintaining the redox state balance. ME1 increased the cytotoxicity of peripheral lymphocytes from patients with advanced cancers. Thus, preserved resilient T cells in patients rebound after salvage therapy and ME1 enhances their resiliency.

The ATF2/miR-3913-5p/CREB5 axis is involved in the cell proliferation and metastasis of colorectal cancer.

Various miRNAs have been shown to participate in the tumor progression and development of colorectal cancer (CRC). However, the role of miR-3913-5p in CRC are yet to be clearly defined. In the present study, we determine that miR-3913-5p is downregulated in CRC cell lines and CRC tissues. Exogenous miR-3913-5p expression weakens the CRC cells growth, migration and invasion. Mechanistically, miR-3913-5p directly targets the 3'UTR of CREB5. Overexpression of CREB5 reverses the suppression of CRC cells proliferation, migration and invasion induced by miR-3913-5p. Furthermore, ATF2 negatively regulates the transcription of miR-3913-5p by binding to its promoter. CREB5 can cooperate with ATF2. CREB5 is required for ATF2 in regulating miR-3913-5p. Finally, inverse correlations can be found between the expressions of miR-3913-5p and CREB5 or ATF2 in CRC tissues. Thus, a plausible mechanism of ATF2/miR-3913-5p/CREB5 axis regulating CRC progression is elucidated. Our findings suggest that miR-3913-5p functions as a tumor suppressor in CRC. ATF2/miR-3913-5p/CREB5 axis might be a potential therapeutic target against CRC progression.

The VAX2-LINC01189-hnRNPF signaling axis regulates cell invasion and migration in gastric cancer.

Transcription factors (TFs) and long noncoding RNAs (lncRNAs) contribute to gastric cancer (GC). However, the roles of TFs and lncRNAs in the invasion and metastasis of GC remain largely unknown. Here, we observed that the transcription factor VAX2 is significantly upregulated in GC cells and tissues and acts as an oncogene. Moreover, high VAX2 expression is associated with the advancement of tumors in GC. In terms of functionality, the enforced expression of VAX2 promotes the proliferation and metastasis of GC cells. Mechanistically, VAX2 specifically interacts with the LINC01189 promoter and represses LINC01189 transcription. Furthermore, LINC01189 exhibits significant downregulation in GC and functions as a suppressor gene. Functionally, it inhibits migratory and invasive abilities in GC cells. In the context of GC metastasis, VAX2 plays a role in modulating it by trans-repressing the expression of LINC01189. Additionally, LINC01189 binds to hnRNPF to enhance hnRNPF degradation through ubiquitination. The cooperation between LINC01189 and hnRNPF regulates GC cell invasion and migration. In addition, both VAX2 and hnRNPF are highly expressed, while LINC01189 is expressed in at low levels in GC tissues compared to normal gastric tissues. Our study suggests that VAX2 expression facilitates, while LINC01189 expression suppresses, metastasis and that the VAX2-LINC01189-hnRNPF axis plays a contributory role in GC development.

[Effect of Fire-deposited Charcoal on Soil Organic Carbon Pools and Associated Enzyme Activities in a Recently Harvested Pinus massoniana Plantation Subjected to Broadcast Burning].

Charcoal is a carbonaceous particulate matter with a highly aromatic structure produced by incomplete combustion, and it can cause persistent long-term effects on soil ecological functions. In this study, we determined soil organic carbon pools and associated enzyme activities following five years of different charcoal treatments[charcoal removal (B0), charcoal retained in situ (B1), and the addition of charcoal removed from B0(B2)] and the unburnt control (UB) in a recently harvested Pinus massoniana plantation subjected to broadcast burning. The results showed that dissolved organic carbon (DOC), microbial biomass carbon (MBC), coarse and fine particulate organic carbon (CPOC and FPOC), and recalcitrant carbon (RC) contents were significantly lower in B1 than those in UB soil (P<0.05). The MBC and FPOC contents of B2 soil were comparable to those of UB soil, which were significantly higher than those of B0 soil (P<0.001). There was no difference in MBC/TC between the B2 and UB soils, whereas MBC/TC was significantly lower in B0 than in UB soil (P<0.05). ß-glucosidase and peroxidase activities of B0, B1, and B2 soils were significantly lower than that of UB soil (P<0.01), and polyphenol oxidase activity was significantly lower in B0 and B2 soils than in UB soil (P<0.01). No significant difference in soil TC, DOC, readily oxidized carbon (ROC), CPOC, and RC content as well as associated enzyme activities was observed among the charcoal treatments (P>0.05). Redundancy analysis showed that sucrose and polyphenol oxidase were the key drivers influencing soil organic carbon fractions, accounting for 16.3% and 12.7% of the total variance, respectively. Overall, our findings indicated that fire-deposited charcoal played a positive role in enhancing soil microbial biomass carbon recovery, soil organic carbon accumulation, and stability, highlighting the importance of charcoal in the management of subtropical plantations in the future.

Secreted ORF8 induces monocytic pro-inflammatory cytokines through NLRP3 pathways in patients with severe COVID-19.

Despite extensive research, the specific factor associated with SARS-CoV-2 infection that mediates the life-threatening inflammatory cytokine response in patients with severe COVID-19 remains unidentified. Herein we demonstrate that the virus-encoded Open Reading Frame 8 (ORF8) protein is abundantly secreted as a glycoprotein in vitro and in symptomatic patients with COVID-19. ORF8 specifically binds to the NOD-like receptor family pyrin domain-containing 3 (NLRP3) in CD14+ monocytes to induce inflammasomal cytokine/chemokine responses including IL1ß, IL8, and CCL2. Levels of ORF8 protein in the blood correlate with severity and disease-specific mortality in patients with acute SARS-CoV-2 infection. Furthermore, the ORF8-induced inflammasome response was readily inhibited by the NLRP3 inhibitor MCC950 in vitro. Our study identifies a dominant cause of pathogenesis, its underlying mechanism, and a potential new treatment strategy for severe COVID-19.

Injection of Viscous Micro-Droplet via Nozzle-Driven Piezoelectric Micro-Jet and Its Performance Control Method.

The inkjet printing technology based on piezoelectric micro-jets can effectively realize the efficient and high-precision processing of special-shaped structures. In this work, a nozzle-driven piezoelectric micro-jet device is proposed, and its structure and micro-jet process are described. ANSYS two-phase, two-way fluid-structure coupling simulation analysis is carried out, and the mechanism of the piezoelectric micro-jet is described in detail. The effects of voltage amplitude, input signal frequency, nozzle diameter and oil viscosity on the injection performance of the proposed device are studied, and a set of effective control methods is summarized. The correctness of the piezoelectric micro-jet mechanism and the feasibility of the proposed nozzle-driven piezoelectric micro-jet device are proved by experiments, and an injection performance test is carried out. The experimental results are consistent with the ANSYS simulation results, which confirms the correctness of the experiment. Finally, the stability and superiority of the proposed device are verified via comparation experiments.

The HOXD9-mediated PAXIP1-AS1 regulates gastric cancer progression through PABPC1/PAK1 modulation.

Long non-coding RNAs (lncRNAs) have been functionally characterised in various diseases. LncRNA PAX-interacting protein 1-antisense RNA 1 (PAXIP1-AS1) has reportedly been associated with cancer development. However, its role in gastric cancer (GC) remains poorly understood. Here, we showed that PAXIP1-AS1 was transcriptionally repressed by homeobox D9 (HOXD9) and was significantly downregulated in GC tissues and cells. Decreased expression of PAXIP1-AS1 was positively correlated with tumour progression, while PAXIP1-AS1 overexpression inhibited cell growth and metastasis both in vitro and in vivo. PAXIP1-AS1 overexpression significantly attenuated HOXD9-enhanced epithelial-to-mesenchymal transition (EMT), invasion and metastasis in GC cells. Poly(A)-binding protein cytoplasmic 1 (PABPC1), an RNA-binding protein, was found to enhance the stability of PAK1 mRNA, leading to EMT progress and GC metastasis. PAXIP1-AS1 was found to directly bind to and destabilise PABPC1, thereby regulating EMT and metastasis of GC cells. In summary, PAXIP1-AS1 suppressed metastasis, and the HOXD9/PAXIP1-AS1/PABPC1/PAK1 signalling axis may be involved in the progression of GC.

The LINC00501-HSP90B1-STAT3 positive feedback loop promotes malignant behavior in gastric cancer cells.

Long non-coding RNAs (lncRNAs) have been implicated in gastric cancer (GC) carcinogenesis and progression. However, the role of LINC00501 in GC growth and metastasis remains unclear. In this study, we found that LINC00501 was frequently upregulated in GC cells and tissues and was closely related to adverse GC clinicopathological features. Aberrant overexpression of LINC00501 promoted GC cell proliferation, invasion, and metastasis both in vitro and in vivo. Mechanistically, LINC00501 stabilized client protein STAT3 from deubiquitylation by directly interacting with cancer chaperone protein HSP90B1. Furthermore, the LINC00501-STAT3 axis modulated GC cell proliferation and metastasis. In turn, STAT3 bound directly to the LINC00501 promoter and positively activated LINC00501 expression, thus forming a positive feedback loop, thereby accelerating tumor growth, invasiveness, and metastasis. In addition, LINC00501 expression was positively correlated with STAT3 and p-STAT3 protein expression levels in gastric clinical samples. Our results reveal that LINC00501 acts as an oncogenic lncRNA and that the LINC00501-HSP90B1-STAT3 positive feedback loop contributes to GC development and progression, suggesting that LINC00501 may be a novel potential biomarker and treatment target for GC.

Ubiquitin-proteasome system reveals clinical subtypes with distinct molecular characteristics and prognoses in gastric cancer.

Modulators in ubiquitin-proteasome system (UPS) has been implicated in regulating cancer-related genes, immune responses, and oncogenesis. However, the global UPS expression pattern and its role in gastric cancer (GC) pathology remain elusive. Herein, we integrated the modulators in UPS and dissected their associations with tumor microenvironment (TME), therapeutic response and prognosis in GC. Ten eligible GC cohorts (n = 2161) were collected in this comprehensive analysis. Unsupervised clustering based on expression profile of ubiquitination regulators was performed to identify distinct expression pattern. Then, pathway activation, and TME characteristics and prognosis were explored for patients in each pattern. Finally, a UPS scoring system in GC, termed UPSGC, is developed for individualized quantification of UPS expression pattern. Two prognosis-distinctive UPS expression patterns were identified and validated. Multiple interdependent characteristics were found in each pattern. Patients in the pattern with poor prognosis were found with activation of EMT, TNFα/NF-κB and IL6/JAK/STAT3 signaling, and more infiltration of immunosuppressive M2 macrophages and Th2 cells in TME. And another pattern was characterized by upregulation of angiogenesis, Notch and Wnt-ß/catenin signaling, as well as enrichment of microvessels in TME. Based on the UPSGC system, two pattern-related clinical subtypes were identified. Finally, the UPSGC subtypes were validated as robust biomarker to predict patient's therapeutic responses and survival outcomes. In conclusion, this study proposes two previously unexplored UPS expression patterns in GC, in which patients have distinct survival outcomes and molecular characteristics. The findings provide new evidences to support the clinical relevance of ubiquitination with personalized therapy.

Investigating the Deterioration of Pavement Skid Resistance Using an Accelerated Pavement Test.

Stone matrix asphalt (SMA) mixture has been widely used in pavement engineering for its preferable in-service performance. However, deterioration of SMA pavement in skid resistance is apparent under traffic loading. There remains lacking attention on skid resistance attenuation of SMA pavement, which in turn is important for skid durability design in practice. Hence, this study aims to perform a thorough investigation to reveal the skid resistance attenuation law of SMA pavement. Multiple types of SMA-13 mixtures prepared by different material designs were selected to conduct a kneading test that simulates real surface states of in-service SMA pavement. Pressure-sensitive film and a 3D laser scanner were utilized for evaluating anti-skid performance and skid durability. The finite element (FE) method is introduced to simulate vehicle braking distance for skid-resistance evaluation. The results show that skid resistance attenuation of SMA pavement consists of two stages: In the first stage, the skid resistance of SMA pavement experiences a short enhancement, followed by a long-term weakening stage. Abundant surface texture of SMA helps to mitigate the impact of traffic load on skid resistance. The FE analysis and pressure-sensitive film results demonstrate the potential of skid durability design of SMA pavement based on the skid resistance attenuation law.

DSGA-Net: Deeply separable gated transformer and attention strategy for medical image segmentation network.

To address the problems of under-segmentation and over-segmentation of small organs in medical image segmentation. We present a novel medical image segmentation network model with Depth Separable Gating Transformer and a Three-branch Attention module (DSGA-Net). Firstly, the model adds a Depth Separable Gated Visual Transformer (DSG-ViT) module into its Encoder to enhance (i) the contextual links among global, local, and channels and (ii) the sensitivity to location information. Secondly, a Mixed Three-branch Attention (MTA) module is proposed to increase the number of features in the up-sampling process. Meanwhile, the loss of feature information is reduced when restoring the feature image to the original image size. By validating Synapse, BraTs2020, and ACDC public datasets, the Dice Similarity Coefficient (DSC) of the results of DSGA-Net reached 81.24%,85.82%, and 91.34%, respectively. Moreover, the Hausdorff Score (HD) decreased to 20.91% and 5.27% on the Synapse and BraTs2020. There are 10.78% and 0.69% decreases compared to the Baseline TransUNet. The experimental results indicate that DSGA-Net achieves better segmentation than most advanced methods.

Classification of Alzheimer's Disease Based on Weakly Supervised Learning and Attention Mechanism.

The brain lesions images of Alzheimer's disease (AD) patients are slightly different from the Magnetic Resonance Imaging of normal people, and the classification effect of general image recognition technology is not ideal. Alzheimer's datasets are small, making it difficult to train large-scale neural networks. In this paper, we propose a network model (WS-AMN) that fuses weak supervision and an attention mechanism. The weakly supervised data augmentation network is used as the basic model, the attention map generated by weakly supervised learning is used to guide the data augmentation, and an attention module with channel domain and spatial domain is embedded in the residual network to focus on the distinctive channels and spaces of images respectively. The location information enhances the corresponding features of related features and suppresses the influence of irrelevant features.The results show that the F1-score is 99.63%, the accuracy is 99.61%. Our model provides a high-performance solution for accurate classification of AD.