Enhanced Hybrid Vision Transformer with Multi-Scale Feature Integration and Patch Dropping for Facial Expression Recognition.

Convolutional neural networks (CNNs) have made significant progress in the field of facial expression recognition (FER). However, due to challenges such as occlusion, lighting variations, and changes in head pose, facial expression recognition in real-world environments remains highly challenging. At the same time, methods solely based on CNN heavily rely on local spatial features, lack global information, and struggle to balance the relationship between computational complexity and recognition accuracy. Consequently, the CNN-based models still fall short in their ability to address FER adequately. To address these issues, we propose a lightweight facial expression recognition method based on a hybrid vision transformer. This method captures multi-scale facial features through an improved attention module, achieving richer feature integration, enhancing the network's perception of key facial expression regions, and improving feature extraction capabilities. Additionally, to further enhance the model's performance, we have designed the patch dropping (PD) module. This module aims to emulate the attention allocation mechanism of the human visual system for local features, guiding the network to focus on the most discriminative features, reducing the influence of irrelevant features, and intuitively lowering computational costs. Extensive experiments demonstrate that our approach significantly outperforms other methods, achieving an accuracy of 86.51% on RAF-DB and nearly 70% on FER2013, with a model size of only 3.64 MB. These results demonstrate that our method provides a new perspective for the field of facial expression recognition.

Prevalence and associated factors of low bone mineral density in people living with HIV: a cross-sectional study.

This study examined low bone mineral density (BMD) prevalence and associated factors among Chinese people living with HIV (PLWH), uncovering a persistent high BMD risk in older individuals, even after adjusting for age and body mass index (BMI). Notably, lopinavir/ritonavir (LPV/r) therapy was linked to reduced BMD, highlighting the imperative need for regular BMD monitoring and interventions in older PLWH. PURPOSE: HIV infection and antiretroviral therapy (ART) have been shown to contribute to lower BMD, resulting in an increased susceptibility to osteopenia and osteoporosis. However, there is limited knowledge about the prevalence of reduced BMD and its associated factors among Chinese PLWH. In this cross-sectional study, we aimed to investigate the prevalence and factors associated with low BMD among PLWH in China. METHODS: We retrospectively enrolled PLWH and non-HIV volunteers who underwent dual-energy X-ray absorptiometry (DXA) scans to measure bone density. Demographic information, laboratory test results, ART regimens, and treatment duration were collected. Univariate and multiple regression analyses were performed to identify factors influencing abnormal bone mass in PLWH. RESULTS: A total of 829 individuals were included in this study, comprising the HIV group (n = 706) and the non-HIV group (n = 123). The prevalence of low BMD among all PLWH was found to be 13.88% (98 out of 706). However, among PLWH aged 50 years and above, the prevalence increased to 65.32% (81 out of 124). In contrast, control subjects in the same age group had a prevalence of 38.21% (47 out of 123). After adjusting for age and BMI, older PLWH still demonstrated a higher prevalence of low BMD compared to the non-HIV group (68.24% vs 34.94%, P < 0.001). Multivariate analysis revealed that older age was strongly associated with a higher risk of low BMD among PLWH, with an odds ratio (OR) of 6.28 for every 10-year increase in age in the ART-naïve population (95% confidence intervals [CIs], 3.12-12.65; P < 0.001) and OR of 4.83 in the ART-experienced population (3.20-7.29, P < 0.001). Within the ART-experienced group, current LPV/r treatment was associated with an increased risk of low BMD (OR = 3.55, 1.24-10.14, P < 0.05), along with lower BMI (OR = 0.84, 0.75-0.95, P < 0.05), and elevated alkaline phosphatase (OR = 1.02, 1.01-1.03, P < 0.01). CONCLUSION: The prevalence of low BMD is higher among PLWH aged 50 years and above compared to non-HIV individuals. The use of LPV/r for ART is associated with reduced BMD. These findings emphasize the importance of regular monitoring of BMD in older PLWH and the need for appropriate interventions to mitigate the risks of osteopenia and osteoporosis in this population.

NECA alleviates inflammatory responses in diabetic retinopathy through dendritic cell toll-like receptor signaling pathway.

Introduction: This study examined the impact of 5'-(N- ethylcarboxamido)adenosine (NECA) in the peripheral blood of healthy individuals, those with diabetes mellitus, diabetic retinopathy (DR), and C57BL/6 mice, both in vivo and in vitro. Methods: Enzyme-linked immunosorbent assay (ELISA) and flow cytometry (FCM) were used to evaluate the effects of NECA on dendritic cells (DCs) and mouse bone marrow-derived dendritic cells (BMDCs) and the effects of NECA-treated DCs on Treg and Th17 cells. The effect of NECA on the Toll-like receptor (TLR) pathway in DCs was evaluated using polymerase chain reaction (PCR) and western blotting (WB). Results: FCM and ELISA showed that NECA inhibited the expression of surface markers of DCs and BMDCs, increased anti-inflammatory cytokines and decreased proinflammatory cytokines. PCR and WB showed that NCEA decreased mRNA transcription and protein expression in the TLR-4-MyD88-NF-kß pathway in DCs and BMDCs. The DR severity in streptozocin (STZ) induced diabetic mice was alleviated. NECA-treated DCs and BMDCs were co-cultivated with CD4+T cells, resulting in modulation of Treg and Th17 differentiation, along with cytokine secretion alterations. Conclusion: NECA could impair DCs' ability to present antigens and mitigate the inflammatory response, thereby alleviating the severity of DR.

A Multi-Scale Natural Scene Text Detection Method Based on Attention Feature Extraction and Cascade Feature Fusion.

Scene text detection is an important research field in computer vision, playing a crucial role in various application scenarios. However, existing scene text detection methods often fail to achieve satisfactory results when faced with text instances of different sizes, shapes, and complex backgrounds. To address the challenge of detecting diverse texts in natural scenes, this paper proposes a multi-scale natural scene text detection method based on attention feature extraction and cascaded feature fusion. This method combines global and local attention through an improved attention feature fusion module (DSAF) to capture text features of different scales, enhancing the network's perception of text regions and improving its feature extraction capabilities. Simultaneously, an improved cascaded feature fusion module (PFFM) is used to fully integrate the extracted feature maps, expanding the receptive field of features and enriching the expressive ability of the feature maps. Finally, to address the cascaded feature maps, a lightweight subspace attention module (SAM) is introduced to partition the concatenated feature maps into several sub-space feature maps, facilitating spatial information interaction among features of different scales. In this paper, comparative experiments are conducted on the ICDAR2015, Total-Text, and MSRA-TD500 datasets, and comparisons are made with some existing scene text detection methods. The results show that the proposed method achieves good performance in terms of accuracy, recall, and F-score, thus verifying its effectiveness and practicality.

Prevalence and transmission of pretreatment drug resistance in people living with HIV-1 in Shanghai China, 2017-2021.

The implementation of pretreatment drug-resistance (PDR) surveillance among people living with HIV-1 (PLWH) is a top priority in countries using efavirenz (EFV)/nevirapine (NVP) for first-line ART. In this study, we assessed the prevalence of PDR among PLWH in Shanghai, China during 2017-2021, and to reveal PDR transmission between Shanghai and other regions of China. A total of 5050 PLWH not on ART during 2017-2021 were included. Partial HIV-1 pol sequences were amplified, sequenced, and analysed for drug-resistance mutations (DRMs). Besides, transmission network of PDR variants was inferred using HIV-TRACE. The overall prevalence of PDR was 4.8% (242/5050; 95% CI, 4.2-5.4). Prevalence of NNRTI-associated PDR was 3.9% (95% CI, 3.4-4.5), higher than those of NRTI-associated (0.8%; 95% CI, 0.5-1.1) and PI-associated PDR (0.9%; 95% CI, 0.6-1.2). High prevalence of PDR (especially high-level resistance) to EFV (132/5050, 2.6%) and NVP (137/5050, 2.7%) were found. CRF01_AE (46.0%) was the predominant HIV-1 genotype with any DRMs, followed by CRF55_01B (21.0%), and CRF07_BC (15.1%). Two NRTI-associated (S68G/N/R and T215A/N/S/Y), five NNRTI-associated (V179D/E/T/L, K103N/R/S/T, E138A/G/K, V106M/I/A and Y181C/I) and two PI-associated mutations (M46I/L/V and Q58E) were the most common observed DRMs in PDR patients in Shanghai. The vast majority of S68G occurred in CRF01_AE (45%). M46I/L/V and Q58E showed a relatively high prevalence in CRF01_AE (4.1%) and CRF07_BC (12.6%). Transmission network analyses demonstrated cross-regional transmission links of PDR variants between Shanghai and other regions of China, which was mainly driven by the potential low-level DRM V179D/E. These results provide crucial information for clinical decision making of first-line ART in PLWH with PDR.

Carbon export from submarine groundwater discharge in a semi-enclosed bay: Impact for the buffering capacity against coastal ocean acidification.

Submarine groundwater discharge (SGD) serves as an important pathway for the transport of dissolved carbon from land to ocean, significantly affecting the coastal biogeochemical cycles. However, the impact of SGD-derived dissolved carbon on the coastal carbon budget remains poorly understood. This study first quantified SGD and associated dissolved organic carbon (DOC), dissolved inorganic carbon (DIC) and total alkalinity (TA) fluxes in Daya Bay using mass balance models based on radium isotopes (223Ra, 224Ra, 226Ra and 228Ra). We then constructed carbon mass balance models to evaluate the impact of SGD-derived carbon on the buffering capacity against coastal ocean acidification. The estimated SGD fluxes ranged from 0.80 × 107 to 2.64 × 107 m3d-1. The DIC, DOC and TA fluxes from SGD were 17.90-36.44 mmol m-2d-1, 0.93-2.13 mmol m-2d-1, and 21.19-28.47 mmol m-2d-1, respectively. Based on carbon mass balances, the DIC flux from SGD was 19-39 times the riverine input, accounting for 27.16 % ∼ 37.64 % of the total carbon source. These results suggest that SGD is a major contributor to DIC, significantly affecting the coastal carbon budget. Furthermore, the average TA:DIC ratio of groundwater discharging into Daya Bay was approximately 1.13. High TA exports enhance the buffering capacity of the coastal ocean and contribute bicarbonate to the ocean, playing a significant role in the ocean carbon sequestration process. This study demonstrates the importance of SGD-derived dissolved carbon in the assessment of coastal carbon budgets.

Trends in and Risk Factors for Drug Resistance in Mycobacterium tuberculosis in HIV-Infected Patients.

Trends in and risk factors for drug resistance in Mycobacterium tuberculosis (M. tuberculosis) in human immunodeficiency virus (HIV)-infected patients with active tuberculosis were analyzed. The clinical data of M. tuberculosis and HIV-coinfected patients treated at the Shanghai Public Health Clinical Center between 2010 and 2022 were collected. The diagnosis of tuberculosis was confirmed by solid or liquid culture. The phenotypic drug susceptibility test was carried out via the proportional method, and the resistance to first-line and second-line drugs was analyzed. Logistic regression analysis was performed to identify associated risk factors for drug resistance in M. tuberculosis. Of the 304 patients with a M. tuberculosis-positive culture and first-line drug susceptibility test results, 114 (37.5%) were resistant to at least one first-line anti-tuberculosis drug. Of the 93 patients with first-line and second-line drug susceptibility test results, 40 (43%) were resistant to at least one anti-tuberculosis drug, and 20 (21.5%), 27 (29.0%), 19 (20.4%), 16 (17.2%), and 14 (15.1%) were resistant to rifampicin, streptomycin, ofloxacin, levofloxacin, and moxifloxacin, respectively; 17 patients (18.3%) had multidrug-resistant tuberculosis (MDR-TB). Between 2010 and 2021, the rate of resistance to streptomycin and rifampicin ranged from 14.3% to 40.0% and from 8.0% to 26.3%, respectively, showing an increasing trend year by year. From 2016 to 2021, the rate of resistance to quinolones fluctuated between 7.7% and 27.8%, exhibiting an overall upward trend. Logistic regression analysis showed that being aged <60 years old was a risk factor for streptomycin resistance, mono-drug resistance, and any-drug resistance (RR 4.139, p = 0.023; RR 7.734, p = 0.047; RR 3.733, p = 0.009). Retreatment tuberculosis was a risk factor for resistance to rifampicin, ofloxacin, of levofloxacin (RR 2.984, p = 0.047; RR 4.517, p = 0.038; RR 6.277, p = 0.014). The drug resistance rates of M. tuberculosis to rifampicin and to quinolones in HIV/AIDS patients were high and have been increasing year by year. Age and a history of previous anti-tuberculosis treatment were the main factors associated with the development of drug resistance in HIV/AIDS patients with tuberculosis.

The profile of HIV-1 drug resistance in Shanghai, China: a retrospective study from 2017 to 2021.

BACKGROUND: HIV-1 drug resistance is a huge challenge in the era of ART. OBJECTIVES: To investigate the prevalence and characteristics of acquired HIV-1 drug resistance (ADR) in Shanghai, China. METHODS: An epidemiological study was performed among people living with human immunodeficiency virus (PLWH) receiving ART in Shanghai from January 2017 to December 2021. A total of 8669 PLWH were tested for drug resistance by genotypic resistance testing. Drug resistance mutations (DRMs) were identified using the Stanford University HIV Drug Resistance Database program. RESULTS: Ten HIV-1 subtypes/circulating recombinant forms (CRFs) were identified, mainly including CRF01_AE (46.8%), CRF07_BC (35.7%), B (6.4%), CRF55_01B (2.8%) and CRF08_BC (2.4%). The prevalence of ADR was 48% (389/811). Three NRTI-associated mutations (M184V/I/L, S68G/N/R and K65R/N) and four NNRTI-associated mutations (V179D/E/T/L, K103N/R/S/T, V106M/I/A and G190A/S/T/C/D/E/Q) were the most common DRMs. These DRMs caused high-level resistance to lamivudine, emtricitabine, efavirenz and nevirapine. The DRM profiles appeared to be significantly different among different subtypes. CONCLUSIONS: We revealed HIV-1 subtype characteristics and the DRM profile in Shanghai, which provide crucial guidance for clinical treatment and management of PLWH.

Comparative evaluation of mechanical injury methods for establishing stable tracheal stenosis animal models.

The study aimed to assess the stability of various mechanical injury techniques in creating tracheal stenosis animal models using endoscopic assistance and investigate the viability of tracheal stoma in this process. Twenty-six healthy adult New Zealand white rabbits were randomly assigned to an experimental and control group. The experimental group underwent tracheal incision followed by steel brush scraping with endoscopic assistance, while the control group received nylon brush scraping. Within the control group, two subgroups were formed: Group A underwent scraping without tracheal stoma, and Group B underwent scraping followed by tracheal stoma. Additionally, a sham operation was performed on a separate group without subsequent scratching, resulting in no stenosis formation. Endoscopic observations were conducted at 7, 14, and 21 days post-scraping, followed by histological examinations of euthanized rabbits on the 21st day. Notably, all rabbits in the non-stoma group survived without complications, whereas Group B rabbits faced mortality post-operation. Histological assessments revealed inflammatory cell infiltration, fibroblast proliferation, and collagen fiber deposition in narrowed tracheal specimens. Steel brush scraping with endoscopic assistance proved more effective in inducing stable tracheal stenosis compared to nylon brush scraping. However, the survival challenges of rabbits with tracheal fistula require further investigation.

Seawater intrusion-triggered high fluoride groundwater development on the eastern coast of China.

High-fluoride groundwater is commonly found in coastal areas worldwide, while its formation mechanism remains elusive. Herein, a comprehensive study was performed to identify the major controlling factor of high-fluoride groundwater occurrence along the eastern coast of China. Hydrogeochemical methods were employed to examine the distribution patterns of seawater intrusion and fluoride concentration and the impact of seawater intrusion on the fluoride concentration. The results indicate that seawater intrusion significantly influences the groundwater evolution process in the study areas. The groundwater in Laizhou Bay was affected by brine, and the groundwater in Tianjin and Jiangsu was affected by seawater with a mixing ratio lower than 40% and 20%, respectively. And the fluoride concentration in groundwater from Tianjin, Laizhou Bay, and Jiangsu generally exceeded 1 mg/L, with the average of 2.3 mg/L, 24.9 mg/L, and 34.6 mg/L, respectively. Both the degree of seawater intrusion and the fluoride concentration exhibit a consistent pattern: Laizhou Bay > Tianjin > Jiangsu. Cl- concentration in groundwater varies positively with the F- concentration (y = 0.66x - 1.31). Moreover, the spatial distribution of areas with high-fluoride groundwater mirrors that of seawater intrusion. The high-fluoride groundwater varies spatially and is related to the degree, stage, and type of seawater intrusion. In other words, when seawater intrusion intensifies more or groundwater in the freshwater renewal phase with higher Na+/Ca2+ or the presence of paleo-seawater intrusion with higher fluoride concentration of brine, the concentration of fluoride in groundwater is higher. As seawater intrusion intensifies, the high-fluoride groundwater in the study areas generally poses a higher health risk to human. These findings enhance our comprehension of the mechanisms underpinning high-fluoride groundwater in coastal regions and the environmental ramifications of seawater intrusion.

Impact of Recombinant Human Granulocyte Colony-Stimulating Factor Combined with Aspirin on Clinical Pregnancy Outcomes and Endometrial Receptivity in Patients with Recurrent Abortion: A Retrospective Study.

Objective: This study investigates the impact of recombinant human granulocyte colony-stimulating factor (rhG-CSF) and aspirin on endometrial receptivity and clinical pregnancy outcomes in individuals with a history of recurrent abortions. Methods: In this retrospective study, 131 individuals with recurrent abortions treated at our facility from July 2019 to December 2020 were split into two groups: mixed therapy and control. The mixed therapy group received aspirin and rhG-CSF, while the control group had no specific treatment. Primary endpoint: live birth rate; secondary: pregnancy rate at 20 weeks. We also evaluated abortion rates, newborn weight, pre-eclampsia, premature delivery, fetal/newborn congenital malformations, and maternal drug adverse reactions. Additionally, we analyzed endometrial blood flow three weeks post-treatment. Results: The analysis encompassed 131 individuals, with 65 in the control group and 66 in the mixed therapy group. Notably, the mixed therapy group (n = 54) exhibited a markedly higher live birth rate than the control group (P < .05). In terms of medication-related side effects, the control group showed no adverse reactions, while the mixed therapy group reported mild effects (skin itching in three cases, leukocytosis in seven, and bone pain in one case) that did not significantly impact outcomes. Pre-treatment, the mixed therapy group had a notably lower resistive index, pulsatility index, and systolic-to-diastolic ratio compared to the control group, with statistical significance (P < .05). The control group's indices remained unchanged (P > .05). Conclusions: In women with a history of recurrent abortions, the administration of recombinant human granulocyte colony-stimulating factor and aspirin can effectively and safely improve live birth rates. This improvement may be associated with enhanced endometrial receptivity.

Survival and prognostic factors of progressive multifocal leukoencephalopathy in people living with HIV in modern ART era.

Background: The incidence of progressive multifocal leukoencephalopathy (PML) in people living with HIV (PLWH) is 2%-4%. Currently, there is no effective therapeutic strategy for the treatment of PML in PLWH, resulting in a mortality of up to 50%. This study aimed to identify risk factors of death and prognostic markers in people living with HIV with PML. Methods: A retrospective cohort study of AIDS-related PML individuals was conducted from January 1, 2015, to October 1, 2022, in Shanghai, China. PLWH who were diagnosed with PML for the first time were included. Kaplan-Meier curve and Cox regression were used to analyze the survival and its predictors. Levels of inflammatory markers and immune checkpoint inhibitors in blood and cerebrospinal fluid (CSF) were measured in the prestored samples using bead-based multiplex assay Indolamine 2,3-dioxygenase was determined using ELISA. Results: Twenty of 71 subjects had initiated antiretroviral therapy (ART) before PML onset and no patients discontinued ART during this period. In total, 34 patients (47.9%) had opportunistic infections (OIs), the median CD4+ T cell count was 73.0 (33.0-149.0) cells/µL. The estimated probability of survival at six months was 78% (95% confidential intervals [CIs]:0.63-0.85). OIs, low CD4+ T cell count were associated with lower estimated six-month survival (hazard ratio 8.01, 95% CIs: 1.80-35.00, P=0.006 and 5.01, 95% CIs:1.57-16.03, p=0.007). Indolamine 2,3-dioxygenase activity in CSF of non-survivors group were higher than survivors group (p<0.05). Conclusions: The survival rate of AIDS-related PML in the modern ART era was higher than the survival rate a decade ago. Low CD4+T cell count, OIs, were all associated with death of individuals with AIDS-related PML. The role of IDO in AIDS-related PML warrant further investigation.

Elevated indoleamine 2,3-dioxygenase activity is associated with endothelial dysfunction in people living with HIV and ROS production in human aortic endothelial cells in vitro.

The precise role of indoleamine 2,3-dioxygenase (IDO) in cardiovascular diseases (CVD) among people living with HIV (PLWH) is still under debate, despite recognized links. This study aimed to investigate the impact of elevated IDO activity on endothelial dysfunction in PLWH. A total of 38 PLWH, who had not previously received anti-retroviral therapy (ART), were enrolled in the study. These participants were monitored for 36 months following the initiation of ART. Measurements including plasma levels of IDO activity, markers of endothelial dysfunction, inflammatory factors, and lipids. In vitro, human aortic endothelial cells (HAEC) were exposed to interferon-γ, an IDO inhibitor, a kynurenine 3-hydroxylase (KMO) inhibitor, as well as different concentrations of kynurenine. Pre-ART, PLWH demonstrated notably elevated plasma concentrations of soluble intercellular adhesion molecule 1 (sICAM-1), soluble vascular cell adhesion molecule 1(sVCAM-1), and IDO activity in comparison to healthy controls. Post-ART, both IDO activity and sICAM-1 levels experienced a significant decrease, with IDO activity reaching levels comparable to those observed in healthy controls. Furthermore, a positive correlation was observed between IDO activity and sICAM-1 (p = 0.0002), as well as sVCAM-1 (p < 0.0001) before ART. In vitro, the augmentation of kynurenine concentration in the medium and the induction of IDO expression in HAEC resulted in increased production of reactive oxygen species (ROS), with minimal impact on endothelial dysfunction. From these findings, it can be concluded that long-term ART has the potential to restore the heightened IDO activity observed in PLWH. The overexpression of IDO primarily influences the expression of ROS in HAEC.

TGFß1-RCN3-TGFBR1 loop facilitates pulmonary fibrosis by orchestrating fibroblast activation.

BACKGROUND: Idiopathic pulmonary fibrosis (IPF) bears high mortality due to unclear pathogenesis and limited therapeutic options. Therefore, identifying novel regulators is required to develop alternative therapeutic strategies. METHODS: The lung fibroblasts from IPF patients and Reticulocalbin 3 (RCN3) fibroblast-selective knockdown mouse model were used to determine the importance of Rcn3 in IPF; the epigenetic analysis and protein interaction assays, including BioID, were used for mechanistic studies. RESULTS: Reticulocalbin 3 (RCN3) upregulation is associated with the fibrotic activation of lung fibroblasts from IPF patients and Rcn3 overexpression blunts the antifibrotic effects of pirfenidone and nintedanib. Moreover, repressing Rcn3 expression in mouse fibroblasts ameliorates bleomycin-induced lung fibrosis and pulmonary dysfunction in vivo. Mechanistically, RCN3 promotes fibroblast activation by maintaining persistent activation of TGFß1 signalling via the TGFß1-RCN3-TGFBR1 positive feedback loop, in which RCN3 upregulated by TGFß1 exposure detains EZH2 (an epigenetic methyltransferase) in the cytoplasm through RCN3-EZH2 interaction, leading to the release of the EZH2-H3K27me3 epigenetic repression of TGFBR1 and the persistent expression of TGFBR1. CONCLUSIONS: These findings introduce a novel regulating mechanism of TGFß1 signalling in fibroblasts and uncover a critical role of the RCN3-mediated loop in lung fibrosis. RCN3 upregulation may cause resistance to IPF treatment and targeting RCN3 could be a novel approach to ameliorate pulmonary fibrosis.

Zanubrutinib, lenalidomide, and rituximab (ZR2 regimen) for HIV-associated diffuse large B-cell lymphoma: a real-world analysis from China.

Seven patients with HIV-associated diffuse large B-cell lymphoma (HIV-DLBCL) who did not derive benefit from traditional first-line or second-line chemotherapy were all eventually treated with zanubrutinib, rituximab, and lenalidomide (the ZR2 regimen). Three patients had a complete response, three had a partial response, and one showed stable disease. The complete response rate was 42.9%, the overall response rate was 85.7%. Three patients developed either neutropenia or thrombocytopenia, and one died of lung infection 3 months after diagnosis.

OEDL: an optimized ensemble deep learning method for the prediction of acute ischemic stroke prognoses using union features.

Background: Early stroke prognosis assessments are critical for decision-making regarding therapeutic intervention. We introduced the concepts of data combination, method integration, and algorithm parallelization, aiming to build an integrated deep learning model based on a combination of clinical and radiomics features and analyze its application value in prognosis prediction. Methods: The research steps in this study include data source and feature extraction, data processing and feature fusion, model building and optimization, model training, and so on. Using data from 441 stroke patients, clinical and radiomics features were extracted, and feature selection was performed. Clinical, radiomics, and combined features were included to construct predictive models. We applied the concept of deep integration to the joint analysis of multiple deep learning methods, used a metaheuristic algorithm to improve the parameter search efficiency, and finally, developed an acute ischemic stroke (AIS) prognosis prediction method, namely, the optimized ensemble of deep learning (OEDL) method. Results: Among the clinical features, 17 features passed the correlation check. Among the radiomics features, 19 features were selected. In the comparison of the prediction performance of each method, the OEDL method based on the concept of ensemble optimization had the best classification performance. In the comparison to the predictive performance of each feature, the inclusion of the combined features resulted in better classification performance than that of the clinical and radiomics features. In the comparison to the prediction performance of each balanced method, SMOTEENN, which is based on a hybrid sampling method, achieved the best classification performance than that of the unbalanced, oversampled, and undersampled methods. The OEDL method with combined features and mixed sampling achieved the best classification performance, with 97.89, 95.74, 94.75, 94.03, and 94.35% for Macro-AUC, ACC, Macro-R, Macro-P, and Macro-F1, respectively, and achieved advanced performance in comparison with that of methods in previous studies. Conclusion: The OEDL approach proposed herein could effectively achieve improved stroke prognosis prediction performance, the effect of using combined data modeling was significantly better than that of single clinical or radiomics feature models, and the proposed method had a better intervention guidance value. Our approach is beneficial for optimizing the early clinical intervention process and providing the necessary clinical decision support for personalized treatment.

Spatio-temporal pattern, matching level and prediction of ageing and medical resources in China.

OBJECTIVE: Population ageing, as a hot issue in global development, increases the burden of medical resources in society. This study aims to assess the current spatiotemporal evolution and interaction between population ageing and medical resources in mainland China; evaluate the matching level of medical resources to population ageing; and forecast future trends of ageing, medical resources, and the indicator of ageing-resources (IAR). METHODS: Data on ageing (EPR) and medical resources (NHI, NBHI, and NHTP) were obtained from China Health Statistics Yearbook and China Statistical Yearbook (2011-2020). We employed spatial autocorrelation to examine the spatial-temporal distribution trends and analyzed the spatio-temporal interaction using a Bayesian spatio-temporal effect model. The IAR, an improved evaluation indicator, was used to measure the matching level of medical resources to population ageing with kernel density analysis for visualization. Finally, an ETS-DNN model was used to forecast the trends in population ageing, medical resources, and their matching level over the next decade. RESULTS: The study found that China's ageing population and medical resources are growing annually, yet distribution is uneven across districts. There is a spatio-temporal interaction effect between ageing and medical resources, with higher levels of both in Eastern China and lower levels in Western China. The IAR is relatively high in Northwest, North China, and the Yangtze River Delta, but showed a declining trend in North China and the Yangtze River Delta. The hybrid model (ETS-DNN) gained an R2 of 0.9719, and the predicted median IAR for 2030 (0.99) across 31 regions was higher than the median IAR for 2020 (0.93). CONCLUSION: This study analyzes the relationship between population ageing and medical resources, revealing a spatio-temporal interaction between them. The IAR evaluation indicator highlights the need to address ageing population challenges and cultivate a competent health workforce. The ETS-DNN forecasts indicate higher concentrations of both medical resources and ageing populations in eastern China, emphasizing the need for region-specific ageing security systems and health service industries. The findings provide valuable policy insights for addressing a hyper-aged society in the future.