Enhanced predictive validity of integrative models for refractory hyperthyroidism considering baseline and early therapy characteristics: a prospective cohort study.

BACKGROUND: A subset of Graves' disease (GD) patients develops refractory hyperthyroidism, posing challenges in treatment decisions. The predictive value of baseline characteristics and early therapy indicators in identifying high risk individuals is an area worth exploration. METHODS: A prospective cohort study (2018-2022) involved 597 newly diagnosed adult GD patients undergoing methimazole (MMI) treatment. Baseline characteristics and 3-month therapy parameters were utilized to develop predictive models for refractory GD, considering antithyroid drug (ATD) dosage regimens. RESULTS: Among 346 patients analyzed, 49.7% developed ATD-refractory GD, marked by recurrence and sustained Thyrotropin Receptor Antibody (TRAb) positivity. Key baseline factors, including younger age, Graves' ophthalmopathy (GO), larger goiter size, and higher initial free triiodothyronine (fT3), free thyroxine (fT4), and TRAb levels, were all significantly associated with an increased risk of refractory GD, forming the baseline predictive model (Model A). Subsequent analysis based on MMI cumulative dosage at 3 months resulted in two subgroups: a high cumulative dosage group (average ≥ 20 mg/day) and a medium-low cumulative dosage group (average < 20 mg/day). Absolute values, percentage changes, and cumulative values of thyroid function and autoantibodies at 3 months were analyzed. Two combined predictive models, Model B (high cumulative dosage) and Model C (medium-low cumulative dosage), were developed based on stepwise regression and multivariate analysis, incorporating additional 3-month parameters beyond the baseline. In both groups, these combined models outperformed the baseline model in terms of discriminative ability (measured by AUC), concordance with actual outcomes (66.2% comprehensive improvement), and risk classification accuracy (especially for Class I and II patients with baseline predictive risk < 71%). The reliability of the above models was confirmed through additional analysis using random forests. This study also explored ATD dosage regimens, revealing differences in refractory outcomes between predicted risk groups. However, adjusting MMI dosage after early risk assessment did not conclusively improve the prognosis of refractory GD. CONCLUSION: Integrating baseline and early therapy characteristics enhances the predictive capability for refractory GD outcomes. The study provides valuable insights into refining risk assessment and guiding personalized treatment decisions for GD patients.

Real-world effectiveness of dupilumab in patients with asthma: Findings from the US ADVANTAGE study.

BACKGROUND: Dupilumab is approved as an add-on maintenance therapy for patients (≥6 years) with moderate-to-severe asthma. Better understanding of real-world effectiveness is needed. OBJECTIVE: To characterize the real-world effectiveness of dupilumab in asthma management. METHODS: This retrospective study included patients (≥12 years of age) diagnosed with asthma, initiating dupilumab between November 2018 and September 2020. The study used a US electronic medical record database (TriNetX Dataworks, Cambridge, Massachusetts). Asthma exacerbation rates before and after the initiation of dupilumab were analyzed using generalized estimating equations models with Poisson probabilistic link to estimate incidence rate ratios (IRRs). Sensitivity analyses were conducted based on previous exacerbation data, eosinophil levels, history of atopic dermatitis or chronic rhinosinusitis with nasal polyps, previous use of biologics, and presence of SARS-CoV-2 (COVID-19). RESULTS: A total of 2400 patients initiating dupilumab met all study criteria. After initiation of dupilumab, risk of asthma exacerbation was reduced by 44% (IRR, 0.56; 95% CI, 0.47-0.57; P = <0.0001) and systemic corticosteroid prescriptions by 48% (IRR, 0.52; 95% CI, 0.48, 0.56; P = <0.0001) compared with those before initiation of dupilumab. Adjustment for COVID-19 showed a greater reduction in asthma exacerbations (IRR, 0.50; 95% CI, 0.45-0.55; P = <0.0001). CONCLUSION: Current real-world efficacy evidence indicates that dupilumab reduces asthma exacerbations and total systemic corticosteroid prescriptions in clinical practice. The effectiveness of dupilumab was observed independent of exacerbation history, eosinophil levels, or COVID-19 impact.

An Explainable and Personalized Cognitive Reasoning Model Based on Knowledge Graph: Toward Decision Making for General Practice.

General practice plays a prominent role in primary health care (PHC). However, evidence has shown that the quality of PHC is still unsatisfactory, and the accuracy of clinical diagnosis and treatment must be improved in China. Decision making tools based on artificial intelligence can help general practitioners diagnose diseases, but most existing research is not sufficiently scalable and explainable. An explainable and personalized cognitive reasoning model based on knowledge graph (CRKG) proposed in this article can provide personalized diagnosis, perform decision making in general practice, and simulate the mode of thinking of human beings utilizing patients' electronic health records (EHRs) and knowledge graph. Taking abdominal diseases as the application point, an abdominal disease knowledge graph is first constructed in a semiautomated manner. Then, the CRKG designed referring to dual process theory in cognitive science involves the update strategy of global graph representations and reasoning on a personal cognitive graph by adopting the idea of graph neural networks and attention mechanisms. For the diagnosis of diseases in general practice, the CRKG outperforms all the baselines with a precision@1 of 0.7873, recall@10 of 0.9020 and hits@10 of 0.9340. Additionally, the visualization of the reasoning process for each visit of a patient based on the knowledge graph enhances clinicians' comprehension and contributes to explainability. This study is of great importance for the exploration and application of decision making based on EHRs and knowledge graph.

Key protein identification by integrating protein complex information and multi-biological features.

Identifying key proteins based on protein-protein interaction networks has emerged as a prominent area of research in bioinformatics. However, current methods exhibit certain limitations, such as the omission of subcellular localization information and the disregard for the impact of topological structure noise on the reliability of key protein identification. Moreover, the influence of proteins outside a complex but interacting with proteins inside the complex on complex participation tends to be overlooked. Addressing these shortcomings, this paper presents a novel method for key protein identification that integrates protein complex information with multiple biological features. This approach offers a comprehensive evaluation of protein importance by considering subcellular localization centrality, topological centrality weighted by gene ontology (GO) similarity and complex participation centrality. Experimental results, including traditional statistical metrics, jackknife methodology metric and key protein overlap or difference, demonstrate that the proposed method not only achieves higher accuracy in identifying key proteins compared to nine classical methods but also exhibits robustness across diverse protein-protein interaction networks.

H3K18 lactylation of senescent microglia potentiates brain aging and Alzheimer's disease through the NFκB signaling pathway.

Cellular senescence serves as a fundamental and underlying activity that drives the aging process, and it is intricately associated with numerous age-related diseases, including Alzheimer's disease (AD), a neurodegenerative aging-related disorder characterized by progressive cognitive impairment. Although increasing evidence suggests that senescent microglia play a role in the pathogenesis of AD, their exact role remains unclear. In this study, we quantified the levels of lactic acid in senescent microglia, and hippocampus tissues of naturally aged mice and AD mice models (FAD4T and APP/PS1). We found lactic acid levels were significantly elevated in these cells and tissues compared to their corresponding counterparts, which increased the level of pan histone lysine lactylation (Kla). We aslo identified all histone Kla sites in senescent microglia, and found that both the H3K18 lactylation (H3K18la) and Pan-Kla were significantly up-regulated in senescent microglia and hippocampus tissues of naturally aged mice and AD modeling mice. We demonstrated that enhanced H3K18la directly stimulates the NFκB signaling pathway by increasing binding to the promoter of Rela (p65) and NFκB1(p50), thereby upregulating senescence-associated secretory phenotype (SASP) components IL-6 and IL-8. Our study provides novel insights into the physiological function of Kla and the epigenetic regulatory mechanism that regulates brain aging and AD. Specifically, we have identified the H3K18la/NFκB axis as a critical player in this process by modulating IL-6 and IL-8. Targeting this axis may be a potential therapeutic strategy for delaying aging and AD by blunting SASP.

Integrating Medical Domain Knowledge for Early Diagnosis of Fever of Unknown Origin: An Interpretable Hierarchical Multimodal Neural Network Approach.

Accurate and interpretable differential diagnostic technologies are crucial for supporting clinicians in decision-making and treatment-planning for patients with fever of unknown origin (FUO). Existing solutions commonly address the diagnosis of FUO by transforming it into a multi-classification task. However, after the emergence of COVID-19 pandemic, clinicians have recognized the heightened significance of early diagnosis in patients with FUO, particularly for practical needs such as early triage. This has resulted in increased demands for identifying a wider range of etiologies, shorter observation windows, and better model interpretability. In this article, we propose an interpretable hierarchical multimodal neural network framework (iHMNNF) to facilitate early diagnosis of FUO by incorporating medical domain knowledge and leveraging multimodal clinical data. The iHMNNF comprises a top-down hierarchical reasoning framework (Td-HRF) built on the class hierarchy of FUO etiologies, five local attention-based multimodal neural networks (La-MNNs) trained for each parent node of the class hierarchy, and an interpretable module based on layer-wise relevance propagation (LRP) and attention mechanism. Experimental datasets were collected from electronic health records (EHRs) at a large-scale tertiary grade-A hospital in China, comprising 34,051 hospital admissions of 30,794 FUO patients from January 2011 to October 2020. Our proposed La-MNNs achieved area under the receiver operating characteristic curve (AUROC) values ranging from 0.7809 to 0.9035 across all five decomposed tasks, surpassing competing machine learning (ML) and single-modality deep learning (DL) methods while also providing enhanced interpretability. Furthermore, we explored the feasibility of identifying FUO etiologies using only the first N-hour time series data obtained after admission.

Long-Term Effectiveness of Dupilumab in Patients with Atopic Dermatitis: Results up to 3 Years from the RELIEVE-AD Study.

INTRODUCTION: Atopic dermatitis (AD) can require long-term therapy. Few real-world studies have evaluated long-term effectiveness from the patients' perspective. The aim of this study was to evaluate patient-reported outcomes (PROs) during long-term dupilumab treatment. METHODS: Adults with moderate-to-severe AD who initiated dupilumab through the US manufacturer patient support program and participated in RELIEVE-AD (a prospective patient survey study with a 12-month follow-up) were recontacted 30-36 months post-initiation regardless of current dupilumab use. The online questionnaire consisted of PROs, including the Atopic Dermatitis Control Tool (ADCT), use of concomitant AD therapies, satisfaction with current therapy, global change in itch relative to before dupilumab initiation, non-itch skin symptoms (skin pain/soreness, hot/burning feeling, and sensitivity to touch), flares, Dermatology Life Quality Index, sleep problems, and the AD-specific Work Productivity and Activity Impairment Questionnaire. RESULTS: Of 698 patients who initiated dupilumab (baseline) and were recontacted, 425 completed the 30-36-month survey. Significant reductions from baseline were reported in concomitant AD therapy use (P < 0.05); 54.4% reported not using other AD medications vs. 12.8% at baseline. At 30-36 months, all results (non-itch skin symptoms, flares, sleep problems, health-related quality of life work/activity impairment, disease control, and treatment satisfaction) were similar to or incrementally better than the 12-month timepoint, with significant improvements vs. baseline (P < 0.001). Global change in itch was reported as "very much better" by 75.3% of respondents. Adequate disease control (score < 7 on ADCT) was reported by 80.7% of respondents, and 86.8% were satisfied with the treatment. CONCLUSIONS: In clinical practice settings, patient-reported benefits of dupilumab were maintained in survey respondents during long-term treatment up to 36 months while the use of concomitant AD therapies reduced.

Atopic dermatitis (also known as eczema) is a chronic skin disease that can have a profoundly negative effect on patients' quality of life. To control disease symptoms, patients often need long-term treatment. Dupilumab is a treatment that has shown benefits in adults with moderate-to-severe atopic dermatitis (AD) when used in long-term (under 4 years) clinical trials; however, few studies have evaluated patients' experiences of long-term dupilumab treatment outside of a clinical trial setting. This study was conducted to do so: 425 adults with moderate-to-severe AD who received dupilumab through a US manufacturer patient support program filled in an online questionnaire 30­36 months after starting treatment. The questionnaire included items on use of additional AD therapies, AD symptoms, quality of life, disease control, and satisfaction with treatment. Patients' responses showed that, at 30­36 months after starting dupilumab treatment, 54% of patients reported not using any other medications for AD vs. 13% of patients when starting dupilumab treatment. In addition, since starting dupilumab, 75% of patients reported one of the most burdensome AD symptoms, itch, as being "very much better" vs. before starting treatment; 81% reported control of AD symptoms; 85% reported a meaningful improvement in quality of life; and 76% were "extremely" or "very" satisfied with the treatment. In summary, this study showed that long-term dupilumab treatment provides continued improvement in symptoms, treatment satisfaction, disease control, and quality of life in adults with moderate-to-severe AD while reducing the need for other AD treatments. Video abstract: How do patients with atopic dermatitis perceive long-term dupilumab treatment in the real world? (MP4 31888 kb).

Interaction between the GCKR rs1260326 variant and serum HDL cholesterol contributes to HOMA-ß and ISIMatusda in the middle-aged T2D individuals.

This study aims to investigate the correlations between islet function/ insulin resistance and serum lipid levels, as well as to assess whether the strength of such correlations is affected by the GCKR rs1260326 variant in healthy and T2D individuals. We performed an oral glucose tolerance test (OGTT) on 4889 middle-aged adults, including 3135 healthy and 1754 T2D individuals from the REACTION population study in the Nanjing region. We also measured their serum lipid levels and genotyped for rs1260326. We found that serum high-density lipoprotein (HDL) cholesterol and triglyceride (TG) levels were independently correlated with indexes of islet function (HOMA-ß and IGI [insulinogenic index]) and insulin resistance (HOMO-IR and ISIMatsuda) in both healthy and T2D individuals. The correlations were significantly decreased in T2D individuals, with significant heterogeneities compared to healthy controls (I2 > 75%, Phet < 0.05). Although no correlation was observed between serum total cholesterol (TC) level and islet function/ insulin resistance in healthy controls, significant correlations were found in T2D individuals, with significant heterogeneity to healthy controls in the correlation with ISIMatsuda(I2 = 85.3%, Phet = 0.009). Furthermore, we found significant interactions of the GCKR rs1260326 variant for the correlations between serum HDL cholesterol and HOMA-ß/ISIMatsuda in T2D subjects (P = 0.015 and 0.038, respectively). These findings illustrate that distinct correlations between serum lipid levels and islet function/ insulin resistance occurred in T2D subjects compared to healthy individuals. Common gene variants, such as rs1260326, might interact substantially when studied in specific populations, especially T2D disease status.

Epileptic detection in single and multi-lead EEG signals using persistent homology based on bi-directional weighted visibility graphs.

Epilepsy is a widespread neurological disorder, and its recurrence and suddenness are making automatic detection of seizure an urgent necessity. For this purpose, this paper performs topological data analysis (TDA) of electroencephalographic (EEG) signals by the medium of graphs to explore the potential brain activity information they contain. Through our innovative method, we first map the time series of epileptic EEGs into bi-directional weighted visibility graphs (BWVGs), which give more comprehensive reflections of the signals compared to previous existing structures. Traditional graph-theoretic measurements are generally partial and mainly consider differences or correlations in vertices or edges, whereas persistent homology (PH), the essential part of TDA, provides an alternative way of thinking by quantifying the topology structure of the graphs and analyzing the evolution of these topological properties with scale changes. Therefore, we analyze the PH for BWVGs and then obtain the two indicators of persistence and birth-death for homology groups to reflect the topology of the mapping graphs of EEG signals and reveal the discrepancies in brain dynamics. Furthermore, we adopt neural networks (NNs) for the automatic detection of epileptic signals and successfully achieve a classification accuracy of 99.67% when distinguishing among three different sets of EEG signals from seizure, seizure-free, and healthy subjects. In addition, to accommodate multi-leads, we propose a classifier that incorporates graph structure to distinguish seizure and seizure-free EEG signals. The classification accuracies of the two subjects used in the classifier are as high as 99.23% and 94.76%, respectively, indicating that our proposed model is useful for the analysis of EEG signals.

Deep learning-based small object detection: A survey.

Small object detection (SOD) is significant for many real-world applications, including criminal investigation, autonomous driving and remote sensing images. SOD has been one of the most challenging tasks in computer vision due to its low resolution and noise representation. With the development of deep learning, it has been introduced to boost the performance of SOD. In this paper, focusing on the difficulties of SOD, we analyze the deep learning-based SOD research papers from four perspectives, including boosting the resolution of input features, scale-aware training, incorporating contextual information and data augmentation. We also review the literature on crucial SOD tasks, including small face detection, small pedestrian detection and aerial image object detection. In addition, we conduct a thorough performance evaluation of generic SOD algorithms and methods for crucial SOD tasks on four well-known small object datasets. Our experimental results show that network configuring to boost the resolution of input features can enable significant performance gains on WIDER FACE and Tiny Person. Finally, several potential directions for future research in the area of SOD are provided.

NSUN2 alleviates doxorubicin-induced myocardial injury through Nrf2-mediated antioxidant stress.

Doxorubicin (DOX) is a commonly used antitumor drug, but its application has been limited because of its strong cardiac damage. This study aims to explore the role of NSUN2 in DOX-induced heart injury. C57BL/6J mice were intraperitoneally injected with 20 mg/Kg DOX to induce heart injury. After 3 days, the cardiac function, cardiac histopathology, myocardial apoptosis, and the expression level of NSUN2 were detected. In vitro, H9C2 cells were transfected with NSUN2 siRNA or overexpressed lentivirus and then treated with 500 ng/ml DOX. After 24 h, the changes in reactive oxygen species (ROS), apoptosis, and NSUN2 expression were detected. After DOX treatment, both in vitro and in vivo experiments showed that the cardiac function decreased, the number of apoptotic cells increased, and the expression level of NSUN2 increased. Interfering the expression of NSUN2 by siRNA promoted DOX-induced heart injury, while overexpression of NSUN2 could inhibit DOX-induced heart injury. Further study showed that NSUN2 promoted antioxidative stress by upregulating the Nrf2 protein level. In addition, NSUN2 overexpression could increase the half-life of Nrf2 mRNA. m5C RNA methylation immunoprecipitation (MeRIP) also showed that the level of Nrf2 m5C mRNA was significantly increased in NSUN2 overexpressed group when compared to the GFP group. NSUN2 enhances the expression of Nrf2 by promoting Nrf2 mRNA m5C modification and enhances its antioxidative stress effect to alleviate DOX-induced myocardial injury.

A Multimodel-Based Deep Learning Framework for Short Text Multiclass Classification with the Imbalanced and Extremely Small Data Set.

Text classification plays an important role in many practical applications. In the real world, there are extremely small datasets. Most existing methods adopt pretrained neural network models to handle this kind of dataset. However, these methods are either difficult to deploy on mobile devices because of their large output size or cannot fully extract the deep semantic information between phrases and clauses. This paper proposes a multimodel-based deep learning framework for short-text multiclass classification with an imbalanced and extremely small dataset. Our framework mainly includes five layers: the encoder layer, the word-level LSTM network layer, the sentence-level LSTM network layer, the max-pooling layer, and the SoftMax layer. The encoder layer uses DistilBERT to obtain context-sensitive dynamic word vectors that are difficult to represent in traditional feature engineering methods. Since the transformer part of this layer is distilled, our framework is compressed. Then, we use the next two layers to extract deep semantic information. The output of the encoder layer is sent to a bidirectional LSTM network, and the feature matrix is extracted hierarchically through the LSTM at the word and sentence level to obtain the fine-grained semantic representation. After that, the max-pooling layer converts the feature matrix into a lower-dimensional matrix, preserving only the obvious features. Finally, the feature matrix is taken as the input of a fully connected SoftMax layer, which contains a function that can convert the predicted linear vector into the output value as the probability of the text in each classification. Extensive experiments on two public benchmarks demonstrate the effectiveness of our proposed approach on an extremely small dataset. It retains the state-of-the-art baseline performance in terms of precision, recall, accuracy, and F1 score, and through the model size, training time, and convergence epoch, we can conclude that our method can be deployed faster and lighter on mobile devices.

Dupilumab in children aged 6 months to younger than 6 years with uncontrolled atopic dermatitis: a randomised, double-blind, placebo-controlled, phase 3 trial.

BACKGROUND: Current systemic treatments for children younger than 6 years with moderate-to-severe atopic dermatitis that is uncontrolled with topical therapies might have suboptimal efficacy and safety. Dupilumab is approved for older children and adults with atopic dermatitis and for other type 2 inflammatory conditions. We aimed to evaluate efficacy and safety of dupilumab with concomitant low-potency topical corticosteroids in children aged 6 months to younger than 6 years with moderate-to-severe atopic dermatitis. METHODS: This randomised, double-blind, placebo-controlled, parallel-group, phase 3 trial was conducted in 31 hospitals, clinics, and academic institutions in Europe and North America. Eligible patients were aged 6 months to younger than 6 years, with moderate-to-severe atopic dermatitis (Investigator's Global Assessment [IGA] score 3-4) diagnosed according to consensus criteria of the American Academy of Dermatology, and an inadequate response to topical corticosteroids. Patients were randomly assigned (1:1) to subcutaneous placebo or dupilumab (bodyweight ≥5 kg to <15 kg: 200 mg; bodyweight ≥15 kg to <30 kg: 300 mg) every 4 weeks plus low-potency topical corticosteroids (hydrocortisone acetate 1% cream) for 16 weeks. Randomisation was stratified by age, baseline bodyweight, and region. Patient allocation was done via a central interactive web response system, and treatment allocation was masked. The primary endpoint at week 16 was the proportion of patients with IGA score 0-1 (clear or almost clear skin). The key secondary endpoint (coprimary endpoint for the EU and EU reference market) at week 16 was the proportion of patients with at least a 75% improvement from baseline in Eczema Area and Severity Index (EASI-75). Primary analyses were done in the full analysis set (ie, all randomly assigned patients, as randomly assigned) and safety analyses were done in all patients who received any study drug. This study was registered with ClinicalTrials.gov, NCT03346434. FINDINGS: Between June 30, 2020, and Feb 12, 2021, 197 patients were screened for eligibility, 162 of whom were randomly assigned to receive dupilumab (n=83) or placebo (n=79) plus topical corticosteroids. At week 16, significantly more patients in the dupilumab group than in the placebo group had IGA 0-1 (23 [28%] vs three [4%], difference 24% [95% CI 13-34]; p<0·0001) and EASI-75 (44 [53%] vs eight [11%], difference 42% [95% CI 29-55]; p<0·0001). Overall prevalence of adverse events was similar in the dupilumab group (53 [64%] of 83 patients) and placebo group (58 [74%] of 78 patients). Conjunctivitis incidence was higher in the dupilumab group (four [5%]) than the placebo group (none). No dupilumab-related adverse events were serious or led to treatment discontinuation. INTERPRETATION: Dupilumab significantly improved atopic dermatitis signs and symptoms versus placebo in children younger than 6 years. Dupilumab was well tolerated and showed an acceptable safety profile, similar to results in older children and adults. FUNDING: Sanofi and Regeneron Pharmaceuticals.

A Weighted Symmetric Graph Embedding Approach for Link Prediction in Undirected Graphs.

Link prediction is an important task in social network analysis and mining because of its various applications. A large number of link prediction methods have been proposed. Among them, the deep learning-based embedding methods exhibit excellent performance, which encodes each node and edge as an embedding vector, enabling easy integration with traditional machine learning algorithms. However, there still remain some unsolved problems for this kind of methods, especially in the steps of node embedding and edge embedding. First, they either share exactly the same weight among all neighbors or assign a completely different weight to each node to obtain the node embedding. Second, they can hardly keep the symmetry of edge embeddings obtained from node representations by direct concatenation or other binary operations such as averaging and Hadamard product. In order to solve these problems, we propose a weighted symmetric graph embedding approach for link prediction. In node embedding, the proposed approach aggregates neighbors in different orders with different aggregating weights. In edge embedding, the proposed approach bidirectionally concatenates node pairs both forwardly and backwardly to guarantee the symmetry of edge representations while preserving local structural information. The experimental results show that our proposed approach can better predict network links, outperforming the state-of-the-art methods. The appropriate aggregating weight assignment and the bidirectional concatenation enable us to learn more accurate and symmetric edge representations for link prediction.

The Frequency of Intrathyroidal Follicular Helper T Cells Varies with the Progression of Graves' Disease and Hashimoto's Thyroiditis.

OBJECTIVE: Autoimmune thyroid diseases (AITD), mainly Graves' disease (GD) and Hashimoto's thyroiditis (HT), are common organ-specific autoimmune diseases characterized by circulating antibodies and lymphocyte infiltration. Follicular helper T (Tfh) cell dysregulation is involved in the development of autoimmune pathologies. We aimed to explore the role of intrathyroidal and circulating Tfh cells in patients with GD and HT. METHODS: Ultrasound-guided thyroid fine-needle aspiration (FNA) was conducted in 35 patients with GD, 40 patients with HT, and 22 patients with nonautoimmune thyroid disease (nAITD). Peripheral blood samples were also obtained from 40 patients with GD, 40 patients with HT, and 40 healthy controls. The frequencies of intrathyroidal and circulating Tfh cells from FNA and peripheral blood samples were assessed by flow cytometry. Additionally, the correlations between the frequencies of the Tfh cells and the levels of autoantibodies and hormones or disease duration were analyzed. RESULTS: The frequency of intrathyroidal CD4+CXCR5+ICOShigh Tfh cells was higher in HT patients than in GD patients. Significant correlations were identified between the percentages of circulating and intrathyroidal Tfh cells and the serum concentrations of thyroid autoantibodies, especially thyroglobulin antibodies (TgAb), in AITD. Intrathyroidal CD4+CXCR5+ICOShigh Tfh cells were positively correlated with free triiodothyronine (FT3) in HT patients but negatively correlated with FT3 in GD patients. In addition, HT patients with a longer disease duration had an increased frequency of intrathyroidal CD4+CXCR5+ICOShigh and CD4+CXCR5+PD-1+ Tfh cells. In contrast, in the GD patients, a longer disease duration did not affect the frequency of intrathyroidal CD4+CXCR5+ICOShigh but was associated with a lower frequency of CD4+CXCR5+PD-1+ Tfh cells. CONCLUSIONS: Our data suggest that intrathyroidal Tfh cells might play a role in the pathogenesis of AITD and they are potential immunobiomarkers for AITD.

Automatic Segmentation of Cervical Cells Based on Star-Convex Polygons in Pap Smear Images.

Cervical cancer is one of the most common cancers that threaten women's lives, and its early screening is of great significance for the prevention and treatment of cervical diseases. Pathologically, the accurate segmentation of cervical cells plays a crucial role in the diagnosis of cervical cancer. However, the frequent presence of adherent or overlapping cervical cells in Pap smear images makes separating them individually a difficult task. Currently, there are few studies on the segmentation of adherent cervical cells, and the existing methods commonly suffer from low segmentation accuracy and complex design processes. To address the above problems, we propose a novel star-convex polygon-based convolutional neural network with an encoder-decoder structure, called SPCNet. The model accomplishes the segmentation of adherent cells relying on three steps: automatic feature extraction, star-convex polygon detection, and non-maximal suppression (NMS). Concretely, a new residual-based attentional embedding (RAE) block is suggested for image feature extraction. It fuses the deep features from the attention-based convolutional layers with the shallow features from the original image through the residual connection, enhancing the network's ability to extract the abundant image features. And then, a polygon-based adaptive NMS (PA-NMS) algorithm is adopted to screen the generated polygon proposals and further achieve the accurate detection of adherent cells, thus allowing the network to completely segment the cell instances in Pap smear images. Finally, the effectiveness of our method is evaluated on three independent datasets. Extensive experimental results demonstrate that the method obtains superior segmentation performance compared to other well-established algorithms.