搜索结果：Keep

Many organisations collect sensitive data that cannot be freely shared. Hospitals store brain magnetic resonance imaging (MRI) scans on internal servers; banks keep transaction records behind strict firewalls; agricultural services retain crop images in isolated repositories. Federated learning (FL) allows models to be trained without centralising raw data, yet most existing systems address a single domain and offer limited insight into model behaviour and provenance over time. BlockFedX is a cross-domain federated learning system designed to address three simultaneous tasks: credit card fraud detection on tabular data, brain tumour detection on MRI images, and plant disease recognition on leaf images. These three domains were deliberately selected because they represent the principal data modalities in real-world privacy-sensitive deployments-structured tabular records, greyscale medical images, and colour natural images-and because public benchmark datasets exist for all three, enabling reproducible evaluation. The system uses a shared backbone that is updated only where model layers have compatible tensor shapes, while domain-specific output layers remain local at each client. Explanations are computed at the clients using SHAP feature-attribution for tabular data and Grad-CAM visual heatmaps for images; the server receives only compact statistical summaries. The server also applies a distance-based anomaly test on client updates and records model hashes, explanation summaries, and anomaly flags in a hash-chained ledger. Experiments on three public datasets under non-identical client data distributions show that BlockFedX achieves an average fraud-detection F1-score of 0.92, 74.32% mean validation accuracy on BrainMRI, and 77% test accuracy on PlantVillage, while keeping all raw data local. These results are below strong centralised baselines, as expected under compact models and non-IID splits, but the system simultaneously provides three properties rarely combined in prior work: cross-domain federated training via a shape-safe backbone, client-side explanations integrated into the learning loop, and a lightweight tamper-evident record of model evolution across rounds.

Forgiveness can help preserve valuable relationships after harm, but it is also a self-regulatory challenge: victims must inhibit retaliatory impulses, shift away from hostile appraisals, and keep longer-term relational goals in mind. However, little is known about the self-regulatory capacities associated with individual differences in young children's responses to transgressions. Japanese 4- and 5-year-olds viewed two live puppet vignettes depicting severe physical transgressions without apology or repair and responded from the victim's perspective. After each vignette, children selected an immediate outward response toward the transgressor: an explicit conciliatory response, sad protest, angry protest, or silence. They also reported their anger and retaliatory desire. Children completed three executive-function tasks indexing cognitive flexibility (Dimensional Change Card Sort), delay of gratification, and reward-punishment decision making (child gambling task). Results indicated that executive-function components were differentially associated with children's post-transgression responses: higher cognitive flexibility was associated with a lower likelihood of selecting the explicit conciliatory response and a higher likelihood of angry protest, suggesting that cognitive flexibility may support boundary-setting protest rather than immediate conciliation in high-severity, no-repair contexts. Better reward-punishment decision making was linked to weaker retaliatory desire, whereas delay of gratification was not uniquely associated with anger or retaliation. Together, these findings refine developmental accounts of forgiveness by suggesting that preschoolers' responses to harm are not governed by a single regulatory capacity: cognitive flexibility tracked outward protest versus conciliation, whereas reward-punishment decision making tracked retaliatory motivation.

Conversations are an important part of our social lives, although for people with hearing impairment (HI), conversations can pose a considerable challenge and can often lead to miscommunications. In conversations recorded from 25 groups consisting of one HI interlocutor and two normal hearing (NH) interlocutors, all miscommunications were identified to evaluate how HI, background noise level, and hearing-aid signal processing affected miscommunications. A subset of miscommunications is so-called other-initiated repairs (OIRs) where one interlocutor signals a communication breakdown, using unspecific open requests (e.g., "What?") or increasingly more specific restricted requests or offers (e.g., "Who?" or "Did you say blue?"). An "open request" signals a problem without specifying what is misheard/understood, while the restricted request specifies what part of the sentence is not heard, and finally, the most specific "restricted offer" is asking the conversation partner to confirm the supposed word/phrase. With increasing communication difficulty, open OIRs are expected to be more frequently used due to poorer speech understanding. The results showed that HI interlocutors generally had more miscommunications and that, across interlocutors, open OIRs were mostly used. At low noise levels, the HI interlocutor had fewer miscommunications and used more specific OIRs when wearing a hearing aid relative to being unaided. At the high noise level, all interlocutors had more miscommunications. When listening to directional sound processing (dir) at the high noise level, the HI interlocutor produced fewer open OIRs. It is interesting that the NH interlocutors were also affected by having more miscommunications and more open OIRs when the HI interlocutor listened to dir. The speech levels of the HI interlocutors were reduced in dir, and a relationship was found between the HI speech levels and the number of OIRs made by the NH interlocutors. Our results highlight how sensitive the number and nature of miscommunications and OIRs in conversations are, not only to the sound environment, but also to the experience of a single (HI) interlocutor, even in triadic interactions, where only two interlocutors are strictly needed to keep the conversation going.

Computational methods are central to the life sciences. The rapid growth and diversification of software tools and databases make it difficult to find, compare, and reuse methods for a given task. bio.tools is a community-driven registry designed to improve the visibility of research software and allow researchers to simplify access to the software ecosystem through structured, interoperable, and accessible metadata. Tools are annotated using the EDAM ontology and additional controlled vocabularies, enabling users to search and filter by scientific topics, operations, input/output data types, and data formats. bio.tools supports interactive exploration via rich tool landing pages and provides programmatic access through a documented API for search, retrieval, and registry statistics. The registry has expanded to almost 33,000 annotated tools through the combined contributions of thousands of community members and semi-automated literature mining that keep the registry up to date. Recent improvements to the registry include machine-assisted scoring to prioritise curator review, and consolidation of both its standards stack and software architecture. bio.tools has also become a foundational upstream metadata source that is reused by other services in the ELIXIR Research Software Ecosystem and beyond, to support synchronisation, cross-linking, and additional downstream services. bio.tools is freely available at https://bio.tools.

Radio frequency identification (RFID) technology was evaluated as a method for electronically tracking human remains and associated identifying information during mass fatality incidents. Conventional tags are prone to physical damage and failure, which can compromise identification processes and hinder the repatriation of bodies. Nine types of uniquely numbered RFID tags were deployed in decompositional and burial environments to assess the survivability of different types of RFID tags buried, inserted, and attached to pig carcasses. The best method of insertion or attachment to keep tags in situ, and whether physical writing in pencil or permanent marker remained on the tags' surfaces. In a series of four studies of whole carcasses, pieces of tissue, as human proxies, and burials in soil, all but two tags were determined to be viable throughout the decomposition process. All tags displayed movement issues, so bone attachment sites are recommended. Handwritten labels on the tags prepared under dry conditions and when protected from decompositional fluids held up better than writing interacting with the fluids. We found the most useful tags were the LF Yellow Ear Tag and the UHF White Tag.

The generation of heat during osteotomy preparation is a significant risk factor for thermal osteonecrosis, which may lead to implant failure. While static surgical guides enhance placement accuracy, they can physically obstruct the flow of external coolant, paradoxically increasing the risk for thermal injury. (1) Review of thermal necrosis and critical temperatures during osteotomy preparation. Advantages and disadvantages of static guided surgery. (2) Design and fabrication of static surgical guides with direct coolant channels to the osteotomy site. Variations include: Hollow-Guide, internal irrigation channels, external irrigation channels, and mesh-based guide fabrication. A new, simplified workflow is proposed to design and fabricate surgical guides for implant placement with the delivery of coolant directly to the osteotomy site with no obstruction and decreased risk of thermal injury and implant failure. This report details a novel approach to mitigating thermal injury during guided dental implant surgery. The application of a patient-specific, three-dimensional-printed surgical guide featuring an integrated Hollow-Guide design bypasses conventional coolant obstruction by delivering sterile saline directly to the osteotomy site via segmentation and selective hollowing. A simple workflow is described to integrate multidirectional irrigation to the surgical site and reduce the risk of thermal injury, subsequent osteonecrosis, and implant failure. While traditional static guides serve as a passive positioning jig, the Hollow-Guide acts as an active therapeutic device during implant preparation and placement. Static surgical guides significantly improve implant placement accuracy, but impede traditional delivery methods of irrigation and coolants to the surgical site. Maintenance of bone temperatures below a critical threshold is necessary to prevent thermal necrosis and implant failure. A digital workflow has been proposed to allow for irrigation directly at the osteotomy site without the need for complex internal or external irrigation channels. Preparing the jawbone for dental implants generates significant heat from drilling. If the bone gets too hot, the tissue can die, ultimately leading to failure and loss of the implant. Dentists often use three-dimensional-printed surgical guides to ensure implants are placed in the correct anatomical position, but these templates can physically block external irrigation from reaching the drill site, increasing the risk of thermal damage. This article introduces a new design technique called the "Hollow-Guide." This method modifies the surgical guide to include built-in channels that allow sterile water or saline to flow directly to the bone during drilling. By solving the problem of blocked water flow, this new design helps keep the bone cool while maintaining the high accuracy of guided surgery. This simple workflow transforms the guide from a passive positioning tool into an active safety device that helps prevent bone damage and ensures better long-term success for dental implants.

The circular economy advocates for a fundamental shift, focusing on the reuse, recycling, and valorization of waste to keep materials in use for as long as possible, thereby reducing the need for new resources and minimizing waste. In this context, the concept of the Blue Circular Economy emerges, emphasizing the preservation and sustainable use of marine and coastal resources. This paper presents and explains a dataset collected through a survey administered in two regions: Algarve-Portugal, and Oslo-Norway. The aim was to address the existing data gap on the production, consumption, and disposal of bivalve shells from aquaculture, as well as to assess stakeholders' perceptions of the reuse and circularity of products. In Portugal, data collection focused on analyzing the mollusk supply chain in the Algarve region and assessing the potential for oyster shell reuse within the country. In Norway, data collection targeted the construction industry and academia to understand their perspectives on the reuse of shells in building applications. The dataset comprises responses from 112 voluntary participants from different countries, categorized into three main groups. The questions addressed the quantity of bivalves produced and consumed, shell disposal practices, perceptions regarding the potential reuse of this waste, and the level of awareness of public policies promoting circular economy strategies. Upon completion of the survey, respondents were given the opportunity to receive a summary of the study findings by voluntarily providing their e-mail address in a designated field of the survey. The dataset integrates quantitative data from online and local face-to-face surveys across different groups. This dataset provides a valuable resource for stakeholders in the aquaculture and construction sectors seeking to explore the potential of this biomaterial to valorise waste into innovative products, as well as for public authorities aiming to support the transition towards a circular economy.

Timely and contextually relevant evidence is essential for decision-making in policy domains, especially during crisis or outbreaks like the COVID-19 pandemic. While systematic reviews provide methodological rigor, their cost and duration limit their policy utility. Rapid reviews offer a pragmatic alternative but are typically static and quickly become outdated as new studies, grey literature, and local knowledge emerge, leading to duplication of effort and losing track of policy decisions and use of evidence. This paper presents a pragmatic framework for converting rapid reviews into living evidence synthesis (LES) systems within policy contexts, especially in low- and middle-income countries (LMICs). Drawing on the experience of eBASE Africa, the framework demonstrates how LES functions can be incrementally embedded within conventional rapid-review workflows. It integrates PRISMA-aligned methods with continuous evidence surveillance across academic literature, grey literature, and non-traditional evidence sources such as social media and Indigenous Ways of Knowing, treating these as structured and systematically incorporated evidence streams, supported by responsible use of artificial intelligence and digital tools with human oversight. The framework reconceptualizes rapid reviews as entry points to living evidence systems, offering a scalable pathway to improve efficiency, continuity, and policy relevance, particularly in LMIC settings. Good decisions in health, education, and other policy areas depend on having the right information at the right time. This became especially clear during the COVID-19 pandemic, when policymakers needed quick answers to urgent questions. Researchers often provide this information through systematic reviews, which carefully collect and analyze existing studies. While these reviews are thorough and reliable, they usually take a long time to complete and can be expensive. This makes them less useful when decisions need to be made quickly. To respond to urgent needs, many organizations use rapid reviews. These are faster versions of systematic reviews that provide timely insights for decision-makers. However, rapid reviews have an important limitation: they are usually done once and then left unchanged. As new studies, reports, and local experiences become available, the findings of these reviews can quickly become outdated. This means that policymakers may rely on information that no longer reflects the latest evidence. It can also lead to duplication of effort, as new reviews are started from scratch instead of building on existing work. This paper explores a practical way to solve this problem by transforming rapid reviews into what are called “living evidence synthesis” systems. A living evidence system is one that is continuously updated as new information becomes available. Instead of treating a review as a one-time product, it becomes an ongoing process that evolves over time. Drawing on the experience of eBASE Africa, this study presents a step-by-step framework for making this transition, particularly in low- and middle-income countries. These settings often face resource constraints, making it even more important to use time and funding efficiently. The framework shows how organizations can gradually build living evidence systems into their existing rapid review processes, rather than starting from scratch. One key feature of this approach is the use of continuous evidence monitoring. This means regularly searching for and incorporating new information from a wide range of sources. In addition to academic research, this includes grey literature such as policy reports, as well as non-traditional sources like social media and Indigenous knowledge. These sources are treated as valuable forms of evidence and are systematically included in the review process. The framework also emphasizes the careful use of digital tools and artificial intelligence to support this work. These tools can help identify new studies, organize information, and reduce the workload for researchers. However, human oversight remains essential to ensure that the information is accurate, relevant, and ethically used. Another important aspect is maintaining transparency and methodological rigor. The approach aligns with established standards such as PRISMA, ensuring that the process remains systematic and credible even as it becomes more flexible and continuous. Overall, this paper argues that rapid reviews should not be seen as final products, but as starting points for ongoing learning. By turning them into living systems, it is possible to keep evidence up to date, reduce duplication, and make better use of available resources. This approach can improve the relevance and usefulness of evidence for policymakers, especially in fast-changing situations and resource-limited settings. Thus, the proposed framework offers a practical and scalable way to ensure that evidence remains current, inclusive, and responsive to real-world needs. It highlights how combining traditional research methods with continuous updating and diverse knowledge sources can lead to better-informed decisions and more effective policies.

BackgroundCaregivers of children with tic disorders experience substantial psychological and social burdens; however, existing research is predominantly cross-sectional, and their lived emotional experiences remain insufficiently explored through qualitative approaches.ObjectiveThis study aimed to explore the emotional experiences of primary caregivers of children with tic disorders throughout the caregiving process.MethodsA descriptive phenomenological design was employed. Semi-structured in-depth interviews were conducted with 13 primary caregivers recruited through purposive sampling. Data collection continued until thematic saturation was reached. Interview transcripts were analyzed using Colaizzi's seven-step method.ResultsThree main themes emerged: multiple caregiving burdens, emotional and psychological distress, and caregiving challenges and coping strategies. Caregivers reported persistent negative emotions, including anxiety, guilt, and emotional exhaustion, which were intensified by financial strain, family conflict, and limited access to professional medical services.ConclusionCaregivers of children with tic disorders experience considerable caregiving burdens and psychological distress, which may hinder both caregiver well-being and the child's rehabilitation. Strengthening psychological support systems and improving access to professional healthcare resources are essential to address caregivers' multifaceted needs. Why was the study done? Researchers conducted this study to map the emotional “lived experiences” of parents in China caring for children with tic disorders. While previous studies often focus on high-level numbers or medical outcomes, they frequently overlook the deeper, daily stories of these families. The goal was to uncover hidden emotional needs so that healthcare professionals can develop personalized support strategies. What did the researchers do? The researchers used a “descriptive phenomenological” design to focus on the emotions parents consciously perceive and express. They conducted in-depth, one-on-one interviews with 13 primary caregivers at a clinic in Shandong, China. To ensure authenticity, they analyzed these conversations using the systematic Colaizzi’s seven-step method to identify key themes. What did the researchers find? The study found that caregiver stress falls into three categories: multiple life burdens, emotional distress, and coping challenges. Parents face “social” embarrassment when symptoms are misunderstood and “personal” guilt, often blaming their own parenting for the disorder. Many caregivers also suppress their own anxiety to keep their child calm, which frequently leads to total emotional exhaustion. What do the findings mean? These findings suggest that medical care must move beyond just treating the child to supporting the whole family. There is an urgent need for professional psychological support for parents and better public education to reduce the social stigma that makes caregiving so much harder.

Joy is an underexplored dimension in autism research, particularly within family life. This qualitative study examined how autistic adolescents and young adults (AYAs) and their families experience and co-create joy. Using a phenomenological design and reflexive thematic analysis, we analyzed 2-week family joy journals from six families (17 participants: six AYAs, two siblings, and nine caregivers). Journal entries revealed joy as a relational process embedded in everyday routines, creative engagement, and shared experiences. Three overarching themes emerged: Experiences of Everyday Joy (sensory pleasures, creativity, and togetherness), the Relational Nature of Joy (co-experienced joy and creativity), and the Layered Nature of Joy (complex intersections of sensory, emotional, and social dimensions). Families described joy as spontaneous yet deeply meaningful, often heightened by the journaling process. Findings challenge deficit-based narratives by highlighting emotional richness and enjoyment experienced within autistic AYAs and their families. Journaling may hold promise as a supportive routine by drawing attention to protective relational processes such as empathy, reciprocity, and connection. This study highlights the significance of acknowledging joy as a potential catalyst for well-being and flourishing in autism research and practice.Lay AbstractJoy is an essential part of life, yet it is rarely studied in autism research. This study explored how autistic adolescents and young adults (AYAs) and their families experience joy in everyday life. We invited six families to keep a "joy journal" for 2 weeks, asking them to write down what brought them joy each day. Seventeen people participated, including six autistic AYAs, two siblings, and nine caregivers. The family joy journals revealed that joy often stemmed from simple, familiar activities, such as reading, listening to music, sharing meals, or spending time with pets. Creative activities, such as drawing, dancing, and playing games, were also familiar sources of joy. Families described joy as something that happens together, not just alone. Many caregivers wrote about feeling joy when they saw their child happy or proud, and AYAs often mentioned enjoying time with family and friends. Three main themes emerged: Experiences of Everyday Joy, the Relational Nature of Joy, and Layered Experiences of Joy, in which sensory pleasure, emotional connection, and shared routines combine. Families reported that the journaling process helped them notice and appreciate joyful moments more frequently. These findings challenge stereotypes that autistic people lack positive emotions. Instead, they show that joy is abundant and deeply relational in autistic AYAs and their families. Recognizing and nurturing joy can strengthen family bonds and support well-being. Simple practices like joy journaling may help families focus on their strengths and create more supportive environments.

Determining an optimal treatment strategy after tyrosine kinase inhibitor (TKI) failure remains challenging in oncogene-driven non-small cell lung cancer (NSCLC). We aimed to evaluate and compare the efficacy and safety of atezolizumab, bevacizumab, carboplatin, and paclitaxel (ABCP) versus pemetrexed plus carboplatin or cisplatin (PC) in patients with progressive NSCLC harboring EGFR, ALK, or ROS1 alterations after TKI failure. Multicenter retrospective study. We analyzed 114 patients with NSCLC treated with either ABCP (n&#x2009;=&#x2009;48) or PC (n&#x2009;=&#x2009;66) post-TKI failure between November 2016 and July 2023. Treatment response, progression-free survival (PFS), overall survival (OS), and safety profiles were assessed. Among 106 evaluable participants, the ABCP arm demonstrated a higher response rate (50.0%) than the PC arm (35.0%). During the median follow-up of 29.9&#x2009;months, PFS was significantly prolonged in the ABCP arm compared with that in the PC arm (7.3 vs 3.0&#x2009;months, respectively; hazard ratio (HR) 0.643; p&#x2009;=&#x2009;0.026), whereas OS showed no significant difference (p&#x2009;=&#x2009;0.165). Subgroup analysis revealed notable improvements in PFS (7.2 vs 2.1&#x2009;months, HR 0.235; p&#x2009;<&#x2009;0.001) and OS (11.0 vs 4.1&#x2009;months, HR 0.418; p&#x2009;=&#x2009;0.001) in patients with three or more metastatic sites. ABCP significantly improved PFS compared with PC in patients with NSCLC that progressed despite prior TKI therapy. An OS benefit was observed in patients with three or more metastatic sites. Combination of atezolizumab, bevacizumab, carboplatin, and paclitaxel (ABCP) improves progression-free survival compared to pemetrexed and carboplatin (PC) in patients with non-small cell lung cancer after targeted therapy fails. Why was this study done? Many patients with non-small cell lung cancer (NSCLC) have specific genetic alterations (mutations like EGFR, ALK, or ROS1). The standard first line treatment for this population is targeted therapy. However, many patients discontinue these drugs because cancer eventually gains resistance to the target therapy. There are unmet needs for patients whose target therapies are exhausted since there are no guideline for subsequent treatment. What did the researchers do? This study compared the effectiveness of two different chemotherapy combinations for patients whose cancer grew after taking targeted pills. The researchers retrospectively analyzed data from 114 patients treated between 2016 and 2023. The patients received either: ABCP: A four-drug combination of atezolizumab (immunotherapy), bevacizumab (anti-VEGF therapy), paclitaxel, and carboplatin (cytotoxic chemotherapy) PC: A two-drug cytotoxic chemotherapy regimen (pemetrexed plus carboplatin). What were the results? The study found that the four-drug combination (ABCP) was more effective at controlling the cancer: Delayed Growth: On average, patients on ABCP went 7.3&#x2009;months before their cancer started growing again, compared to only 3.0&#x2009;months for those on PC. Tumor Shrinkage: Tumors shrank in 50% of patients taking ABCP, compared to 35% of patients taking PC. Survival: While the overall survival time was similar for both groups, patients with more widespread cancer (spread to 3 or more sites) lived significantly longer if they received the ABCP treatment. What do these findings mean? For patients with NSCLC whose targeted therapy has been exhausted, using the ABCP combination appears to keep the cancer under control significantly longer than PC. It may be a particularly good option for patients whose cancer has spread to multiple parts of the body.

Bridge decks are continuously subjected to high environmental exposure, traffic loading, and material aging, leading to progressive delamination which can negatively affect structural integrity and public safety. More specifically, subsurface delamination of concrete and corroded steel reinforcement must be repaired to keep the decks operational. Among non-destructive evaluation techniques, Ground-Penetrating Radar (GPR) and Infrared Thermography (IRT) offer complementary capabilities for detecting subsurface and near-surface defects; however, effective GPR-IRT data fusion remains challenging due to fundamental differences in sensing principles, spatial resolution and sensitivity. This study introduces a Physics-Enhanced Multi-Modal Fusion (PE-MMF) framework that integrates GPR and IRT data to improve delamination detection in reinforced concrete bridge decks. The proposed approach leverages transfer learning, cross-modal attention mechanisms, and gated fusion to enable robust learning from heterogeneous sensor inputs. Furthermore, a systematic feature selection protocol is integrated to identify physically meaningful indicators that remain consistent across different bridges, enhancing generalization capability. The framework is trained and validated using the publicly available SDNET2021 dataset, comprising co-registered GPR and IRT measurements from five in-service bridge decks with verified delamination ground truth. Results demonstrate substantial performance improvements, with average F1-score gains of up to 55% over IRT-based methods and 25% over GPR-based methods across all tested bridges. Comparative analysis against state-of-the-art methods confirmed the superior generalization capability of the proposed multi-modal approach over single-modality approaches. The findings highlight the potential of deep learning-based sensor fusion as a scalable and data-efficient decision-support tool to prioritize regions for detailed physical investigation during long-term infrastructure monitoring.

Bryopsidales green macroalgae can induce non-photochemical quenching (NPQ) only slowly, presumably due to the lack of the proton gradient-induced component (qE) of NPQ and xanthophyll cycle. Here, two morphologically rather similar siphonous macroalgae, a Bryopsidales alga Bryopsis sp. and a Dasycladales alga Acetabularia acetabulum, latter of which is capable of qE, were given high light treatments with both constant and fluctuating intensity. No differences in the rate of photoinhibition of Photosystem II (PSII), estimated with the chlorophyll a fluorescence parameter FV/FM, in the absence or presence of lincomycin, were observed between constant and fluctuating light, nor between the two algae. Bryopsis sp. showed slower PSII recovery than A. acetabulum, possibly reflecting a regulatory response rather than increased oxidative stress, as the recovery rates increased with increasing amounts of PSII photoinhibition in both algae. In Bryopsis sp., however, high light treatments led to decreased electron transfer rates, estimated by both chlorophyll a fluorescence and net oxygen production, whereas a stimulation was observed in A. acetabulum. Nigericin, which prevents the formation of qE, increased photoinhibition in A. acetabulum but not in Bryopsis sp. Microoxic conditions as well as inhibitors of plastid terminal oxidase and mitochondrial respiration, on the other hand, enhanced photoinhibition only in Bryopsis sp., suggesting that, in the absence of qE, oxygen-dependent pathways (including flavodiiron proteins) are important for photoprotection. Near-infra-red absorption measurements suggest decreased Photosystem I (PSI) donor side limitation in Bryopsis sp., compared to A. acetabulum, and a lower capacity to keep P700 oxidised.

To identify a rare but potentially life-threatening cause of delayed postpartum hemorrhage. A 36-year-old woman, gravida 1, para 1, had just underwent cesarean section at a local clinic and the immediate postpartum recovery was well. At two months postpartum, following strenuous exercise, two sudden episodes of massive vaginal bleeding occurred, accompanied by tachycardia, blood pressure drop to 81/50 mmHg and anemia of 8.7 g/dl in hemoglobin. Transvaginal ultrasonography and computed tomography revealed a 1.6 cm pulsating mass near the left uterine artery. After shared decision-making and fertility counseling following another episode of bleeding, transcatheter arterial embolization was performed, where the left uterine artery pseudoaneurysm was successfully identified and occluded using a combination of Gelfoam and coil. She was discharged on post-procedure day 2. Care providers must keep in mind the diagnosis of pseudoaneurysm in the postpartum stage when a pulsating mass is detected in the uterus.

Accurate prediction of peri-operative cardiac complications is critical to optimise pre-operative decision-making. Traditional risk prediction scores, such as the Revised Cardiac Risk Index, show only modest discrimination. Machine learning can model complex, non-linear relationships but their predictive performance compared with traditional scores remains unclear. We performed a systematic review and Bayesian network meta-analysis. The primary outcome was postoperative adverse cardiac events following non-cardiac surgery. Prediction models were assessed relative to the Revised Cardiac Risk Index. As many studies evaluated multiple versions of each model type, the highest performing ('best version') and lowest performing ('worst version') results were analysed. Models were ranked using the surface under the cumulative ranking curve (SUCRA). Thirteen studies evaluating 54 models and 927,113 patients were included. Machine learning approaches generally outperformed traditional risk scores. Automated machine learning ranked highest (SUCRA 96.6) showed the greatest improvement in the best version analysis (mean difference (MD) 0.28 (95%CrI 0.16-0.40)) and remained superior in the sensitivity analysis (MD 0.30 (95%CrI 0.14-0.45)). Gradient boosting models showed superior performance over the Revised Cardiac Risk Index across analysis (best version: MD 0.20 (95%CrI 0.14-0.26), worst version: MD 0.18 (95%CrI 0.12-0.25), SUCRA 82.4). The Gupta Perioperative Risk for Myocardial Infarction or Cardiac Arrest score outperformed the Revised Cardiac Risk Index in the best version analysis (MD 0.16 (95%CrI 0.01-0.32)). Between-study heterogeneity was low. None of the included studies externally validated their machine learning models and only six were judged to be at low risk of bias. Most machine learning models showed better discrimination than traditional risk scores, with automated machine learning and gradient boosting models ranking highest. However, study quality, calibration reporting and absence of external validation limit immediate clinical adoption. Prospective, multicentre evaluation is required before integration of these models into peri-operative practice. We looked at many studies that tried to predict heart problems after surgery. We compared older scoring tools, like the Revised Cardiac Risk Index, with newer computer&#x2010;based methods called machine learning. We checked which ones were better at predicting who might have problems after surgery. Doctors need good tools to predict heart problems before surgery so they can keep patients safe and make the best plans. Older tools are helpful but not always very accurate. We wanted to see if newer computer methods could do a better job. We found that machine learning methods were usually better at predicting heart problems than the older tools. Some types, like automated machine learning and gradient boosting, worked the best. However, many of these newer models have not been tested enough in real hospitals yet. This means more research is needed before doctors can fully rely on them in everyday care.

Inflammatory bowel disease (IBD) is global chronical colonic phlogosis closely related to immunity system dysfunction. IBD includes Crohn's disease (CD) and ulcerative colitis (UC). However, the current treatments for UC still have significant limitations such as insufficient efficacy, obvious side effects and large individual differences. As a natural pigment protein extracted from dry Spirulina platensis powder, phycocyanin (PC) attracted wide focus due to significant antioxidant and anti-inflammatory abilities. However, the role and value of PC in treatment of UC have not been fully studied. This paper aimed to elucidate the effects of Spirulina PC on UC from multiple aspects. This paper revealed that PC treatment could alleviate consequences of colonitis (intestinal mucosal atrophy, loss of weight, the increase in organ indices). After PC treatment, the contents of proinflammatory factors including interleukin-1&#x3b2; (IL-1&#x3b2;), tumor necrosis factor-&#x3b1; (TNF-&#x3b1;) and interleukin-6 (IL-6) reduced and the contents of anti-inflammatory factor interleukin-10 (IL-10) increased in colitis mice. Oxidative stress was regulated by PC. After PC treatment, enzyme activities of catalase (CAT), total superoxide dismutase (T-SOD) and glutathione peroxidase (GSH-Px) increased apparently, levels of myeloperoxidase (MPO) and malondialdehyde (MDA) reduced significantly in colitis mice. PC elevated levels of tight junction proteins including Occludin, zona occludens-1 (ZO-1), Claudin-1, and played positive role to keep integrity of mouse colonic mucosa. Microbial composition and metabolites analysis of colonic contents revealed that PC improved ratio of beneficent germs and reduced ratio of harmful germs in colitis mice. The contents of beneficial metabolites increased in colitis mice, these beneficial metabolites involved in metabolism of amino acids, lipids, carbohydrates, nucleotides and secondary metabolites. This study suggested that PC had great potential to prevent and alleviate UC as dietary supplement.