Surgical Outcomes in Stage IV Pancreatic Cancer with Liver Metastasis Current Evidence and Future Directions: A Systematic Review and Meta-Analysis of Surgical Resection

BACKGROUND/OBJECTIVES

There is increasing evidence that a subset of patients with stage IV pancreatic ductal adenocarcinoma (PDAC) and liver-only metastasis may benefit from surgical resection of both the primary tumor and metastatic lesions.

METHODS

A meta-analysis and systematic review were conducted in patients with stage IV PDAC and hepatic-only metastasis. A comprehensive literature search (January 2015-June 2023) was performed using PubMed with keywords including "pancreatic cancer", "oligometastatic", and "surgery".

RESULTS

Sixteen articles were included in the final review and characterized based on patient selection factors and prognostic indicators. Seven studies reported hazard ratios (HRs) or Kaplan-Meier curves for survival in synchronous resected cohorts versus chemotherapy/palliation alone, which indicated a statistically significant survival benefit in the resection cohorts (pooled HR: 0.41, 95% CI: 0.31-0.53, p < 0.01). Prognostic indicators for synchronous and metachronous resection included lower pre-operative CA19-9, negative margin status of the primary tumor, moderate-to-well-differentiated tumors (grades I-II), and receiving pre-operative chemotherapy.

CONCLUSIONS

Surgical/ablation selection factors are evolving, with priorities on (1) response to induction chemotherapy, (2) ability to achieve R0 resection, and (3) minimally invasive approaches remaining critical to optimal patient selection. Standardized radiologic and tumor marker evaluation and response to neoadjuvant therapy and optimizing performance status are critical to improved outcomes.

Promises and Pitfalls of Internet Search Data in Mental Health: Critical Review

Unlabelled

The internet is now integral to everyday life, and users' web-based search data could be of strategic importance in mental health care. As shown by previous studies, internet searches may provide valuable insights into an individual's mental state and could be of great value in early identification and helping in pathways to care. Internet search data can potentially provide real-time identification (eg, alert mechanisms for timely interventions). In this paper, we discuss the various problems related to the use of these data in research and clinical practice, including privacy concerns, integration with clinical information, and technical limitations. We also propose solutions to address these issues and provide possible future directions.

Predicting preterm birth using machine learning methods

Preterm birth is a significant public health concern, given its correlation with neonatal mortality and morbidity. The aetiology of preterm birth is complex and multifactorial. The objective of this study was to develop and compare machine learning models for predicting the risk of preterm birth. Data were collected from 50 patients in a maternity ward, with an analysis performed based on the timing of delivery (preterm vs. term). The applicability of XGBoost, CatBoost, logistic regression, support vector machines (SVM), and decision trees for predicting preterm delivery was evaluated through training. The linear SVM with boosted parameters demonstrated the highest performance, achieving an accuracy of 82%, precision of 83%, recall of 86%, and an F1-score of 84%. The logistic regression model, also boosted, demonstrated comparable performance to the linear SVM, with similar accuracy (80%), precision (82%), recall (82%), and F1-score (82%). The performance of other models, including decision trees and more complex algorithms, was inferior, which is likely attributable to the limited dataset and the number of parameters involved. In particular, machine learning models, most notably the linear SVM, can be effectively employed to assess the risk of preterm birth. The findings indicate that the linear SVM model exhibits the greatest efficacy among the tested models.

Measuring water pollution effects on antimicrobial resistance through explainable artificial intelligence

Antimicrobial resistance refers to the ability of pathogens to develop resistance to drugs designed to eliminate them, making the infections they cause more difficult to treat and increasing the likelihood of disease diffusion and mortality. As such, antimicrobial resistance is considered as one of the most significant and universal challenges to both health and society, as well as the environment. In our research, we employ the explainable artificial intelligence paradigm to identify the factors that most affect the onset of antimicrobial resistance in diversified territorial contexts, which can vary widely from each other in terms of climatic, economic and social conditions. Specifically, we employ a large set of indicators identified through the One Health framework to predict, at the country level, mortality resulting from antimicrobial resistance related to Acinetobacter baumannii, Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Streptococcus pneumoniae. The analysis reveals the outstanding importance of indicators related to water accessibility and quality in determining mortality due to antimicrobial resistance to the considered pathogens across countries, providing perspective as a potential tool for decision support and monitoring.

Explainable Artificial Intelligence Models for Predicting Depression Based on Polysomnographic Phenotypes

Depression is a common mental health disorder and a leading contributor to mortality and morbidity. Despite several advancements, the current screening methods have limitations in enabling the robust and automated detection of depression, thereby hindering early diagnosis and timely intervention. This study aimed to develop explainable artificial intelligence (AI) models to predict depression using polysomnographic phenotype data, ensuring high predictive performance while providing clear insights into the importance of features influencing the risk of depression. Advanced machine learning algorithms such as random forest, extreme gradient boosting, categorical boosting, and light gradient boosting machines were employed to train and validate the predictive AI models. Phenotype data from subjective health questionnaires, clinical assessments, and demographic factors were analyzed. The explainable AI models identified the important features, and their performance was evaluated using cross-validation. The study population, comprising 114 control participants and 39 individuals with depression, was stratified based on validated depression-scoring methods. The proposed explainable AI models achieved an F1-score of 85%, verifying their high reliability in predicting depression. Key features influencing the risk of depression, such as anxiety disorders, sleep efficiency, and demographic factors, offer actionable insights for clinical practice, highlighting the transparency of these models. This study proposed and developed explainable AI models based on polysomnographic phenotype data for the automated detection of depression and verified that these models help improve mental health diagnostics, enabling timely interventions.

A Risk Warning Model for Anemia Based on Facial Visible Light Reflectance Spectroscopy: Cross-Sectional Study

Background

Anemia is a global public health issue causing symptoms such as fatigue, weakness, and cognitive decline. Furthermore, anemia is associated with various diseases and increases the risk of postoperative complications and mortality. Frequent invasive blood tests for diagnosis also pose additional discomfort and risks to patients.

Objective

This study aims to assess the facial spectral characteristics of patients with anemia and to develop a predictive model for anemia risk using machine learning approaches.

Methods

Between August 2022 and September 2023, we collected facial image data from 78 anemic patients who met the inclusion criteria from the Hematology Department of Shanghai Hospital of Traditional Chinese Medicine. Between March 2023 and September 2023, we collected data from 78 healthy adult participants from Shanghai Jiading Community Health Center and Shanghai Gaohang Community Health Center. A comprehensive statistical analysis was performed to evaluate differences in spectral characteristics between the anemic patients and healthy controls. Then, we used 10 different machine learning algorithms to create a predictive model for anemia. The least absolute shrinkage and selection operator was used to analyze the predictors. We integrated multiple machine learning classification models to identify the optimal model and developed Shapley additive explanations (SHAP) for personalized risk assessment.

Results

The study identified significant differences in facial spectral features between anemic patients and healthy controls. The support vector machine classifier outperformed other classification models, achieving an accuracy of 0.875 (95% CI 0.825-0.925) for distinguishing between the anemia and healthy control groups. In the SHAP interpretation of the model, forehead-570 nm, right cheek-520 nm, right zygomatic-570 nm, jaw-570 nm, and left cheek-610 nm were the features with the highest contributions.

Conclusions

Facial spectral data demonstrated clinical significance in anemia diagnosis, and the early warning model for anemia risk constructed based on spectral information demonstrated a high accuracy rate.

Tongue coating microbial communities vary in children with Henoch-Schönlein purpura

Previous research has shown that microbes play a role in immune-related diseases. Our study reveals that children with Henoch-Schönlein purpura nephritis have distinct and altered tongue coating microbiota, characterized by significant changes in species richness, diversity, and specific microbial compositions compared with healthy controls. Nevertheless, the particular involvement of tongue coating microbiota in Henoch-Schönlein Purpura remains unclear. A total of 26 children were enrolled, including 13 patients with HSP and 13 healthy children. Tongue coating samples were collected for DNA extraction and 16S rRNA gene sequencing. Alpha diversity indices, including ACE, Chao1, Shannon, and Simpson indices, revealed significantly lower richness and diversity of tongue coating microbiota in children with Henoch-Schönlein purpura compared to healthy controls. Beta diversity analysis demonstrated distinct clustering of microbial communities between HSP and healthy children, with significant compositional differences. 16S rRNA gene sequencing showed that the relative abundance of key genera, such as Veillonella and Prevotella, differed between the two groups. A random forest algorithm identified five genera as potential diagnostic biomarkers for HSP. Co-occurrence analysis revealed different hub microbes in HSP and healthy children. BugBase predicted an increased proportion of stress-tolerant bacteria in the HSP group compared to the healthy controls group. PICRUSt analysis indicated alterations in metabolic functions of tongue coating microbiota between HSP and healthy children, with 25 KEGG pathways exhibiting significant differences. Children with HSP exhibit marked differences in their tongue coating microbiota when contrasted with their healthy counterparts.

Advances in Neuroimaging and Deep Learning for Emotion Detection: A Systematic Review of Cognitive Neuroscience and Algorithmic Innovations

Background/Objectives: The following systematic review integrates neuroimaging techniques with deep learning approaches concerning emotion detection. It, therefore, aims to merge cognitive neuroscience insights with advanced algorithmic methods in pursuit of an enhanced understanding and applications of emotion recognition. Methods: The study was conducted following PRISMA guidelines, involving a rigorous selection process that resulted in the inclusion of 64 empirical studies that explore neuroimaging modalities such as fMRI, EEG, and MEG, discussing their capabilities and limitations in emotion recognition. It further evaluates deep learning architectures, including neural networks, CNNs, and GANs, in terms of their roles in classifying emotions from various domains: human-computer interaction, mental health, marketing, and more. Ethical and practical challenges in implementing these systems are also analyzed. Results: The review identifies fMRI as a powerful but resource-intensive modality, while EEG and MEG are more accessible with high temporal resolution but limited by spatial accuracy. Deep learning models, especially CNNs and GANs, have performed well in classifying emotions, though they do not always require large and diverse datasets. Combining neuroimaging data with behavioral and cognitive features improves classification performance. However, ethical challenges, such as data privacy and bias, remain significant concerns. Conclusions: The study has emphasized the efficiencies of neuroimaging and deep learning in emotion detection, while various ethical and technical challenges were also highlighted. Future research should integrate behavioral and cognitive neuroscience advances, establish ethical guidelines, and explore innovative methods to enhance system reliability and applicability.

Public interest in online searching of asthma information: insights from a Google trends analysis

BACKGROUND

Google Trends (GT) is a free tool that provides insights into the public's interest and information-seeking behavior on specific topics. In this study, we utilized GT data on patients' search history to better understand their questions and information needs regarding asthma.

METHODS

We extracted the relative GT search volume (RSV) for keywords associated with asthma to explore information-seeking behaviors and assess internet search patterns regarding asthma disease from 2004 to 2024 in both English and Persian languages. In addition, a correlation analysis was conducted to assess terms correlated with asthma searches. Then, the AutoRegressive predictive models were developed to estimate future patterns of asthma-related searches and the information needs of individuals with asthma.

RESULTS

The analysis revealed that the mean total RSV for asthma-related keywords over the 20-year period was 41.79 ± 6.07. The researchers found that while asthma-related search volume has shown a consistent upward trend in Persian-speaking countries over the last decade, English-speaking countries have experienced less variability in such searches except for a spike during the COVID-19 pandemic. The correlation analysis of related subjects showed that "air pollution", "infection", and "insomnia" have a positive correlation with asthma. Developing AutoRegressive predictive models on retrieved Google Trends data revealed a seasonal pattern in global asthma-related search interest. In contrast, the models forecasted a growing increase in information-seeking behaviors regarding asthma among Persian-speaking patients over the coming decades.

CONCLUSIONS

There are significant differences in how people search for and access asthma information based on their language and regional context. In English-speaking countries, searches tend to focus on broader asthma-related topics like pollution and infections, likely due to the availability of comprehensive asthma resources. In contrast, Persian speakers prioritize understanding specific aspects of asthma-like symptoms, medications, and complementary treatments. To address these divergent information needs, health organizations should tailor content to these divergent needs.

A worldwide itinerary of research ethics in science for a better social responsibility and justice: a bibliometric analysis and review

This study provides a comprehensive overview of research ethics in science using an approach that combine bibliometric analysis and systematic review. The importance of ethical conduct in scientific research to maintain integrity, credibility, and societal relevance has been highlighted. The findings revealed a growing awareness of ethical issues, as evidenced by the development of numerous guidelines, codes of conduct, and oversight institutions. However, significant challenges persist, including the lack of standardized approaches for detecting misconduct, limited understanding of the factors contributing to unethical behavior, and unclear definitions of ethical violations. To address these issues, this study recommends promoting transparency and data sharing, enhancing education, and training programs, establishing robust mechanisms to identify and address misconduct, and encouraging collaborative research and open science practices. This study emphasizes the need for a collaborative approach to restore public confidence in science, protect its positive impact, and effectively address global challenges, while upholding the principles of social responsibility and justice. This comprehensive approach is crucial for maintaining research credibility, conserving resources, and safeguarding both the research participants and the public.

Dual feature-based and example-based explanation methods

A new approach to the local and global explanation based on selecting a convex hull constructed for the finite number of points around an explained instance is proposed. The convex hull allows us to consider a dual representation of instances in the form of convex combinations of extreme points of a produced polytope. Instead of perturbing new instances in the Euclidean feature space, vectors of convex combination coefficients are uniformly generated from the unit simplex, and they form a new dual dataset. A dual linear surrogate model is trained on the dual dataset. The explanation feature importance values are computed by means of simple matrix calculations. The approach can be regarded as a modification of the well-known model LIME. The dual representation inherently allows us to get the example-based explanation. The neural additive model is also considered as a tool for implementing the example-based explanation approach. Many numerical experiments with real datasets are performed for studying the approach. A code of proposed algorithms is available. The proposed results are fundamental and can be used in various application areas. They do not involve specific human subjects and human data.

Commentary on "Preliminary Species Hypotheses" in Entomological Taxonomy: A Global Data and FAIR Infrastructure Perspective

What if early taxonomic findings were treated like preprints, open to iterative improvement or managed with practices from the open-source community, such as Git branching, merging and patch management? Prompted by Buckley's article Charting a Future for Entomological Taxonomy in New Zealand (2024), this commentary explores these possibilities in the context of biodiversity informatics. In response to the need for rapid, scalable biodiversity monitoring, Buckley introduces preliminary species hypotheses (PSH) as a bridge between quick identification tools and the rigorous Linnaean system, leveraging DNA barcoding and AI-assisted image recognition to produce provisional classifications that can later be validated. Expanding on Buckley's framework, this commentary emphasises the critical role of data linking, versioning and integration to support evolving taxonomic data. Borrowing from software and open-source practices, I explore the idea of managing PSH with an infrastructure that treats each taxonomic update as a versioned "commit", which can be tracked, refined and integrated over time. Drawing insights from FAIR (Findable, Accessible, Interoperable, Reusable) principles and Digital Extended Specimens, I identify infrastructure requirements for PSH, including robust data standards, persistent identifiers and interoperability to support global biodiversity repositories. Additionally, Taxonomic Data Objects offer a model for dynamically integrating PSH into adaptable taxonomies that can evolve with new data and tools. By positioning PSH within an open, infrastructure-focused framework, this commentary advocates for scalable, hypothesis-driven biodiversity data that meets modern conservation needs, bridging traditional and emerging practices in taxonomy.

The Heart of Transformation: Exploring Artificial Intelligence in Cardiovascular Disease

The application of artificial intelligence (AI) and machine learning (ML) in medicine and healthcare has been extensively explored across various areas. AI and ML can revolutionize cardiovascular disease management by significantly enhancing diagnostic accuracy, disease prediction, workflow optimization, and resource utilization. This review summarizes current advancements in AI and ML concerning cardiovascular disease, including their clinical investigation and use in primary cardiac imaging techniques, common cardiovascular disease categories, clinical research, patient care, and outcome prediction. We analyze and discuss commonly used AI and ML models, algorithms, and methodologies, highlighting their roles in improving clinical outcomes while addressing current limitations and future clinical applications. Furthermore, this review emphasizes the transformative potential of AI and ML in cardiovascular practice by improving clinical decision making, reducing human error, enhancing patient monitoring and support, and creating more efficient healthcare workflows for complex cardiovascular conditions.

A Systematic Study of Popular Software Packages and AI/ML Models for Calibrating In Situ Air Quality Data: An Example with Purple Air Sensors

Accurate air pollution monitoring is critical to understand and mitigate the impacts of air pollution on human health and ecosystems. Due to the limited number and geographical coverage of advanced, highly accurate sensors monitoring air pollutants, many low-cost and low-accuracy sensors have been deployed. Calibrating low-cost sensors is essential to fill the geographical gap in sensor coverage. We systematically examined how different machine learning (ML) models and open-source packages could help improve the accuracy of particulate matter (PM) 2.5 data collected by Purple Air sensors. Eleven ML models and five packages were examined. This systematic study found that both models and packages impacted accuracy, while the random training/testing split ratio (e.g., 80/20 vs. 70/30) had minimal impact (0.745% difference for R2). Long Short-Term Memory (LSTM) models trained in RStudio and TensorFlow excelled, with high R2 scores of 0.856 and 0.857 and low Root Mean Squared Errors (RMSEs) of 4.25 µg/m3 and 4.26 µg/m3, respectively. However, LSTM models may be too slow (1.5 h) or computation-intensive for applications with fast response requirements. Tree-boosted models including XGBoost (0.7612, 5.377 µg/m3) in RStudio and Random Forest (RF) (0.7632, 5.366 µg/m3) in TensorFlow offered good performance with shorter training times (<1 min) and may be suitable for such applications. These findings suggest that AI/ML models, particularly LSTM models, can effectively calibrate low-cost sensors to produce precise, localized air quality data. This research is among the most comprehensive studies on AI/ML for air pollutant calibration. We also discussed limitations, applicability to other sensors, and the explanations for good model performances. This research can be adapted to enhance air quality monitoring for public health risk assessments, support broader environmental health initiatives, and inform policy decisions.

Algorithmic emergence? Epistemic in/justice in AI-directed transformations of healthcare

Moves toward integration of Artificial Intelligence (AI), particularly deep learning and generative AI-based technologies, into the domains of healthcare and public health have recently intensified, with a growing body of literature tackling the ethico-political implications of this. This paper considers the interwoven epistemic, sociopolitical and technical ramifications of healthcare-AI entanglements, examining how AI materialities shape emergence of particular modes of healthcare organization, governance and roles, and reflecting on how to embed participatory engagement within these entanglements. We discuss the implications of socio-technical entanglements between AI and Evidence-Based Medicine (EBM) for equitable development and governance of health AI. AI applications invariably center on the domains of medical knowledge and practice that are amenable to computational workings. This, in turn, intensifies the prioritization of these medical domains and furthers the assumptions which support the development of AI, a move which decontextualizes the qualitative nuances and complexities of healthcare while simultaneously advancing infrastructure to support these medical domains. We sketch the material and ideological reconfiguration of healthcare which is being shaped by the move toward embedding health AI assemblages in real-world contexts. We then consider the implications of this, how AI might be best employed in healthcare, and how to tackle the algorithmic injustices which become reproduced within health AI assemblages.

Data-Sharing Statements Requested from Clinical Trials by Public, Environmental, and Occupational Health Journals: Cross-Sectional Study

BACKGROUND

Data sharing plays a crucial role in health informatics, contributing to improving health information systems, enhancing operational efficiency, informing policy and decision-making, and advancing public health surveillance including disease tracking. Sharing individual participant data in public, environmental, and occupational health trials can help improve public trust and support by enhancing transparent reporting and reproducibility of research findings. The International Committee of Medical Journal Editors (ICMJE) requires all papers to include a data-sharing statement. However, it is unclear whether journals in the field of public, environmental, and occupational health adhere to this requirement.

OBJECTIVE

This study aims to investigate whether public, environmental, and occupational health journals requested data-sharing statements from clinical trials submitted for publication.

METHODS

In this bibliometric survey of "Public, Environmental, and Occupational Health" journals, defined by the Journal Citation Reports (as of June 2023), we included 202 journals with clinical trial reports published between 2019 and 2022. The primary outcome was a journal request for a data-sharing statement, as identified in the paper submission instructions. Multivariable logistic regression analysis was conducted to evaluate the relationship between journal characteristics and journal requests for data-sharing statements, with results presented as odds ratios (ORs) and corresponding 95% CIs. We also investigated whether the journals included a data-sharing statement in their published trial reports.

RESULTS

Among the 202 public, environmental, and occupational health journals included, there were 68 (33.7%) journals that did not request data-sharing statements. Factors significantly associated with journal requests for data-sharing statements included open access status (OR 0.43, 95% CI 0.19-0.97), high journal impact factor (OR 2.31, 95% CI 1.15-4.78), endorsement of Consolidated Standards of Reporting Trials (OR 2.43, 95% CI 1.25-4.79), and publication in the United Kingdom (OR 7.18, 95% CI 2.61-23.4). Among the 134 journals requesting data-sharing statements, 26.9% (36/134) did not have statements in their published trial reports.

CONCLUSIONS

Over one-third of the public, environmental, and occupational health journals did not request data-sharing statements in clinical trial reports. Among those journals that requested data-sharing statements in their submission guidance pages, more than one quarter published trial reports with no data-sharing statements. These results revealed an inadequate practice of requesting data-sharing statements by public, environmental, and occupational health journals, requiring more effort at the journal level to implement ICJME recommendations on data-sharing statements.

Simplicity within biological complexity

Motivation

Heterogeneous, interconnected, systems-level, molecular (multi-omic) data have become increasingly available and key in precision medicine. We need to utilize them to better stratify patients into risk groups, discover new biomarkers and targets, repurpose known and discover new drugs to personalize medical treatment. Existing methodologies are limited and a paradigm shift is needed to achieve quantitative and qualitative breakthroughs.

Results

In this perspective paper, we survey the literature and argue for the development of a comprehensive, general framework for embedding of multi-scale molecular network data that would enable their explainable exploitation in precision medicine in linear time. Network embedding methods (also called graph representation learning) map nodes to points in low-dimensional space, so that proximity in the learned space reflects the network's topology-function relationships. They have recently achieved unprecedented performance on hard problems of utilizing few omic data in various biomedical applications. However, research thus far has been limited to special variants of the problems and data, with the performance depending on the underlying topology-function network biology hypotheses, the biomedical applications, and evaluation metrics. The availability of multi-omic data, modern graph embedding paradigms and compute power call for a creation and training of efficient, explainable and controllable models, having no potentially dangerous, unexpected behaviour, that make a qualitative breakthrough. We propose to develop a general, comprehensive embedding framework for multi-omic network data, from models to efficient and scalable software implementation, and to apply it to biomedical informatics, focusing on precision medicine and personalized drug discovery. It will lead to a paradigm shift in the computational and biomedical understanding of data and diseases that will open up ways to solve some of the major bottlenecks in precision medicine and other domains.

Deep Learning-Driven Single-Lead ECG Classification: A Rapid Approach for Comprehensive Cardiac Diagnostics

Background/Objectives: This study aims to address the critical need for accessible, early, and accurate cardiac di-agnostics, especially in resource-limited or remote settings. By shifting focus from traditional multi-lead ECG analysis to single-lead ECG data, this research explores the potential of advanced deep learning models for classifying cardiac conditions, including Nor-mal, Abnormal, Previous Myocardial Infarction (PMI), and Myocardial Infarction (MI). Methods: Five state-of-the-art deep learning architectures-Inception, DenseNet201, MobileNetV2, NASNetLarge, and VGG16-were systematically evaluated on individual ECG leads. Key performance metrics, such as model accuracy, inference time, and size, were analyzed to determine the optimal configurations for practical applications. Results: VGG16 emerged as the most accurate model, achieving an F1-score of 98.11% on lead V4 with a prediction time of 4.2 ms and a size of 528 MB, making it suitable for high-precision clinical settings. MobileNetV2, with a compact size of 13.4 MB, offered a balanced performance, achieving a 97.24% F1-score with a faster inference time of 3.2 ms, positioning it as an ideal candidate for real-time monitoring and telehealth applications. Conclusions: This study bridges a critical gap in cardiac diagnostics by demonstrating the feasibility of lightweight, scalable, single-lead ECG analysis using advanced deep learning models. The findings pave the way for deploying portable diagnostic tools across diverse settings, enhancing the accessibility and efficiency of cardiac care globally.

Metaverse for mental health disorders: Opportunities and challenges

Several articles on the mental health impact of the metaverse and the need to balance its potential benefits with the risks of metaverse use has recently published. The metaverse consists of a combination of immersive technologies and artificial intelligence algorithms. The metaverse differs from the preceding digital psychiatric interventions due to its complex structure and interactions between components. The diverse functions of the metaverse ensure that it may have a substantial impact on mental health. However, the evidence for its efficacy in treating mental health disorders is limited to a few trials. The mental health benefits of immersive technologies are well-documented and suggest that metaverse-based psychiatric treatment may be similarly efficacious. The mental health risks of the metaverse are largely unknown, and it is not clear whether they will be greater than other digital psychiatric interventions. Much more research is needed to determine whether metaverse-based psychiatric treatment will meet the standards of appropriate mental healthcare.

The anti-circumsporozoite antibody response to repeated, seasonal booster doses of the malaria vaccine RTS,S/AS01E

The recently deployed RTS,S/AS01E malaria vaccine induces a strong antibody response to the circumsporozoite protein (CSP) on the surface of the Plasmodium falciparum sporozoite which is associated with protection. The anti-CSP antibody titre falls rapidly after primary vaccination, associated with a decline in efficacy, but the antibody titre and the protective response can be partially restored by a booster dose of vaccine, but this response is also transitory. In many malaria- endemic areas of Africa, children are at risk of malaria, including severe malaria, until they are five years of age or older and to sustain protection from malaria for this period by vaccination with RTS,S/AS01E, repeated booster doses of vaccine may be required. However, there is little information about the immune response to repeated booster doses of RTS,S/AS01E. In many malaria-endemic areas of Africa, the burden of malaria is largely restricted to the rainy season and, therefore, a recent trial conducted in Burkina Faso and Mali explored the impact of repeated annual booster doses of RTS,S/AS01E given immediately prior to the malaria transmission season until children reached the age of five years. Anti-CSP antibody titres were measured in sera obtained from a randomly selected subset of children enrolled in this trial collected before and one month after three priming and four annual booster doses of vaccine using the GSK ELISA developed at the University of Ghent and, in a subset of these samples, by a multiplex assay developed at the University of Oxford. Three priming doses of RTS,S/AS01E induced a strong anti-CSP antibody response (GMT 368.9 IU/mL). Subsequent annual, seasonal booster doses induced a strong, but lower, antibody response; the GMT after the fourth booster was 128.5 IU/mL. Children whose antibody response was in the upper and middle terciles post vaccination had a lower incidence of malaria during the following year than children in the lowest tercile. Results obtained with GSK ELISA and the Oxford Multiplex assay were strongly correlated (Pearson's correlation coefficient, r = 0.94; 95% CI, 0.93-0.95). Although anti-CSP antibody titres declined after repeated booster doses of RTS,S/AS01E a high, although declining, level of efficacy was sustained suggesting that there may have been changes in the characteristics of the anti-CSP antibody following repeated booster doses.Clinical Trials Registration. NCT03143218.

Looking to the Future of Viral Vectors in Ocular Gene Therapy: Clinical Review

Eye diseases can significantly affect the quality of life of patients due to decreased visual acuity. Although modern ophthalmological diagnostic methods exist, some diseases of the visual system are asymptomatic in the early stages. Most patients seek advice from an ophthalmologist as a result of rapidly progressive manifestation of symptoms. A number of inherited and acquired eye diseases have only supportive treatment without eliminating the etiologic factor. A promising solution to this problem may be gene therapy, which has proven efficacy and safety shown in a number of clinical studies. By directly altering or replacing defective genes, this therapeutic approach will stop as well as reverse the progression of eye diseases. This review examines the concept of gene therapy and its application in the field of ocular pathologies, emphasizing the most recent scientific advances and their potential impacts on visual function status.

Examining inclusivity: the use of AI and diverse populations in health and social care: a systematic review

BACKGROUND

Artificial intelligence (AI)-based systems are being rapidly integrated into the fields of health and social care. Although such systems can substantially improve the provision of care, diverse and marginalized populations are often incorrectly or insufficiently represented within these systems. This review aims to assess the influence of AI on health and social care among these populations, particularly with regard to issues related to inclusivity and regulatory concerns.

METHODS

We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Six leading databases were searched, and 129 articles were selected for this review in line with predefined eligibility criteria.

RESULTS

This research revealed disparities in AI outcomes, accessibility, and representation among diverse groups due to biased data sources and a lack of representation in training datasets, which can potentially exacerbate inequalities in care delivery for marginalized communities.

CONCLUSION

AI development practices, legal frameworks, and policies must be reformulated to ensure that AI is applied in an equitable manner. A holistic approach must be used to address disparities, enforce effective regulations, safeguard privacy, promote inclusion and equity, and emphasize rigorous validation.

Large language models generating synthetic clinical datasets: a feasibility and comparative analysis with real-world perioperative data

Background

Clinical data is instrumental to medical research, machine learning (ML) model development, and advancing surgical care, but access is often constrained by privacy regulations and missing data. Synthetic data offers a promising solution to preserve privacy while enabling broader data access. Recent advances in large language models (LLMs) provide an opportunity to generate synthetic data with reduced reliance on domain expertise, computational resources, and pre-training.

Objective

This study aims to assess the feasibility of generating realistic tabular clinical data with OpenAI's GPT-4o using zero-shot prompting, and evaluate the fidelity of LLM-generated data by comparing its statistical properties to the Vital Signs DataBase (VitalDB), a real-world open-source perioperative dataset.

Methods

In Phase 1, GPT-4o was prompted to generate a dataset with qualitative descriptions of 13 clinical parameters. The resultant data was assessed for general errors, plausibility of outputs, and cross-verification of related parameters. In Phase 2, GPT-4o was prompted to generate a dataset using descriptive statistics of the VitalDB dataset. Fidelity was assessed using two-sample t-tests, two-sample proportion tests, and 95% confidence interval (CI) overlap.

Results

In Phase 1, GPT-4o generated a complete and structured dataset comprising 6,166 case files. The dataset was plausible in range and correctly calculated body mass index for all case files based on respective heights and weights. Statistical comparison between the LLM-generated datasets and VitalDB revealed that Phase 2 data achieved significant fidelity. Phase 2 data demonstrated statistical similarity in 12/13 (92.31%) parameters, whereby no statistically significant differences were observed in 6/6 (100.0%) categorical/binary and 6/7 (85.71%) continuous parameters. Overlap of 95% CIs were observed in 6/7 (85.71%) continuous parameters.

Conclusion

Zero-shot prompting with GPT-4o can generate realistic tabular synthetic datasets, which can replicate key statistical properties of real-world perioperative data. This study highlights the potential of LLMs as a novel and accessible modality for synthetic data generation, which may address critical barriers in clinical data access and eliminate the need for technical expertise, extensive computational resources, and pre-training. Further research is warranted to enhance fidelity and investigate the use of LLMs to amplify and augment datasets, preserve multivariate relationships, and train robust ML models.

Differential Pattern of Symptom Correlation With Acute Respiratory Infections in Korea

BACKGROUND

Acute respiratory infections (ARIs) pose global health challenges, with major outbreaks affecting healthcare systems and resulting in significant morbidity and mortality. We aimed to identify distinctive signs or symptoms correlated with ARIs for utilizing syndromic surveillance.

METHODS

We used data from national Korean databases to examine correlations between various symptoms and the reported ARI viruses in children aged under and over 5 years.

RESULTS

In children under 5 years old, respiratory symptoms were strongly correlated with human adenovirus, human respiratory syncytial virus, and human rhinovirus. Patients aged over 5 years displayed more diverse patterns, with varied correlations. The cases of fever were a strong indicator of respiratory viruses (human adenovirus, human parainfluenza viruses, and human rhinovirus) in children under 5 years old, while those over 5 years showed symptoms such as smell and taste disturbances.

CONCLUSIONS

These findings emphasize the correlation between various symptoms and ARIs across different age groups and may help to improve syndromic surveillance systems.

Exploring ethical considerations in medical research: Harnessing pre-generated transformers for AI-powered ethics discussions

INTRODUCTION

In medical research involving human subjects, ethical review is essential to protect individuals. However, concerns have been raised about variations in ethical review opinions and a decline in review quality. Adequately protecting human subjects requires multifaceted opinions from ethics committee members. Despite the need to increase the number of committee members, resources are limited. To address these challenges, we explored the use of a generative pre- learning transformer, an interactive artificial intelligence (AI) tool, to discuss ethical issues in medical research.

METHODS

The generation AI used in the research used ChatGPT3.5, which has learned ethical guidelines from various countries worldwide. We requested the generative AI to provide insights on ethical considerations for virtual research involving individuals. The obtained answers were documented and verified by experts.

RESULTS

The AI successfully highlighted considerations for informed consent regarding individuals with dementia and mental illness, as well as concerns about invasiveness in research. It also raised points about potential side effects of off-label drug use. However, it could not offer specific measures for psychological considerations or broader ethical issues, providing limited ethical insights. This limitation may be attributed to biased opinions resulting from machine learning optimization, preventing comprehensive identification of certain ethical issues.

CONCLUSION

Although the validity of ethical opinions generated by the generative AI requires further examination, our findings suggest that this technology could be employed to prompt reviews and re-evaluate ethical concerns arising in research.

CAREUP: An Integrated Care Platform with Intrinsic Capacity Monitoring and Prediction Capabilities

This paper describes CAREUP, a novel older adult healthy aging support platform based on Intrinsic Capacity (IC) monitoring. Besides standard functionalities like storing health measurement data or providing users with personalized recommendations, the platform includes novel intrinsic capacity assessment and prediction algorithms. Older adults' performance is continuously monitored in all five IC domains-locomotion, psychology, cognition, vitality, and sensory capacity-based on measurement results and answers to questionnaires gathered using the platform's mobile applications. The users are also presented with a machine learning-based prediction of how their intrinsic capacity might change over the following years. The platform's operation was successfully tested with the participation of older adults and their caregivers in three countries: Austria, Italy, and Romania.

MAD-Onto: an ontology design for mobile app development

Introduction

Mobile app development has rapidly evolved into a crucial aspect of modern technology, driving innovation across various industries and transforming user experiences globally. The dynamic nature of mobile technology requires developers to navigate a complex landscape of platforms, devices, and user requirements. Effective management and sharing of knowledge are essential to address these challenges, ensuring streamlined development processes and enhanced collaboration among stakeholders.

Methods

To this end, ontologies have emerged as powerful tools for structuring and standardizing domain-specific knowledge. This paper introduces MAD-onto, a comprehensive ontology designed specifically for the mobile app development domain. The ontology is constructed by identifying key concepts, defining classes and their hierarchies, establishing class properties, and creating instances relevant to mobile app development. To ensure robustness, the ontology is evaluated using a multi-criteria evaluation metric, focusing on consistency, completeness, conciseness, expandability, and sensitiveness. Additionally, SWRL rules are applied to validate and enforce logical constraints within the ontology.

Results

Through these rigorous evaluation methods, MAD-onto demonstrates its utility in providing a structured framework for the mobile app development lifecycle, facilitating better decision-making, collaboration, and efficiency.

Discussion

The findings highlight the significance of ontology-driven approaches in addressing the complexities of mobile app development and set a foundation for future research and advancements in this field.

Hybridization in the Anthropocene - how pollution and climate change disrupt mate selection in freshwater fish

Chemical pollutants and/or climate change have the potential to break down reproductive barriers between species and facilitate hybridization. Hybrid zones may arise in response to environmental gradients and secondary contact between formerly allopatric populations, or due to the introduction of non-native species. In freshwater ecosystems, field observations indicate that changes in water quality and chemistry, due to pollution and climate change, are correlated with an increased frequency of hybridization. Physical and chemical disturbances of water quality can alter the sensory environment, thereby affecting chemical and visual communication among fish. Moreover, multiple chemical compounds (e.g. pharmaceuticals, metals, pesticides, and industrial contaminants) may impair fish physiology, potentially affecting phenotypic traits relevant for mate selection (e.g. pheromone production, courtship, and coloration). Although warming waters have led to documented range shifts, and chemical pollution is ubiquitous in freshwater ecosystems, few studies have tested hypotheses about how these stressors may facilitate hybridization and what this means for biodiversity and species conservation. Through a systematic literature review across disciplines (i.e. ecotoxicology and evolutionary biology), we evaluate the biological interactions, toxic mechanisms, and roles of physical and chemical environmental stressors (i.e. chemical pollution and climate change) in disrupting mate preferences and inducing interspecific hybridization in freshwater fish. Our study indicates that climate change-driven changes in water quality and chemical pollution may impact visual and chemical communication crucial for mate choice and thus could facilitate hybridization among fishes in freshwater ecosystems. To inform future studies and conservation management, we emphasize the importance of further research to identify the chemical and physical stressors affecting mate choice, understand the mechanisms behind these interactions, determine the concentrations at which they occur, and assess their impact on individuals, populations, species, and biological diversity in the Anthropocene.

Tau PET positivity predicts clinically relevant cognitive decline driven by Alzheimer's disease compared to comorbid cases; proof of concept in the ADNI study

β-amyloid (Aβ) pathology is not always coupled with Alzheimer's disease (AD) relevant cognitive decline. We assessed the accuracy of tau PET to identify Aβ(+) individuals who show prospective disease progression. 396 cognitively unimpaired and impaired individuals with baseline Aβ and tau PET and a follow-up of ≥ 2 years were selected from the Alzheimer's Disease Neuroimaging Initiative dataset. The participants were dichotomously grouped based on either clinical conversion (i.e., change of diagnosis) or cognitive deterioration (fast (FDs) vs. slow decliners (SDs)) using data-driven clustering of the individual annual rates of cognitive decline. To assess cognitive decline in individuals with isolated Aβ(+) or absence of both Aβ and tau (T) pathologies, we investigated the prevalence of non-AD comorbidities and FDG PET hypometabolism patterns suggestive of AD. Baseline tau PET uptake was higher in Aβ(+)FDs than in Aβ(-)FD/SDs and Aβ(+)SDs, independently of baseline cognitive status. Baseline tau PET uptake identified MCI Aβ(+) Converters and Aβ(+)FDs with an area under the curve of 0.85 and 0.87 (composite temporal region of interest) respectively, and was linearly related to the annual rate of cognitive decline in Aβ(+) individuals. The T(+) individuals constituted largely a subgroup of those being Aβ(+) and those clustered as FDs. The most common biomarker profiles in FDs (n = 70) were Aβ(+)T(+) (n = 34, 49%) and Aβ(+)T(-) (n = 19, 27%). Baseline Aβ load was higher in Aβ(+)T(+)FDs (M = 83.03 ± 31.42CL) than in Aβ(+)T(-)FDs (M = 63.67 ± 26.75CL) (p-value = 0.038). Depression diagnosis was more prevalent in Aβ(+)T(-)FDs compared to Aβ(+)T(+)FDs (47% vs. 15%, p-value = 0.021), as were FDG PET hypometabolism pattern not suggestive of AD (86% vs. 50%, p-value = 0.039). Our findings suggest that high tau PET uptake is coupled with both Aβ pathology and accelerated cognitive decline. In cases of isolated Aβ(+), cognitive decline may be associated with changes within the AD spectrum in a multi-morbidity context, i.e., mixed AD.

Interpretation of acid-base metabolism on arterial blood gas samples via machine learning algorithms

BACKGROUND

Arterial blood gas evaluation is crucial for critically ill patients, as it provides essential information about acid-base metabolism and respiratory balance, but evaluation can be complex and time-consuming. Artificial intelligence can perform tasks that require human intelligence, and it is revolutionizing healthcare through technological advancements.

AIM

This study aims to assess arterial blood gas evaluation using artificial intelligence algorithms.

METHODS

The study included 21.541 retrospective arterial blood gas samples, categorized into 15 different classes by experts for evaluating acid-base metabolism status. Six machine learning algorithms were utilized; accuracy, balanced accuracy, sensitivity, specificity, precision, and F1 values of the models were determined; and ROC curves were drawn to assess areas under the curve for each class. Evaluation of which sample was estimated in which class was conducted using the confusion matrices of the models.

RESULTS

The bagging classifier (BC) model achieved the highest balanced accuracy with 99.24%, whereas the XGBoost model reached the highest accuracy with 99.66%. The BC model shows 100% sensitivity for nine classes and 100% specificity for 10 classes, and the model correctly predicted 6438 of 6463 test samples and achieved an accuracy of 99.61%, with an area under the curve > 0.9 in all classes on a class basis.

CONCLUSION

The machine learning models developed exhibited remarkable accuracy, sensitivity, and specificity in predicting the status of acid-base metabolism. However, implementing these models can aid clinicians, freeing up their time for more intricate tasks.

A novel case-based reasoning system for explainable lung cancer diagnosis

Lung cancer is a leading cause of cancer death worldwide. The survival rate is generally higher when this disease is detected in its early stages. Advances in artificial intelligence (AI) have enabled the development of decision support systems that help physicians diagnose diseases. However, these systems often provide final predictions without clarifying how those decisions are reached, raising concerns about trust and adaptation in life-threatening diseases. To address these issues, this study proposes an explainable case-based reasoning (XCBR) approach that considers both physicians' tendency to base their decisions on past cases and the case complexity in its predictions and explanations. The proposed XCBR is enhanced with naïve Bayes (NB) and multilayer perceptron (MLP) classifiers which are processed hierarchically: when the NB deems its predictions to be unlikely, the MLP classifier is employed to verify or update the predictions. This approach incorporates Shapley additive explanations values to elucidate the solutions offered by the MLP. Furthermore, it utilizes the Harris hawks optimization algorithm for feature selection and feature weighting. The proposed XCBR achieved high accuracies of 94.47 % and 100 % on two different datasets, demonstrating its generalization capability. Based on Wilcoxon signed-rank test, its classification accuracy is comparable to that of other state-of-the-art approaches and commonly used classifiers. Moreover, since this approach prioritizes case complexity in its predictions and explanations, it offers better explainability and is particularly suited for serious diseases.

Large-scale analysis of fact-checked stories on Twitter reveals graded effects of ambiguity and falsehood on information reappearance

Misinformation disrupts our information ecosystem, adversely affecting individuals and straining social cohesion and democracy. Understanding what causes online (mis)information to (re)appear is crucial for fortifying our information ecosystem. We analyzed a large-scale Twitter (now "X") dataset of about 2 million tweets across 123 fact-checked stories. Previous research suggested a falsehood effect (false information reappears more frequently) and an ambiguity effect (ambiguous information reappears more frequently). However, robust indicators for their existence remain elusive. Using polynomial statistical modeling, we compared a falsehood model, an ambiguity model, and a dual effect model. The data supported the dual effect model ( 13.76 times as likely as a null model), indicating both ambiguity and falsehood promote information reappearance. However, evidence for ambiguity was stronger: the ambiguity model was 6.6 times as likely as the falsehood model. Various control checks affirmed the ambiguity effect, while the falsehood effect was less stable. Nonetheless, the best-fitting model explained < 7 % of the variance, indicating that (i) the dynamics of online (mis)information are complex and (ii) falsehood effects may play a smaller role than previous research has suggested. These findings underscore the importance of understanding the dynamics of online (mis)information, though our focus on fact-checked stories may limit the generalizability to the full spectrum of information shared online. Even so, our results can inform policymakers, journalists, social media platforms, and the public in building a more resilient information environment, while also opening new avenues for research, including source credibility, cross-platform applicability, and psychological factors.

Progress in toxicogenomics to protect human health

Toxicogenomics measures molecular features, such as transcripts, proteins, metabolites and epigenomic modifications, to understand and predict the toxicological effects of environmental and pharmaceutical exposures. Transcriptomics has become an integral tool in contemporary toxicology research owing to innovations in gene expression profiling that can provide mechanistic and quantitative information at scale. These data can be used to predict toxicological hazards through the use of transcriptomic biomarkers, network inference analyses, pattern-matching approaches and artificial intelligence. Furthermore, emerging approaches, such as high-throughput dose-response modelling, can leverage toxicogenomic data for human health protection even in the absence of predicting specific hazards. Finally, single-cell transcriptomics and multi-omics provide detailed insights into toxicological mechanisms. Here, we review the progress since the inception of toxicogenomics in applying transcriptomics towards toxicology testing and highlight advances that are transforming risk assessment.

Making Pathologists Ready for the New Artificial Intelligence Era: Changes in Required Competencies

In recent years, there has been an increasing interest in developing and using artificial intelligence (AI) models in pathology. Although pathologists generally have a positive attitude toward AI, they report a lack of knowledge and skills regarding how to use it in practice. Furthermore, it remains unclear what skills pathologists would require to use AI adequately and responsibly. However, adequate training of (future) pathologists is essential for successful AI use in pathology. In this paper, we assess which entrustable professional activities (EPAs) and associated competencies pathologists should acquire in order to use AI in their daily practice. We make use of the available academic literature, including literature in radiology, another image-based discipline, which is currently more advanced in terms of AI development and implementation. Although microscopy evaluation and reporting could be transferrable to AI in the future, most of the current pathologist EPAs and competencies will likely remain relevant when using AI techniques and interpreting and communicating results for individual patient cases. In addition, new competencies related to technology evaluation and implementation will likely be necessary, along with knowing one's own strengths and limitations in human-AI interactions. Because current EPAs do not sufficiently address the need to train pathologists in developing expertise related to technology evaluation and implementation, we propose a new EPA to enable pathology training programs to make pathologists fit for the new AI era "using AI in diagnostic pathology practice" and outline its associated competencies.

Lifestyle factors and other predictors of common mental disorders in diagnostic machine learning studies: A systematic review

BACKGROUND

Machine Learning (ML) models have been used to predict common mental disorders (CMDs) and may provide insights into the key modifiable factors that can identify and predict CMD risk and be targeted through interventions. This systematic review aimed to synthesise evidence from ML studies predicting CMDs, evaluate their performance, and establish the potential benefit of incorporating lifestyle data in ML models alongside biological and/or demographic-environmental factors.

METHODS

This systematic review adheres to the PRISMA statement (Prospero CRD42023401194). Databases searched included MEDLINE, EMBASE, PsycInfo, IEEE Xplore, Engineering Village, Web of Science, and Scopus from database inception to 28/08/24. Included studies used ML methods with feature importance to predict CMDs in adults. Risk of bias (ROB) was assessed using PROBAST. Model performance metrics were compared. The ten most important variables reported by each study were assigned to broader categories to evaluate their frequency across studies.

RESULTS

117 studies were included (111 model development-only, 16 development and validation). Deep learning methods showed best accuracy for predicting CMD cases. Studies commonly incorporated features from multiple categories (n = 56), and frequently identified demographic-environmental predictors in their top ten most important variables (63/69 models). These tended to be in combination with psycho-social and biological variables (n = 15). Lifestyle data were infrequently examined as sole predictors of CMDs across included studies (4.27 %). Studies commonly had high heterogeneity and ROB ratings.

CONCLUSION

This review is the first to evaluate the utility of diagnostic ML for CMDs, assess their ROB, and evaluate predictor types. CMDs were able to be predicted, however studies had high ROB and lifestyle data were underutilised, precluding full identification of a robust predictor set.

Disentangling the effect of sex from brain size on brain organization and cognitive functioning

Neuroanatomical sex differences estimated in neuroimaging studies are confounded by total intracranial volume (TIV) as a major biological factor. Employing a matching approach widely used for causal modeling, we disentangled the effect of TIV from sex to study sex-differentiated brain aging trajectories, their relation to functional networks and cytoarchitectonic classes, brain allometry, and cognition. Using data from the UK Biobank, we created subsamples that removed, maintained, or exaggerated the TIV differences in the original sample. We compared regional and vertex-level sex estimates across subsamples. The overall sex-related differences diminished in head size-matched subsamples, suggesting that most of the observed variability results from TIV differences. Furthermore, bidirectional sex differences in brain neuroanatomy emerged that were previously masked by the effect of TIV. Allometry remained fairly consistent across lifespan and was not sex-differentiated. Finally, the matching process changed the direction of the estimated sex differences in "verbal and numerical reasoning" and "working memory", suggesting that behavioral sex difference investigations can benefit from additional biological analysis to uncover the underlying factors contributing to cognition. Taken together, we provide new evidence disentangling sex differences from TIV as a relevant biological confound.

Safeguarding Patients in the AI Era: Ethics at the Forefront of Pharmacovigilance

Artificial intelligence is increasingly being used in pharmacovigilance. However, the use of artificial intelligence in pharmacovigilance raises ethical concerns related to fairness, non-discrimination, compliance, and responsibility as the central ethical principles in risk assessment and regulatory requirements. This paper explores these concerns and provides a roadmap to how to address these challenges by considering data collection, privacy protection, transparency and accountability, model training, and explainability in artificial intelligence decision making for drug safety surveillance. A number of responsible approaches have been identified including an ethics framework and best practices to enhance artificial intelligence use in healthcare. The document also recognizes some initiatives that have demonstrated the importance of ethics in artificial intelligence pharmacovigilance. Nevertheless, the major needs mentioned in this paper are transparency, accountability, data protection, and fairness, which stress the necessity of collaboration to construct a cognitive framework aimed at integrating ethical artificial intelligence into pharmacovigilance. In conclusion, innovation should be balanced with ethical responsibility to enhance public health outcomes as well as patient safety.

Artificial Intelligence in Temporomandibular Joint Disorders: An Umbrella Review

OBJECTIVES

Given the complexity of temporomandibular joint disorders (TMDs) and their overlapping symptoms with other conditions, an accurate diagnosis necessitates a thorough examination, which can be time-consuming and resource-intensive. Consequently, innovative diagnostic tools are required to increase TMD diagnosis efficiency and precision. Therefore, the purpose of this umbrella review was to examine the existing evidence about the usefulness of artificial intelligence (AI) in TMD diagnosis.

MATERIAL AND METHODS

A comprehensive search of the literature was performed from inception to November 30, 2024, in PubMed-MEDLINE, Embase, and Scopus databases. This review evaluated systematic reviews (SRs) and meta-analyses (MAs) that reported TMD patients/datasets, any AI model as intervention, no treatment, placebo as comparator and accuracy, sensitivity, specificity, or predictive value of AI models as outcome. The extracted data were complemented with narrative synthesis.

RESULTS

Out of 1497 search results, this umbrella review included five studies. One of the five articles was an SR while the other four were SRMAs. Three studies focused on patients with temporomandibular joint (TMJ) problems as a group, whereas two were specific to temporomandibular joint osteoarthritis (TMJOA). The included studies reported the use of imaging datasets as samples, including cone-beam computed tomography (CBCT), magnetic resonance imaging (MRI), and panoramic radiography. The studies reported an accuracy level ranging from 0.59 to 1. Four studies reported sensitivity levels ranging from 0.76 to 0.80. Four studies reported specificity values ranging from 0.63 to 0.95 for TMJ conditions. However, only one study provided the area under the curve (AUC) in the diagnosis of TMDs.

CONCLUSIONS

AI has the ability to provide faster, more accurate, sensitive, and objective diagnosis of TMJ condition. However, the performance is determined on the AI models and datasets used. Therefore, before implementing AI models in clinical practice, it is essential for researchers to extensively refine and evaluate the AI application.

Compression-enabled interpretability of voxelwise encoding models

Voxelwise encoding models based on convolutional neural networks (CNNs) are widely used as predictive models of brain activity evoked by natural movies. Despite their superior predictive performance, the huge number of parameters in CNN-based models have made them difficult to interpret. Here, we investigate whether model compression can build more interpretable and more stable CNN-based voxelwise models while maintaining accuracy. We used multiple compression techniques to prune less important CNN filters and connections, a receptive field compression method to select receptive fields with optimal center and size, and principal component analysis to reduce dimensionality. We demonstrate that the model compression improves the accuracy of identifying visual stimuli in a hold-out test set. Additionally, compressed models offer a more stable interpretation of voxelwise pattern selectivity than uncompressed models. Finally, the receptive field-compressed models reveal that the optimal model-based population receptive fields become larger and more centralized along the ventral visual pathway. Overall, our findings support using model compression to build more interpretable voxelwise models.

Evaluating generalizability of oncology trial results to real-world patients using machine learning-based trial emulations

Randomized controlled trials (RCTs) evaluating anti-cancer agents often lack generalizability to real-world oncology patients. Although restrictive eligibility criteria contribute to this issue, the role of selection bias related to prognostic risk remains unclear. In this study, we developed TrialTranslator, a framework designed to systematically evaluate the generalizability of RCTs for oncology therapies. Using a nationwide database of electronic health records from Flatiron Health, this framework emulates RCTs across three prognostic phenotypes identified through machine learning models. We applied this approach to 11 landmark RCTs that investigated anti-cancer regimens considered standard of care for the four most prevalent advanced solid malignancies. Our analyses reveal that patients in low-risk and medium-risk phenotypes exhibit survival times and treatment-associated survival benefits similar to those observed in RCTs. In contrast, high-risk phenotypes show significantly lower survival times and treatment-associated survival benefits compared to RCTs. Our results were corroborated by a comprehensive robustness assessment, including examinations of specific patient subgroups, holdout validation and semi-synthetic data simulation. These findings suggest that the prognostic heterogeneity among real-world oncology patients plays a substantial role in the limited generalizability of RCT results. Machine learning frameworks may facilitate individual patient-level decision support and estimation of real-world treatment benefits to guide trial design.

Social Media, Survey, and Medical Literature Data Reveal Escalating Antisemitism Within the United States Healthcare Community

Antisemitism has been rising for decades and worsened following the events of Oct 7, 2023. Although anecdotal evidence suggests that these trends extend into the US medical community, quantitative data have been lacking. To address this gap, we quantitated publications about antisemitism, analyzed social media posts from the accounts of 220,405 healthcare professionals, and disseminated a survey to members of Jewish medical associations. Publications and social media posts about antisemitism rose > fivefold, while posts promoting antisemitic stereotypes increased 2-fourfold. Most Jewish-identifying medical students and professionals (75.4%) reported exposure to antisemitism. Together, our results suggest that antisemitism is escalating within the US healthcare community.

Artificial Intelligence (AI)-driven approach to climate action and sustainable development

Countries have pledged commitment to the 2030 Sustainable Development Goal (SDGs) and the Paris Agreement to combat climate change. To maximize synergies between SDGs and climate actions (CAs), we evaluate the alignment of national commitment to SDGs and emissions reduction targets by comparing action plans embodied in Voluntary National Review (VNR) reports and the Nationally Determined Contributions (NDCs) across 67 countries. An Artificial Intelligence (AI)-based approach is proposed in this study to explore the interconnectedness by applying machine learning classifier and natural language processing. Middle- and low-income countries with high emissions tend to have low NDC targets and contain similar information in VNR reports. High-income countries show less alignment between their NDCs and VNRs. The economic status of countries is found to be connected to their climate actions and SDGs alignment. Here, we demonstrate utility and promise in using AI techniques to unravel interactions between CA and SDG.

Machine Learning-Based Risk Factor Analysis and Prediction Model Construction for the Occurrence of Chronic Heart Failure: Health Ecologic Study

BACKGROUND

Chronic heart failure (CHF) is a serious threat to human health, with high morbidity and mortality rates, imposing a heavy burden on the health care system and society. With the abundance of medical data and the rapid development of machine learning (ML) technologies, new opportunities are provided for in-depth investigation of the mechanisms of CHF and the construction of predictive models. The introduction of health ecology research methodology enables a comprehensive dissection of CHF risk factors from a wider range of environmental, social, and individual factors. This not only helps to identify high-risk groups at an early stage but also provides a scientific basis for the development of precise prevention and intervention strategies.

OBJECTIVE

This study aims to use ML to construct a predictive model of the risk of occurrence of CHF and analyze the risk of CHF from a health ecology perspective.

METHODS

This study sourced data from the Jackson Heart Study database. Stringent data preprocessing procedures were implemented, which included meticulous management of missing values and the standardization of data. Principal component analysis and random forest (RF) were used as feature selection techniques. Subsequently, several ML models, namely decision tree, RF, extreme gradient boosting, adaptive boosting (AdaBoost), support vector machine, naive Bayes model, multilayer perceptron, and bootstrap forest, were constructed, and their performance was evaluated. The effectiveness of the models was validated through internal validation using a 10-fold cross-validation approach on the training and validation sets. In addition, the performance metrics of each model, including accuracy, precision, sensitivity, F1-score, and area under the curve (AUC), were compared. After selecting the best model, we used hyperparameter optimization to construct a better model.

RESULTS

RF-selected features (21 in total) had an average root mean square error of 0.30, outperforming principal component analysis. Synthetic Minority Oversampling Technique and Edited Nearest Neighbors showed better accuracy in data balancing. The AdaBoost model was most effective with an AUC of 0.86, accuracy of 75.30%, precision of 0.86, sensitivity of 0.69, and F1-score of 0.76. Validation on the training and validation sets through 10-fold cross-validation gave an AUC of 0.97, an accuracy of 91.27%, a precision of 0.94, a sensitivity of 0.92, and an F1-score of 0.94. After random search processing, the accuracy and AUC of AdaBoost improved. Its accuracy was 77.68% and its AUC was 0.86.

CONCLUSIONS

This study offered insights into CHF risk prediction. Future research should focus on prospective studies, diverse data, advanced techniques, longitudinal studies, and exploring factor interactions for better CHF prevention and management.

An Explainable Artificial Intelligence Text Classifier for Suicidality Prediction in Youth Crisis Text Line Users: Development and Validation Study

BACKGROUND

Suicide represents a critical public health concern, and machine learning (ML) models offer the potential for identifying at-risk individuals. Recent studies using benchmark datasets and real-world social media data have demonstrated the capability of pretrained large language models in predicting suicidal ideation and behaviors (SIB) in speech and text.

OBJECTIVE

This study aimed to (1) develop and implement ML methods for predicting SIBs in a real-world crisis helpline dataset, using transformer-based pretrained models as a foundation; (2) evaluate, cross-validate, and benchmark the model against traditional text classification approaches; and (3) train an explainable model to highlight relevant risk-associated features.

METHODS

We analyzed chat protocols from adolescents and young adults (aged 14-25 years) seeking assistance from a German crisis helpline. An ML model was developed using a transformer-based language model architecture with pretrained weights and long short-term memory layers. The model predicted suicidal ideation (SI) and advanced suicidal engagement (ASE), as indicated by composite Columbia-Suicide Severity Rating Scale scores. We compared model performance against a classical word-vector-based ML model. We subsequently computed discrimination, calibration, clinical utility, and explainability information using a Shapley Additive Explanations value-based post hoc estimation model.

RESULTS

The dataset comprised 1348 help-seeking encounters (1011 for training and 337 for testing). The transformer-based classifier achieved a macroaveraged area under the curve (AUC) receiver operating characteristic (ROC) of 0.89 (95% CI 0.81-0.91) and an overall accuracy of 0.79 (95% CI 0.73-0.99). This performance surpassed the word-vector-based baseline model (AUC-ROC=0.77, 95% CI 0.64-0.90; accuracy=0.61, 95% CI 0.61-0.80). The transformer model demonstrated excellent prediction for nonsuicidal sessions (AUC-ROC=0.96, 95% CI 0.96-0.99) and good prediction for SI and ASE, with AUC-ROCs of 0.85 (95% CI 0.97-0.86) and 0.87 (95% CI 0.81-0.88), respectively. The Brier Skill Score indicated a 44% improvement in classification performance over the baseline model. The Shapley Additive Explanations model identified language features predictive of SIBs, including self-reference, negation, expressions of low self-esteem, and absolutist language.

CONCLUSIONS

Neural networks using large language model-based transfer learning can accurately identify SI and ASE. The post hoc explainer model revealed language features associated with SI and ASE. Such models may potentially support clinical decision-making in suicide prevention services. Future research should explore multimodal input features and temporal aspects of suicide risk.

Postdoc publications and citations link to academic retention and faculty success

Postdoctoral training is a career stage often described as a demanding and anxiety-laden time when many promising PhDs see their academic dreams slip away due to circumstances beyond their control. We use a unique dataset of academic publishing and careers to chart the more or less successful postdoctoral paths. We build a measure of academic success on the citation patterns two to five years into a faculty career. Then, we monitor how students' postdoc positions-in terms of relocation, change of topic, and early well-cited papers-relate to their early-career success. One key finding is that the postdoc period seems more important than the doctoral training to achieve this form of success. This is especially interesting in light of the many studies of academic faculty hiring that link Ph.D. granting institutions and hires, omitting the postdoc stage. Another group of findings can be summarized as a Goldilocks principle: It seems beneficial to change one's direction, but not too much.

Violence against women and girls research: Leveraging gains across disciplines

Violence against women and girls (VAWG) is a leading cause of mortality and morbidity worldwide, linked to numerous health, economic, and human rights outcomes. Target 5.2 of the Sustainable Development Goals calls for elimination of all forms of VAWG; however, progress toward achieving this goal has been inadequate. A lack of sufficient data and evidence has hindered global efforts to meet this target and hold governments accountable for action. While there have been substantial advancements in VAWG research methodology over the past three decades, researchers from diverse disciplines tend to work in silos, inhibiting progress in VAWG research. To address this challenge, we offer four key recommendations to support researchers in expanding transdisciplinary approaches: 1) leverage insights from a variety of VAWG data sources, 2) improve precision of VAWG definitions and outcomes, 3) create strategies to address underreporting, and 4) advance research ethics and equity. We conclude with a call to action for researchers, institutions, and donors to advance transdisciplinary research and foster collaboration, learning, and cross-fertilization across scientific fields to accelerate VAWG prevention efforts now and for future generations.

OCT-based diagnosis of glaucoma and glaucoma stages using explainable machine learning

Glaucoma poses a growing health challenge projected to escalate in the coming decades. However, current automated diagnostic approaches on Glaucoma diagnosis solely rely on black-box deep learning models, lacking explainability and trustworthiness. To address the issue, this study uses optical coherence tomography (OCT) images to develop an explainable artificial intelligence (XAI) tool for diagnosing and staging glaucoma, with a focus on its clinical applicability. A total of 334 normal and 268 glaucomatous eyes (86 early, 72 moderate, 110 advanced) were included, signal processing theory was employed, and model interpretability was rigorously evaluated. Leveraging SHapley Additive exPlanations (SHAP)-based global feature ranking and partial dependency analysis (PDA) estimated decision boundary cut-offs on machine learning (ML) models, a novel algorithm was developed to implement an XAI tool. Using the selected features, ML models produce an AUC of 0.96 (95% CI: 0.95-0.98), 0.98 (95% CI: 0.96-1.00) and 1.00 (95% CI: 1.00-1.00) respectively on differentiating early, moderate and advanced glaucoma patients. Overall, machine outperformed clinicians in the early stage and overall glaucoma diagnosis with 10.4 -11.2% higher accuracy. The developed user-friendly XAI software tool shows potential as a valuable tool for eye care practitioners, offering transparent and interpretable insights to improve decision-making.

Daily physical activity and cardiorespiratory fitness in adults with Down syndrome with and without congenital heart disease

BACKGROUND

Individuals with Down syndrome (DS) typically exhibit lower cardiorespiratory fitness and reduced moderate-to-vigorous physical activity (MVPA) compared to persons without disability. Approximately 50-55 % of individuals with DS have congenital heart disease (CHD), which is associated with cardiopulmonary deficiencies and reduced MVPA participation in non-DS populations. It is unknown if CHD related comorbidities compound with DS associated deficits in physical activity and fitness.

OBJECTIVE

To compare physical activity, cardiorespiratory fitness, and cardiovascular function, of persons with DS with and without CHD.

METHODS

Baseline data were used from a 12-month randomized controlled physical activity intervention of adults with DS. Participants with DS were age and sex matched based on presence of CHD. Measures of physical activity through accelerometry (n = 42; CHD, n = 21), cardiorespiratory fitness (VO2peak; n = 34, CHD n = 17), and cardiovascular function (anaerobic threshold, chronotropic index, O2 pulse; n = 34, CHD n = 17) were compared by CHD status using Wilcoxon rank sum tests.

RESULTS

There were no differences in VO2peak between those with and without CHD (CHD 20.3 ml/kg/min; no CHD 21.3 ml/kg/min, p = 0.44). MVPA was lower for those with CHD vs. without CHD (10.0 vs 13.3 min/week, p = 0.05). There were no differences in cardiovascular function by group.

CONCLUSION

Fitness and physical activity were low regardless of CHD status. Adults with DS and CHD may engage in less physical activity than those without CHD, however fitness and cardiovascular function were not further impaired by CHD. Given the prevalence of CHD in DS, it is important to include those with CHD in work increasing physical activity and fitness.

A scoping review of privacy and utility metrics in medical synthetic data

The use of synthetic data is a promising solution to facilitate the sharing and reuse of health-related data beyond its initial collection while addressing privacy concerns. However, there is still no consensus on a standardized approach for systematically evaluating the privacy and utility of synthetic data, impeding its broader adoption. In this work, we present a comprehensive review and systematization of current methods for evaluating synthetic health-related data, focusing on both privacy and utility aspects. Our findings suggest that there are a variety of methods for assessing the utility of synthetic data, but no consensus on which method is optimal in which scenario. Moreover, we found that most studies included in this review do not evaluate the privacy protection provided by synthetic data, and those that do often significantly underestimate the risks.