Adjuvant Transarterial Chemoembolization With Sorafenib for Portal Vein Tumor Thrombus: A Randomized Clinical Trial.

Importance: Certain patients with hepatocellular carcinoma with portal vein tumor thrombus could benefit from surgical resection, and postoperative adjuvant therapy may lower the incidence of tumor recurrence. Objective: To compare the efficacy and safety of sorafenib plus transarterial chemoembolization vs sorafenib alone as postoperative adjuvant therapy for patients with hepatocellular carcinoma with portal vein tumor thrombus. Design, Setting, and Participants: This was a phase 3, multicenter, randomized clinical trial conducted in 5 hospitals in China. A total of 158 patients were enrolled and randomized from October 2019 to March 2022, with a median follow-up of 28.4 months. Portal vein tumor thrombus was graded by the Cheng classification. Eligible patients with hepatocellular carcinoma with Cheng grade I to III portal vein tumor thrombus (ie, involving segmental or sectoral branches, right- or left-side branch, or main trunk of portal vein) were included. Interventions: Patients were randomly assigned 1:1 to receive transarterial chemoembolization with sorafenib or sorafenib alone as postoperative adjuvant therapy. Sorafenib treatment was started within 3 days after randomization, with an initial dose of 400 mg orally twice a day. In the transarterial chemoembolization with sorafenib group, transarterial chemoembolization was performed 1 day after the first administration of sorafenib. Main Outcomes and Measures: The primary end point was recurrence-free survival. Efficacy was assessed in the intention-to-treat population and safety was assessed in patients who received at least 1 dose of study treatment. Results: Of 158 patients included, the median (IQR) age was 54 (43-61) years, and 140 (88.6%) patients were male. The median (IQR) recurrence-free survival was significantly longer in the transarterial chemoembolization with sorafenib group (16.8 [12.0-NA] vs 12.6 [7.8-18.1] months; hazard ratio [HR], 0.57; 95% CI, 0.39-0.83; P = .002). The median (IQR) overall survival was also significantly longer with transarterial chemoembolization with sorafenib than with sorafenib alone (30.4 [20.6-NA] vs 22.5 [15.4-NA] months; HR, 0.57; 95% CI, 0.36-0.91; P = .02). The most common grade 3/4 adverse event was hand-foot syndrome (23 of 79 patients in the transarterial chemoembolization with sorafenib group [29.1%] vs 24 of 79 patients in the sorafenib alone group [30.4%]). There were no treatment-related deaths in either group. The transarterial chemoembolization with sorafenib group did not show additional toxicity compared with the sorafenib monotherapy group. Conclusion and Relevance: In this study, the combination of sorafenib and transarterial chemoembolization as postoperative adjuvant therapy in patients with hepatocellular carcinoma with portal vein tumor thrombus resulted in longer recurrence-free survival and overall survival than sorafenib alone and was well tolerated. Trial Registration: ClinicalTrials.gov Identifier: NCT04143191.

The Role of Artificial Intelligence in the Identification and Evaluation of Bone Fractures.

Artificial intelligence (AI), particularly deep learning, has made enormous strides in medical imaging analysis. In the field of musculoskeletal radiology, deep-learning models are actively being developed for the identification and evaluation of bone fractures. These methods provide numerous benefits to radiologists such as increased diagnostic accuracy and efficiency while also achieving standalone performances comparable or superior to clinician readers. Various algorithms are already commercially available for integration into clinical workflows, with the potential to improve healthcare delivery and shape the future practice of radiology. In this systematic review, we explore the performance of current AI methods in the identification and evaluation of fractures, particularly those in the ankle, wrist, hip, and ribs. We also discuss current commercially available products for fracture detection and provide an overview of the current limitations of this technology and future directions of the field.

A review of self-supervised, generative, and few-shot deep learning methods for data-limited magnetic resonance imaging segmentation.

Magnetic resonance imaging (MRI) is a ubiquitous medical imaging technology with applications in disease diagnostics, intervention, and treatment planning. Accurate MRI segmentation is critical for diagnosing abnormalities, monitoring diseases, and deciding on a course of treatment. With the advent of advanced deep learning frameworks, fully automated and accurate MRI segmentation is advancing. Traditional supervised deep learning techniques have advanced tremendously, reaching clinical-level accuracy in the field of segmentation. However, these algorithms still require a large amount of annotated data, which is oftentimes unavailable or impractical. One way to circumvent this issue is to utilize algorithms that exploit a limited amount of labeled data. This paper aims to review such state-of-the-art algorithms that use a limited number of annotated samples. We explain the fundamental principles of self-supervised learning, generative models, few-shot learning, and semi-supervised learning and summarize their applications in cardiac, abdomen, and brain MRI segmentation. Throughout this review, we highlight algorithms that can be employed based on the quantity of annotated data available. We also present a comprehensive list of notable publicly available MRI segmentation datasets. To conclude, we discuss possible future directions of the field-including emerging algorithms, such as contrastive language-image pretraining, and potential combinations across the methods discussed-that can further increase the efficacy of image segmentation with limited labels.

Research on optimization method of railway construction scheme based on multidimensional combination weighting and improved grey theory.

The optimization of railway construction schemes is a complexity system engineering task with multiple dimensions, diverse conditional constraints, and multifaceted objective assessments. The decision-making and scheme evaluation entail subjectivity, randomness, and fuzziness. To address the comprehensive optimization challenge in construction schemes effectively and efficiently, we investigate an optimization method for railway construction schemes. This method is based on multi-dimensional combination weighting and improved grey theory. After analyzing the primary influencing factors, we established a railway construction plan optimization index system comprising 4 dimensions and 18 factors. The weight combination coefficient is determined using the pros and cons solution distance method, and the optimal weight set for the index is determined through the multi-dimensional combination weighting approach. Utilizing the method of superior and inferior solution distance coupled with grey theory, we ascertain the order of advantages and disadvantages for each construction scheme, subsequently achieving construction scheme optimization. To illustrate this, we employ the optimization process for a high-speed railway section in Guangxi as an exemplar. The verification results indicate that the gray relative closeness values for schemes A, B, and C are 0.7089, 0.4813, and 0.4463, respectively. Scheme A has the highest gray relative closeness value, thus making it the optimal route scheme. The optimal results obtained through this method align with the outcomes of expert validation and existing research, thereby validating the effectiveness and practicality of the model. By employing a multidimensional combination weighting method, the deficiencies of traditional indicator weight calculations are mitigated, resulting in indicator weights that are more reflective of the actual circumstances. At the same time, the application of improvements in the grey theory comprehensive evaluation method enables the integration and computation of indicator data for each construction plan. Through the intuitive representation of grey relative closeness, the advantages and disadvantages of each plan are effectively characterized. This enhances the scientific rigor and applicability of the railway construction plan optimization process. The research findings can serve as a reference for similar railway construction scheme selection problems in the future.

Effects of Geographical Origin and Tree Age on the Stable Isotopes and Multi-Elements of Pu-erh Tea.

Pu-erh tea is a famous tea worldwide, and identification of the geographical origin of Pu-erh tea can not only protect manufacture's interests, but also boost consumers' confidence. However, tree age may also influence the fingerprints of Pu-erh tea. In order to study the effects of the geographical origin and tree age on the interactions of stable isotopes and multi-elements of Pu-erh tea, 53 Pu-erh tea leaves with three different age stages from three different areas in Yunnan were collected in 2023. The δ13C, δ15N values and 25 elements were determined and analyzed. The results showed that δ13C, δ15N, Mg, Mn, Fe, Cu, Zn, Rb, Sr, Y, La, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu had significant differences among different geographical origins (p < 0.05). Mn content was significantly influenced by region and tree age interaction. Based on multi-way analysis of variance, principal component analysis and step-wised discriminant analysis, 24 parameters were found to be closely related to the geographical origin rather than tree age, and the geographical origin of Pu-erh tea can be 100.0% discriminated in cross-validation with six parameters (δ13C, δ15N, Mn, Mg, La, and Tb). The study could provide references for the establishment of a database for the traceability of Pu-erh tea, and even the identification of tea sample regions with different tree ages.

Perimeter monitoring of urban buried pipeline threated by construction activities based on distributed fiber optic sensing and real-time object detection.

Urban construction activities seriously jeopardize the security of buried pipeline. Distributed optical fiber vibration monitoring is one of the most promising ways to prevent third-party threats, of which the biggest challenge is to quickly and accurately detect rare abnormal events from extremely large amounts of time-space raw data. By analogy with image recognition, the task here is similar to object detection if considering the time-space optical signals as the grayscale images and the abnormal events as the objects. Given this, what we believe to be a novel monitoring method is proposed, which consists of two Faster R-CNN models, a max pooling layer and a monitoring strategy. In the field tests, the 86-hour optical vibration signals for 5.25 km distance are recognized within 6.6 minutes with the recognition rate of 98.85% for construction activities, and only two false alarms are issued. The proposed method can reduce the recognition time by 99.59% compared to the CNN-based method.

Precision co-composting of multi-source organic solid wastes provide a sustainable waste management strategy with high eco-efficiency: a life cycle assessment.

With the increase of organic solid wastes (OSWs), current waste management practices, such as landfill, incineration, and windrow composting, have shown weaknesses in both resource recycling and environmental protection. Co-composting has been used to achieve nutrient and carbon recycling but is accused of high ammonia emission and low degradation efficiency. Therefore, this study developed a precision co-composting strategy (S3, which adds functional bacteria generated from food processing waste to a co-composting system) and compared it with the current OSW treatment strategy (S1) and traditional co-composting strategy (S2) from a life cycle assessment (LCA) perspective. The results showed that compared with S1, the eco-efficiency increased by 31.3% due to the higher economic profit of S2 but did not directly reduce the environmental cost. The addition of bacterial agents reduced ammonia emissions and shortened composting time, so compared with S1 and S2, the environmental cost of S3 was reduced by 37.9 and 43.6%, while the economic profit increased by 79.8 and 24.4%, respectively. The changes in environmental costs and economic benefits resulted in a huge improvement of S3's eco-efficiency, which was 189.6 and 121.7% higher than S1 and S2. Meanwhile, the adoption of S3 at a national scale in China could reduce the emission of 1,4-dichlorobenzene by 99.9% compared with S1 and increase profits by 6.58 billion USD per year. This study proposes a novel approach that exhibits high eco-efficiency in the treatment of OSWs.

Sintilimab plus bevacizumab combined with radiotherapy as first-line treatment for hepatocellular carcinoma with portal vein tumor thrombus: A multicenter, single-arm, phase 2 study.

BACKGROUND AND AIMS: Systemic treatments are listed as first-line therapies for HCC with portal vein tumor thrombus (PVTT), resulting in modest efficacy. We aimed to evaluate the efficacy and safety of sintilimab plus bevacizumab combined with radiotherapy in HCC with PVTT and to identify prognostic biomarkers. APPROACH AND RESULTS: This open-label, multicenter, single-arm, phase 2 clinical trial was conducted at 3 tertiary hospitals in China. A total of 46 patients with HCC with PVTT were enrolled. All the patients received the first cycle of i.v. sintilimab (200 mg, day 1) plus bevacizumab (15 mg/kg, day 1) within 3 days after enrollment. Radiotherapy (30-50 Gy/10 fractions) was administered after 2 cycles of Sin-Bev. Sin-Bev was disrupted during radiotherapy and resumed 2 weeks after radiotherapy and continued every 3 weeks thereafter until disease progression, unacceptable toxicity, or withdrawal of consent. The primary end point was objective response rate. Patients obtained an objective response rate of 58.7% and a disease control rate of 100%. After a median follow-up time of 26.0 months (95% CI: 24.0-26.0), the median OS was 24.0 months (95% CI: 19.0 to not applicable) and the median progression-free survival was 13.8 months (95% CI: 12.0-21.0), respectively. No unexpected adverse events or treatment-related deaths occurred. Mutations of PCTMD1 were predictive of shorter OS and progression-free survival. CONCLUSIONS: Sintilimab plus bevacizumab combined with radiotherapy provides favorable treatment response and survival outcomes along with an acceptable safety profile in the first-line setting for patients with HCC with PVTT (ClinicalTrials.gov Identifier: NCT05010434).

The oxidized phospholipid PGPC impairs endothelial function by promoting endothelial cell ferroptosis via FABP3.

Ferroptosis is a novel cell death mechanism that is mediated by iron-dependent lipid peroxidation. It may be involved in atherosclerosis development. Products of phospholipid oxidation play a key role in atherosclerosis. 1-palmitoyl-2-glutaroyl-sn-glycero-3-phosphocholine (PGPC) is a phospholipid oxidation product present in atherosclerotic lesions. It remains unclear whether PGPC causes atherosclerosis by inducing endothelial cell ferroptosis. In this study, human umbilical vein endothelial cells (HUVECs) were treated with PGPC. Intracellular levels of ferrous iron, lipid peroxidation, superoxide anions (O2•-), and glutathione were detected, and expression of fatty acid binding protein-3 (FABP3), glutathione peroxidase 4 (GPX4), and CD36 were measured. Additionally, the mitochondrial membrane potential (MMP) was determined. Aortas from C57BL6 mice were isolated for vasodilation testing. Results showed that PGPC increased ferrous iron levels, the production of lipid peroxidation and O2•-, and FABP3 expression. However, PGPC inhibited the expression of GPX4 and glutathione production and destroyed normal MMP. These effects were also blocked by ferrostatin-1, an inhibitor of ferroptosis. FABP3 silencing significantly reversed the effect of PGPC. Furthermore, PGPC stimulated CD36 expression. Conversely, CD36 silencing reversed the effects of PGPC, including PGPC-induced FABP3 expression. Importantly, E06, a direct inhibitor of the oxidized 1-palmitoyl-2-arachidonoyl-phosphatidylcholine IgM natural antibody, inhibited the effects of PGPC. Finally, PGPC impaired endothelium-dependent vasodilation, ferrostatin-1 or FABP3 inhibitors inhibited this impairment. Our data demonstrate that PGPC impairs endothelial function by inducing endothelial cell ferroptosis through the CD36 receptor to increase FABP3 expression. Our findings provide new insights into the mechanisms of atherosclerosis and a therapeutic target for atherosclerosis.

Reprogramming the Intrahepatic Cholangiocarcinoma Immune Microenvironment by Chemotherapy and CTLA-4 Blockade Enhances Anti-PD-1 Therapy.

Intrahepatic cholangiocarcinoma (ICC) has limited therapeutic options and a dismal prognosis. Adding blockade of the anti-programmed cell death protein (PD)-1 pathway to gemcitabine/cisplatin chemotherapy has recently shown efficacy in biliary tract cancers but with low response rates. Here, we studied the effects of anti-cytotoxic T lymphocyte antigen (CTLA)-4 when combined with anti-PD-1 and gemcitabine/cisplatin in orthotopic murine models of ICC. This combination therapy led to substantial survival benefits and reduction of morbidity in two aggressive ICC models that were resistant to immunotherapy alone. Gemcitabine/cisplatin treatment increased tumor-infiltrating lymphocytes and normalized the ICC vessels and, when combined with dual CTLA-4/PD-1 blockade, increased the number of activated CD8+Cxcr3+IFNγ+ T cells. CD8+ T cells were necessary for the therapeutic benefit because the efficacy was compromised when CD8+ T cells were depleted. Expression of Cxcr3 on CD8+ T cells is necessary and sufficient because CD8+ T cells from Cxcr3+/+ but not Cxcr3-/- mice rescued efficacy in T cell‒deficient mice. Finally, rational scheduling of anti-CTLA-4 "priming" with chemotherapy followed by anti-PD-1 therapy achieved equivalent efficacy with reduced overall drug exposure. These data suggest that this combination approach should be clinically tested to overcome resistance to current therapies in ICC patients.

Enhancing the Thermal Stability of Zein Particle-Stabilized Aeratable Emulsions Through Genipin-Protein Cross-Linking and Its Possible Mechanism of Action.

Protein particle-stabilized emulsions often lack thermal stability, impacting their industrial use. This study investigated the effects of genipin (GP)-zein cross-linked particles with varying GP-to-protein weight ratios (0/0.02/0.1:1) on emulsion thermal stability. Enhanced stability was observed at the GP level of 0.1. Heat treatment increased the covalent cross-linking in raw particles and emulsions. Isolated particles from heated emulsions grew in size (micrometer scale) with higher GP levels, unlike heated raw particles (nanoscale). GP-protein cross-linking reduced the droplet-droplet and particle-emulsifier interactions in the heated emulsion. Spectroscopic analysis and electrophoresis revealed that GP-zein cross-linking increased protein structural stability and inhibited nondisulfide and non-GP cross-linking reactions in heated emulsions. The GP-zein bridges between particles at the oil-water interface create strong connections in the particle layer (shell), referred to as "particle-shell locking", enhancing the thermal stability of emulsion significantly. This insight aids the future design of protein-particle-based emulsions, preserving properties like aeratability during thermal processing.

Endothelial extracellular vesicles induce acute lung injury via follistatin-like protein 1.

Cardiopulmonary bypass has been speculated to elicit systemic inflammation to initiate acute lung injury (ALI), including acute respiratory distress syndrome (ARDS), in patients after cardiac surgery. We previously found that post-operative patients showed an increase in endothelial cell-derived extracellular vesicles (eEVs) with components of coagulation and acute inflammatory responses. However, the mechanism underlying the onset of ALI owing to the release of eEVs after cardiopulmonary bypass, remains unclear. Plasma plasminogen-activated inhibitor-1 (PAI-1) and eEV levels were measured in patients with cardiopulmonary bypass. Endothelial cells and mice (C57BL/6, Toll-like receptor 4 knockout (TLR4-/-) and inducible nitric oxide synthase knockout (iNOS-/-)) were challenged with eEVs isolated from PAI-1-stimulated endothelial cells. Plasma PAI-1 and eEVs were remarkably enhanced after cardiopulmonary bypass. Plasma PAI-1 elevation was positively correlated with the increase in eEVs. The increase in plasma PAI-1 and eEV levels was associated with post-operative ARDS. The eEVs derived from PAI-1-stimulated endothelial cells could recognize TLR4 to stimulate a downstream signaling cascade identified as the Janus kinase 2/3 (JAK2/3)-signal transducer and activator of transcription 3 (STAT3)-interferon regulatory factor 1 (IRF-1) pathway, along with iNOS induction, and cytokine/chemokine production in vascular endothelial cells and C57BL/6 mice, ultimately contributing to ALI. ALI could be attenuated by JAK2/3 or STAT3 inhibitors (AG490 or S3I-201, respectively), and was relieved in TLR4-/- and iNOS-/- mice. eEVs activate the TLR4/JAK3/STAT3/IRF-1 signaling pathway to induce ALI/ARDS by delivering follistatin-like protein 1 (FSTL1), and FSTL1 knockdown in eEVs alleviates eEV-induced ALI/ARDS. Our data thus demonstrate that cardiopulmonary bypass may increase plasma PAI-1 levels to induce FSTL1-enriched eEVs, which target the TLR4-mediated JAK2/3/STAT3/IRF-1 signaling cascade and form a positive feedback loop, leading to ALI/ARDS after cardiac surgery. Our findings provide new insight into the molecular mechanisms and therapeutic targets for ALI/ARDS after cardiac surgery.

Generating a multimodal artificial intelligence model to differentiate benign and malignant follicular neoplasms of the thyroid: A proof-of-concept study.

BACKGROUND: Machine learning has been increasingly used to develop algorithms that can improve medical diagnostics and prognostication and has shown promise in improving the classification of thyroid ultrasound images. This proof-of-concept study aims to develop a multimodal machine-learning model to classify follicular carcinoma from adenoma. METHODS: This is a retrospective study of patients with follicular adenoma or carcinoma at a single institution between 2010 and 2022. Demographics, imaging, and perioperative variables were collected. The region of interest was annotated on ultrasound and used to perform radiomics analysis. Imaging features and clinical variables were then used to create a random forest classifier to predict malignancy. Leave-one-out cross-validation was conducted to evaluate classifier performance using the area under the receiver operating characteristic curve. RESULTS: Patients with follicular adenomas (n = 7) and carcinomas (n = 11) with complete imaging and perioperative data were included. A total of 910 features were extracted from each image. The t-distributed stochastic neighbor embedding method reduced the dimension to 2 primary represented components. The random forest classifier achieved an area under the receiver operating characteristic curve of 0.76 (clinical only), 0.29 (image only), and 0.79 (multimodal data). CONCLUSION: Our multimodal machine learning model demonstrates promising results in classifying follicular carcinoma from adenoma. This approach can potentially be applied in future studies to generate models for preoperative differentiation of follicular thyroid neoplasms.

The national, regional, and global impact of glaucoma as reported in the 2019 Global Burden of Disease Study.

Introduction: We investigated the disability-adjusted life years (DALYs) of glaucoma. Methods: The estimated annual percentage change (EAPC) was measured to assess trends in the age-standardized DALY rate from 1990 to 2019. Results: The global age-standardized DALY rate of glaucoma decreased with an EAPC of -1.00. The age-standardized DALY rate decreased least in high-SDI regions. Eastern sub-Saharan Africa had highest age-standardized DALY rate in 2019. At the national level, Mali had the highest age-standardized DALY rate in 2019. Conclusions: Although the global burden of glaucoma has decreased, the burden remain high in regions with low SDI values and in sub-Saharan Africa.

Discovery Viewer (DV): Web-Based Medical AI Model Development Platform and Deployment Hub.

The rapid rise of artificial intelligence (AI) in medicine in the last few years highlights the importance of developing bigger and better systems for data and model sharing. However, the presence of Protected Health Information (PHI) in medical data poses a challenge when it comes to sharing. One potential solution to mitigate the risk of PHI breaches is to exclusively share pre-trained models developed using private datasets. Despite the availability of these pre-trained networks, there remains a need for an adaptable environment to test and fine-tune specific models tailored for clinical tasks. This environment should be open for peer testing, feedback, and continuous model refinement, allowing dynamic model updates that are especially important in the medical field, where diseases and scanning techniques evolve rapidly. In this context, the Discovery Viewer (DV) platform was developed in-house at the Biomedical Engineering and Imaging Institute at Mount Sinai (BMEII) to facilitate the creation and distribution of cutting-edge medical AI models that remain accessible after their development. The all-in-one platform offers a unique environment for non-AI experts to learn, develop, and share their own deep learning (DL) concepts. This paper presents various use cases of the platform, with its primary goal being to demonstrate how DV holds the potential to empower individuals without expertise in AI to create high-performing DL models. We tasked three non-AI experts to develop different musculoskeletal AI projects that encompassed segmentation, regression, and classification tasks. In each project, 80% of the samples were provided with a subset of these samples annotated to aid the volunteers in understanding the expected annotation task. Subsequently, they were responsible for annotating the remaining samples and training their models through the platform's "Training Module". The resulting models were then tested on the separate 20% hold-off dataset to assess their performance. The classification model achieved an accuracy of 0.94, a sensitivity of 0.92, and a specificity of 1. The regression model yielded a mean absolute error of 14.27 pixels. And the segmentation model attained a Dice Score of 0.93, with a sensitivity of 0.9 and a specificity of 0.99. This initiative seeks to broaden the community of medical AI model developers and democratize the access of this technology to all stakeholders. The ultimate goal is to facilitate the transition of medical AI models from research to clinical settings.

A Forest Wildlife Detection Algorithm Based on Improved YOLOv5s.

A forest wildlife detection algorithm based on an improved YOLOv5s network model is proposed to advance forest wildlife monitoring and improve detection accuracy in complex forest environments. This research utilizes a data set from the Hunan Hupingshan National Nature Reserve in China, to which data augmentation and expansion methods are applied to extensively train the proposed model. To enhance the feature extraction ability of the proposed model, a weighted channel stitching method based on channel attention is introduced. The Swin Transformer module is combined with a CNN network to add a Self-Attention mechanism, thus improving the perceptual field for feature extraction. Furthermore, a new loss function (DIOU_Loss) and an adaptive class suppression loss (L_BCE) are adopted to accelerate the model's convergence speed, reduce false detections in confusing categories, and increase its accuracy. When comparing our improved algorithm with the original YOLOv5s network model under the same experimental conditions and data set, significant improvements are observed, in particular, the mean average precision (mAP) is increased from 72.6% to 89.4%, comprising an accuracy improvement of 16.8%. Our improved algorithm also outperforms popular target detection algorithms, including YOLOv5s, YOLOv3, RetinaNet, and Faster-RCNN. Our proposed improvement measures can well address the challenges posed by the low contrast between background and targets, as well as occlusion and overlap, in forest wildlife images captured by trap cameras. These measures provide practical solutions for enhanced forest wildlife protection and facilitate efficient data acquisition.

Foreground Segmentation-Based Density Grading Networks for Crowd Counting.

Estimating object counts within a single image or video frame represents a challenging yet pivotal task in the field of computer vision. Its increasing significance arises from its versatile applications across various domains, including public safety and urban planning. Among the various object counting tasks, crowd counting is particularly notable for its critical role in social security and urban planning. However, intricate backgrounds in images often lead to misidentifications, wherein the complex background is mistaken as the foreground, thereby inflating forecasting errors. Additionally, the uneven distribution of crowd density within the foreground further exacerbates predictive errors of the network. This paper introduces a novel architecture with a three-branch structure aimed at synergistically incorporating hierarchical foreground information and global scale information into density map estimation, thereby achieving more precise counting results. Hierarchical foreground information guides the network to perform distinct operations on regions with varying densities, while global scale information evaluates the overall density level of the image and adjusts the model's global predictions accordingly. We also systematically investigate and compare three potential locations for integrating hierarchical foreground information into the density estimation network, ultimately determining the most effective placement.Through extensive comparative experiments across three datasets, we demonstrate the superior performance of our proposed method.

A critical application of different methods for the vulnerability assessment of shallow aquifers in Zhengzhou City.

Groundwater vulnerability can partially reflect the possibility of groundwater contamination, which is crucial for ensuring human health and a good ecological environment. The current study seeks to assess the groundwater vulnerability of Zhengzhou City by adopting an amended version of the traditional DRASTIC model, i.e., the DRASTICL model, which incorporates land use type indicators. More specifically, the AHP-DRASTICL, entropy-DRASTICL, and AE-DRASTICL models were established by optimizing weights using the analytic hierarchy process (AHP) and entropy weight method. The evaluation results for these five models were divided into five levels: very low, low, medium, high, and very high. Using Spearman's rank correlation coefficient, the nitrate concentration was used to verify the groundwater vulnerability assessment results. The AE-DRASTICL model was found to perform the best, with a Spearman correlation coefficient of 0.78. However, the AHP and entropy weight method effectively improved the accuracy of vulnerability assessment results, making it more suitable for the study area. This study provides important insights to inform the design of strategies to protect groundwater in Zhengzhou.

High-density lipoprotein regulates angiogenesis by long non-coding RNA HDRACA.

Normal high-density lipoprotein (nHDL) can induce angiogenesis in healthy individuals. However, HDL from patients with coronary artery disease undergoes various modifications, becomes dysfunctional (dHDL), and loses its ability to promote angiogenesis. Here, we identified a long non-coding RNA, HDRACA, that is involved in the regulation of angiogenesis by HDL. In this study, we showed that nHDL downregulates the expression of HDRACA in endothelial cells by activating WW domain-containing E3 ubiquitin protein ligase 2, which catalyzes the ubiquitination and subsequent degradation of its transcription factor, Kruppel-like factor 5, via sphingosine 1-phosphate (S1P) receptor 1. In contrast, dHDL with lower levels of S1P than nHDL were much less effective in decreasing the expression of HDRACA. HDRACA was able to bind to Ras-interacting protein 1 (RAIN) to hinder the interaction between RAIN and vigilin, which led to an increase in the binding between the vigilin protein and proliferating cell nuclear antigen (PCNA) mRNA, resulting in a decrease in the expression of PCNA and inhibition of angiogenesis. The expression of human HDRACA in a hindlimb ischemia mouse model inhibited the recovery of angiogenesis. Taken together, these findings suggest that HDRACA is involved in the HDL regulation of angiogenesis, which nHDL inhibits the expression of HDRACA to induce angiogenesis, and that dHDL is much less effective in inhibiting HDRACA expression, which provides an explanation for the decreased ability of dHDL to stimulate angiogenesis.

Deep Learning for Automated Measurement of Patellofemoral Anatomic Landmarks.

BACKGROUND: Patellofemoral anatomy has not been well characterized. Applying deep learning to automatically measure knee anatomy can provide a better understanding of anatomy, which can be a key factor in improving outcomes. METHODS: 483 total patients with knee CT imaging (April 2017-May 2022) from 6 centers were selected from a cohort scheduled for knee arthroplasty and a cohort with healthy knee anatomy. A total of 7 patellofemoral landmarks were annotated on 14,652 images and approved by a senior musculoskeletal radiologist. A two-stage deep learning model was trained to predict landmark coordinates using a modified ResNet50 architecture initialized with self-supervised learning pretrained weights on RadImageNet. Landmark predictions were evaluated with mean absolute error, and derived patellofemoral measurements were analyzed with Bland-Altman plots. Statistical significance of measurements was assessed by paired t-tests. RESULTS: Mean absolute error between predicted and ground truth landmark coordinates was 0.20/0.26 cm in the healthy/arthroplasty cohort. Four knee parameters were calculated, including transepicondylar axis length, transepicondylar-posterior femur axis angle, trochlear medial asymmetry, and sulcus angle. There were no statistically significant parameter differences (p > 0.05) between predicted and ground truth measurements in both cohorts, except for the healthy cohort sulcus angle. CONCLUSION: Our model accurately identifies key trochlear landmarks with ~0.20-0.26 cm accuracy and produces human-comparable measurements on both healthy and pathological knees. This work represents the first deep learning regression model for automated patellofemoral annotation trained on both physiologic and pathologic CT imaging at this scale. This novel model can enhance our ability to analyze the anatomy of the patellofemoral compartment at scale.