20-Hydroxyecdysone Modulates Bmp53-Mediated Apoptosis Regulation by Suppressing Mdm2-like-Dependent Ubiquitination in Silkworm, Bombyx mori

In the silkworm, 20-hydroxyecdysone (20E) induces apoptosis and autophagy, driving larval organ degeneration and remodeling. This mechanism may be a potential target for eco-friendly strategy for insect pests. However, a major challenge in harnessing this approach lies in the insufficient understanding of 20E's regulatory process in such a cell death mechanism. Our previous research has identified Bmp53 as a crucial gene in promoting the development of Bombyx mori during the pupal stage by inducing apoptosis, and has predicted the potential apoptotic regulatory network of Bmp53, wherein Mdm2-like ubiquitinating structural protein serves as a key component of this network. This study demonstrates that Mdm2-like acts as a ubiquitination regulatory protein, controlling its apoptosis-inducing activity via interaction with Bmp53. Moreover, co-expression of Mdm2-like and Bmp53 indirectly affects gene expression in the 20E-mediated apoptosis pathway. Further investigation revealed that Mdm2-like suppresses 20E-induced apoptosis by downregulating Bmp53 expression. This study reveals that the ubiquitination-mediated Mdm2-like/Bmp53 apoptosis pathway is a novel mechanism regulating silkworm apoptosis, with 20E playing a crucial role in this process. These findings enhance our understanding of the genetic mechanisms underlying tissue degradation during the metamorphic stage of the Bombyx mori. Additionally, these insights provide a theoretical reference for the development of environmentally friendly, hormone-based control strategies targeting protein modification as a means of managing lepidopteran pests.

Artificial Intelligence Chatbots as Sources of Implant Dentistry Information for the Public: Validity and Reliability Assessment

This study assessed the reliability and validity of responses from three chatbot systems-OpenAI's GPT-3.5, Gemini, and Copilot-concerning frequently asked questions (FAQs) in implant dentistry posed by patients.Twenty FAQs were prompted to three chatbots in three different times utilizing their respective application programming interfaces. The responses were assessed for validity (low and high threshold) and reliability by two prosthodontic consultants using a five-point Likert scale.The test of normality was utilized using the Shapiro-Wilk test. Differences between different chatbots regarding the quantitative variables in a given (fixed) time point and between the same chatbots in different time points were assessed using Friedman's two-way analysis of variance by ranks, followed by pairwise comparisons. All statistical analyses were conducted using the SPSS (Statistical Package for Social Sciences) Version 26.0 software program.GPT-3.5 provided the longest responses, while Gemini was the most concise. All chatbots advised consulting dental professionals more frequently. Validity was high under the low-threshold test but low under the high-threshold test, with Copilot scoring the highest. Reliability was high for all, with Gemini achieving perfect consistency.Chatbots showed consistent and generally valid responses with some variability in accuracy and details. While the chatbots demonstrated a high degree of reliability, their validity-especially under high-threshold criterion-remains limited. Improvements in accuracy and comprehensiveness are necessary for more effective use in providing information about dental implants.

Deep learning models based on multiparametric magnetic resonance imaging and clinical parameters for identifying synchronous liver metastases from rectal cancer

OBJECTIVES

To establish and validate deep learning (DL) models based on pre-treatment multiparametric magnetic resonance imaging (MRI) images of primary rectal cancer and basic clinical data for the prediction of synchronous liver metastases (SLM) in patients with Rectal cancer (RC).

METHODS

In this retrospective study, 176 and 31 patients with RC who underwent multiparametric MRI from two centers were enrolled in the primary and external validation cohorts, respectively. Clinical factors, including sex, primary tumor site, CEA level, and CA199 level were assessed. A clinical feature (CF) model was first developed by multivariate logistic regression, then two residual network DL models were constructed based on multiparametric MRI of primary cancer with or without CF incorporation. Finally, the SLM prediction models were validated by 5-fold cross-validation and external validation. The performance of the models was evaluated by decision curve analysis (DCA) and receiver operating characteristic (ROC) analysis.

RESULTS

Among three SLM prediction models, the Combined DL model integrating primary tumor MRI and basic clinical data achieved the best performance (AUC = 0.887 in primary study cohort; AUC = 0.876 in the external validation cohort). In the primary study cohort, the CF model, MRI DL model, and Combined DL model achieved AUCs of 0.816 (95% CI: 0.750, 0.881), 0.788 (95% CI: 0.720, 0.857), and 0.887 (95% CI: 0.834, 0.940) respectively. In the external validation cohort, the CF model, DL model without CF, and DL model with CF achieved AUCs of 0.824 (95% CI: 0.664, 0.984), 0.662 (95% CI: 0.461, 0.863), and 0.876 (95% CI: 0.728, 1.000), respectively.

CONCLUSION

The combined DL model demonstrates promising potential to predict SLM in patients with RC, thereby making individualized imaging test strategies.

CLINICAL RELEVANCE STATEMENT

Accurate synchronous liver metastasis (SLM) risk stratification is important for treatment planning and prognosis improvement. The proposed DL signature may be employed to better understand an individual patient's SLM risk, aiding in treatment planning and selection of further imaging examinations to personalize clinical decisions.

CLINICAL TRIAL NUMBER

Not applicable.

LitSense 2.0: AI-powered biomedical information retrieval with sentence and passage level knowledge discovery

LitSense 2.0 (https://www.ncbi.nlm.nih.gov/research/litsense2/) is an advanced biomedical search system enhanced with dense vector semantic retrieval, designed for accessing literature on sentence and paragraph levels. It provides unified access to 38 million PubMed abstracts and 6.6 million full-length articles in the PubMed Central (PMC) Open Access subset, encompassing 1.4 billion sentences and ∼300 million paragraphs, and is updated weekly. Compared to PubMed and PMC, the primary platforms for biomedical information search, LitSense offers cross-platform functionality by searching seamlessly across both PubMed and PMC and returning relevant results at a more granular level. Building on the success of the original LitSense launched in 2018, LitSense 2.0 introduces two major enhancements. The first is the addition of paragraph-level search: users can now choose to search either against sentences or against paragraphs. The second is improved retrieval accuracy via a state-of-the-art biomedical text encoder, ensuring more reliable identification of relevant results across the entire biomedical literature.

Superficial Neuromodulation in Dysautonomia in Women with Post-COVID-19 Condition: A Pilot Study

Post-COVID-19 condition involves persistent symptoms after acute infection, often linked to dysautonomia, which affects heart rate variability, pain perception, fatigue, and sleep. Superficial neuromodulation has been proposed as a treatment. Objective: To assess the effects of superficial neuromodulation on symptoms, sleep quality, and autonomic function in post-COVID-19 condition patients. Methods: A pilot study was conducted based on a triple-blind randomized controlled trial methodology involving 16 female participants. The experimental group received neuromodulation, while the control group used a placebo device. The intervention spanned 15 sessions over two months. Primary outcomes included heart rate variability, pain threshold, cortisol levels, fatigue, sleep quality, and quality of life, analyzed using repeated-measures ANOVA. Results: Both groups improved over time. Heart rate variability (SDNN) increased in the experimental group (30.42 to 39.11 ms) but decreased in controls (31.88 to 28.73 ms) (p < 0.05). Pain threshold at C5-C6 improved in the experimental group (2.1 to 3.5 kg/cm2) but remained stable in controls (p = 0.032). Fatigue decreased significantly in both groups (p = 0.002). Sleep quality improved, with Pittsburgh Sleep Quality Index scores decreasing similarly in both groups. Cortisol levels increased, with a non-significant trend favoring controls. Conclusions: While improvements were seen, both groups benefited, suggesting a possible placebo effect. Superficial neuromodulation appears safe, but further studies with larger samples are needed to confirm efficacy.

Proteomic Analysis of Invasive Breast Cancer Cells Treated with CBD Reveals Proteins Associated with the Reversal of Their Epithelial-Mesenchymal Transition Induced by IL-1β

Cannabidiol (CBD) has shown promise in treating cancers with an inflammatory microenvironment. Although it has been demonstrated that IL-1β induces epithelial-mesenchymal transition (EMT) of MCF-7 cells and CBD reverts this process, in restoring the epithelial non-invasive phenotype, there is limited understanding of how this cannabinoid regulates these processes. In this work, MCF-7 cells were induced to adopt an aggressive phenotype (6D cells), which was reversed by CBD. Then, protein expression was analyzed by mass spectrometry to compare 6D vs. MCF-7 cells and 6D+CBD vs. 6D cells proteomes. Novel proteins associated with EMT and CBD signaling were identified. Twenty-four of them were oppositely regulated by IL-1β and CBD, suggesting new points of crosstalk between the IL-1β and CBD signaling pathways. From the data, two protein networks were constructed: one related to EMT with 58 up-regulated proteins and another with 21 related to CBD signaling. The first one showed the proteins BRCA1, MSN, and CORO1A as the key axis that contributes to the establishment of a mesenchymal phenotype. In the CBD signaling, the key axis was formed by SUPT16H, SETD2, and H2BC12, which suggests epigenetic regulation by CBD in the restoration of an epithelial phenotype of breast cancer cells, providing new targets for anticancer therapy.

Immune-metabolic crosstalk in HNSCC: mechanisms and therapeutic opportunities

Head and neck squamous cell carcinoma (HNSCC) is a prevalent malignancy, characterized by metabolic reprogramming. This reprogramming creates an acidic and hypoxic environment within tumor cells to adapt to metabolic changes. Experimental data indicate that in HNSCC, the metabolic reprogramming of tumor cells regulates immune cells via metabolites or signaling pathways, thereby promoting cancer progression or immune evasion. This article reviews the metabolic reprogramming in HNSCC, including glucose, fatty acids, amino acids, and nucleotide metabolism. These metabolic pathways play crucial roles in the proliferation, differentiation, and effector functions of immune cells, and influence immunosuppressive checkpoints. Additionally, this review explores the potential relationships between metabolic reprogramming, tumor immunity, and related treatments. Thus, targeting metabolic reprogramming and interactions between immune cells may help overcome therapeutic resistance in HNSCC patients.

Comparative evaluation of the accuracy and reliability of ChatGPT versions in providing information on Helicobacter pylori infection

Objective

This study aimed to evaluate the accuracy and reliability of responses provided by three versions of ChatGPT (ChatGPT-3.5, ChatGPT-4, and ChatGPT-4o) to questions related to Helicobacter pylori (Hp) infection, as well as to explore their potential applications within the healthcare domain.

Methods

A panel of experts compiled and refined a set of 27 clinical questions related to Hp. These questions were presented to each ChatGPT version, generating three distinct sets of responses. The responses were evaluated and scored by three gastroenterology specialists utilizing a 5-point Likert scale, with an emphasis on accuracy and comprehensiveness. To assess response stability and reliability, each question was submitted three times over three consecutive days.

Results

Statistically significant differences in the Likert scale scores were observed among the three ChatGPT versions (p < 0.0001). ChatGPT-4o demonstrated the best performance, achieving an average score of 4.46 (standard deviation 0.82) points. Despite its high accuracy, ChatGPT-4o exhibited relatively low repeatability. In contrast, ChatGPT-3.5 exhibited the highest stability, although it occasionally provided incorrect answers. In terms of readability, ChatGPT-4 achieved the highest Flesch Reading Ease score of 24.88 (standard deviation 0.44), however, no statistically significant differences in readability were observed among the versions.

Conclusion

All three versions of ChatGPT were effective in addressing Hp-related questions, with ChatGPT-4o delivering the most accurate information. These findings suggest that artificial intelligence-driven chat models hold significant potential in healthcare, facilitating improved patient awareness, self-management, and treatment compliance, as well as supporting physicians in making informed medical decisions by providing accurate information and personalized recommendations.

The impact of different task contexts on emergency responders' trust and usage intention of artificial intelligence

Proper use of artificial intelligence (AI) can significantly enhance emergency responders' performance. However, they do not always trust or appropriately use AI. This study examined emergency responders' trust in AI and usage intention under different rescue pressures and uncertainty from the perspective of perceived capability. This study was conducted in two phases: first, questionnaire data were collected from 99 firefighters; second, semi-structured interviews were conducted with 12 participants. Results revealed that rescue pressure affected perceived AI capability, whereas uncertainty influenced perceived self-capability. Rescue pressure affected perceived AI capability, which subsequently impacted trust, and ultimately, usage intention. These findings explain the process through which rescue pressure impacts willingness to use AI. This study also explores the psychological mechanisms of trust and usage and provides valuable recommendations for designers to develop AI systems suitable for emergency responders.

Multilevel risk analysis of postoperative pulmonary complications following mandibular fractures: a retrospective cohort study based on patient characteristics and healthcare system factors

BACKGROUND

Pulmonary complications (PPCs) following mandibular fractures are serious post-surgery problems. This study analyzed risk factors of PPCs following mandibular fractures using the National Inpatient Sample (NIS) database, aiming to help clinicians specify surgical protocols and postoperative care for patients.

METHOD

A retrospective cohort study was conducted to examine patient demographics, hospital characteristics and preoperative comorbidities for identifying risk factors associated with postoperative pulmonary complications (PPCs). The analysis utilized data from the National Inpatient Sample (NIS) database containing patients undergoing mandibular surgery between 2010 and 2019. The cohort was stratified into two groups: those with PPCs and non-PPC cases. Statistical associations were evaluated through univariate and multivariate logistic regression analyses. A threshold of P ≤ 0.001 was set for statistical significance.

RESULTS

The study included 41,984 adult patients (33,017 male; 8,967 female; aged ≥ 18 years), with 3,514 cases of postoperative pulmonary complications (PPCs) subclassified as: 1,347 pneumonia, 2,452 acute respiratory failure (ARF), and 212 pulmonary embolism (PE). For patients with PPCs, there was a significant increase in the age by 8 years, length of stay (LOS) by 12 days, the total charge (TOTCHG) by $163,579, and the mortality rate by 8.9%. Following the analysis, the following risk factors and their incidence were identified: number of comorbidities ≥ 3 (OR = 3.72, 40.4%), fluid and electrolyte disorders (OR = 2.66, 46.7%), obesity (OR = 1.38, 5.0%), congestive heart failure (OR = 1.24, 4.4%), coagulopathy (OR = 1.94, 12.4%), peripheral vascular disorders (OR = 1.53, 5.7%), pulmonary circulation disorders (OR = 7 .93, 4.1%), respiratory diseases (OR = 3.93, 5.2%), other neurological disorders (OR = 1.57, 15.2%), and paralysis (OR = 2.43, 5.0%).

CONCLUSION

In this study, statistical methods were employed to identify the risk factors for pulmonary complications following mandibular fractures, which can aid in the establishment of a sound surgical procedure and postoperative care.

Post-acute sequelae SARS-CoV-2 infection and neuropathic pain: a narrative review of the literature and future directions

PURPOSE OF REVIEW

Neuropathic pain is a recognized and debilitating symptom of SARS-CoV-2 infection across acute, post-acute, and long-COVID phases. Initially emerging as acute or subacute symptoms, these neuropathic manifestations can evolve into chronic conditions, with approximately 10% of all SARS-CoV-2 cases (estimated 65 million individuals globally) developing post-acute SARS-CoV-2 (PASC) neuropathic sequalae. Given the limited literature specifically addressing neuropathic pain related to PASC, a deeper understanding is needed to improve management and reduce patient burden.

RECENT FINDINGS

PASC symptoms are associated with disease severity, elevated body mass indexes, preexisting psychological conditions, and addiction history. Sex differences appear to influence prevalence, and the multisystem nature of PASC complicates symptom presentation, with mood disorders, fatigue, and cognitive dysfunction contributing to altered pain perception. Proposed mechanisms include immune dysregulation, persistent viral protein effects, and neuroanatomical changes. Management typically involves a multimodal approach.

SUMMARY

This review examines SARS-CoV-2 neuropathic pain across the illness trajectory, examining its pathophysiology, prevalence, and treatment. It highlights the potential for subacute neuropathic symptoms to become chronic and calls for future research to refine long-term management strategies and assess broader healthcare implications.

Peri-intubation complications in critically ill obese patients: a secondary analysis of the international INTUBE cohort

BACKGROUND

Airway management in critically ill obese patients is potentially associated with a higher risk of adverse events due to a constellation of physiological and anatomical challenges. Data from international prospective studies on peri-intubation adverse events in obese critically ill patients are lacking.

METHODS

INTUBE (International Observational Study to Understand the Impact and Best Practices of Airway Management In Critically Ill Patients) was an international multicentre prospective cohort study enrolling critically ill adult patients undergoing in-hospital tracheal intubation in 197 sites from 29 countries worldwide from October 1, 2018, to July 31, 2019. This secondary analysis compares airway management practices and outcomes between obese (body mass index-BMI ≥ 30 kg/m2) and non-obese patients (BMI < 30 kg/m2).

RESULTS

A total of 2946 patients met inclusion criteria for this secondary analysis, 639 (21.7%) obese and 2307 (78.3%) non-obese. Severe peri-intubation hypoxemia was more frequently reported in obese compared to non-obese patients (12.1% vs 8.6% respectively, p = 0.01). Variables independently associated with a higher risk of peri-intubation hypoxemia were baseline SpO2/FiO2 (OR 0.996, 95% CI 0.994-0.997), 30-45° head-up position (OR 1.53, 95% CI 1.04-2.26) and first-pass intubation failure (OR for first-pass success 0.21, 95% CI 0.15-0.29). Obesity (OR 0.71, 95% CI 0.56-0.91) and 20° head-up position (OR 0.67, 95% CI 0.47-0.95) were independently associated with higher likelihood of first-pass intubation failure. In contrast, intubation by staff physician/consultant (OR 1.70, 95% CI 1.30-2.21) or anesthesiologists (OR 1.98, 95% CI 1.55-2.53) were associated with higher first-pass success.

CONCLUSIONS

Compared to non-obese patients, obese critically ill exhibit a higher incidence of peri-intubation severe hypoxemia. In this population, worse baseline oxygenation and first-pass intubation failure significantly increase the risk of peri-intubation severe hypoxemia. As obesity is linked to a higher likelihood of first-pass intubation failure, likely driven by more challenging airway features, in this high-risk population first attempt should be performed by an expert operator to minimize peri-intubation complications.

TRIAL REGISTRATION

Clinicaltrials.gov NCT03616054 . Registered 3 August 2018.

Nucleolar sequestration of cannabinoid type-2 receptors in triple-negative breast cancer cells

Multiple investigations have shown that the different types of cannabinoids, phytocannabinoids, synthetic cannabinoids, and endocannabinoids, possess antiproliferative and anticancer properties. The cannabinoid type-2 receptor (CB2R) has been proposed as a central player in tumor progression and has been correlated with the aggressiveness of breast cancer. Using immunocytochemistry and confocal microscopy, in the present work, we studied the expression level and subcellular localization of CB2R in two human triple-negative breast cancer (TNBC) cell lines, corresponding to early (stage I, HCC-1395) and metastatic (MDA-MB-231) stages, and they were compared with a non-tumoral mammary epithelial cell line (MCF-10A). We found that although CB2R was detected at the plasma membrane, it was mainly localized intracellularly, with ~40-fold higher expression in both TNBC cell lines than in MCF-10A (P < 0.0001). Notably, double staining with DAPI or with the nucleoli-specific fluorescent marker (3xnls-mTurquoise2) showed that most of the CB2R overexpressed in the nucleoli of cancer cells. This finding is supported by the fact that CB2R expression was markedly lower in mitotic cells compared to interphase cells (P < 0.0001). Interestingly, exposure of cancer cells to the specific agonist HU-308 reversed the nucleolar sequestration of CB2R while increasing the presence of the receptor in the nucleoplasm and cytoplasm (P < 0.0001). In addition, we found that this agonist reduced both the cell migration (P < 0.05-0.0001) and proliferation (P < 0.001) of TNBC cells. It remains to determine the function and signaling ability of CB2R in the nucleolus. Although our study only includes cell lines (tumoral and non-tumoral), we consider that this feature of nucleolar sequestration of CB2R could be a potential diagnostic marker for TNBC from the early stage.

Cortical signatures linked to behavior quantitatively track arousal levels

While current arousal level assessments in patients with disorders of consciousness discriminate altered states of consciousness, there are significant limitations in characterizing the transition from one state to another or quantifying the frequent arousal level fluctuations observed in a patient. Here, we identified a repeated, temporally discrete, dynamical pattern evident in the recovery of consciousness from anesthesia and brain injury coma models in rodents. We prospectively validated these features we label "Arousal Units" (AU) in neonatal humans recovering from static hypoxic injuries and senior patients emerging from anesthesia indicating their generalizability. The AUs lawfully link changes in spectral power and breathing frequency and reliably associate with motor changes. Distinctive cortical patterns within AUs can be transformed into arousal indices, determining arousal levels. The reliability of these events is demonstrated across intact and brain-injured states and translates to the human brain; extracting these stereotyped dynamics could aid anesthesia monitoring, tracking coma recovery, and identifying cognitive motor dissociation.

Large Language Models and Artificial Neural Networks for Assessing 1-Year Mortality in Patients With Myocardial Infarction: Analysis From the Medical Information Mart for Intensive Care IV (MIMIC-IV) Database

BACKGROUND

Accurate mortality risk prediction is crucial for effective cardiovascular risk management. Recent advancements in artificial intelligence (AI) have demonstrated potential in this specific medical field. Qwen-2 and Llama-3 are high-performance, open-source large language models (LLMs) available online. An artificial neural network (ANN) algorithm derived from the SWEDEHEART (Swedish Web System for Enhancement and Development of Evidence-Based Care in Heart Disease Evaluated According to Recommended Therapies) registry, termed SWEDEHEART-AI, can predict patient prognosis following acute myocardial infarction (AMI).

OBJECTIVE

This study aims to evaluate the 3 models mentioned above in predicting 1-year all-cause mortality in critically ill patients with AMI.

METHODS

The Medical Information Mart for Intensive Care IV (MIMIC-IV) database is a publicly available data set in critical care medicine. We included 2758 patients who were first admitted for AMI and discharged alive. SWEDEHEART-AI calculated the mortality rate based on each patient's 21 clinical variables. Qwen-2 and Llama-3 analyzed the content of patients' discharge records and directly provided a 1-decimal value between 0 and 1 to represent 1-year death risk probabilities. The patients' actual mortality was verified using follow-up data. The predictive performance of the 3 models was assessed and compared using the Harrell C-statistic (C-index), the area under the receiver operating characteristic curve (AUROC), calibration plots, Kaplan-Meier curves, and decision curve analysis.

RESULTS

SWEDEHEART-AI demonstrated strong discrimination in predicting 1-year all-cause mortality in patients with AMI, with a higher C-index than Qwen-2 and Llama-3 (C-index 0.72, 95% CI 0.69-0.74 vs C-index 0.65, 0.62-0.67 vs C-index 0.56, 95% CI 0.53-0.58, respectively; all P<.001 for both comparisons). SWEDEHEART-AI also showed high and consistent AUROC in the time-dependent ROC curve. The death rates calculated by SWEDEHEART-AI were positively correlated with actual mortality, and the 3 risk classes derived from this model showed clear differentiation in the Kaplan-Meier curve (P<.001). Calibration plots indicated that SWEDEHEART-AI tended to overestimate mortality risk, with an observed-to-expected ratio of 0.478. Compared with the LLMs, SWEDEHEART-AI demonstrated positive and greater net benefits at risk thresholds below 19%.

CONCLUSIONS

SWEDEHEART-AI, a trained ANN model, demonstrated the best performance, with strong discrimination and clinical utility in predicting 1-year all-cause mortality in patients with AMI from an intensive care cohort. Among the LLMs, Qwen-2 outperformed Llama-3 and showed moderate predictive value. Qwen-2 and SWEDEHEART-AI exhibited comparable classification effectiveness. The future integration of LLMs into clinical decision support systems holds promise for accurate risk stratification in patients with AMI; however, further research is needed to optimize LLM performance and address calibration issues across diverse patient populations.

Global Health care Professionals' Perceptions of Large Language Model Use In Practice: Cross-Sectional Survey Study

Background

ChatGPT is a large language model-based chatbot developed by OpenAI. ChatGPT has many potential applications to health care, including enhanced diagnostic accuracy and efficiency, improved treatment planning, and better patient outcomes. However, health care professionals' perceptions of ChatGPT and similar artificial intelligence tools are not well known. Understanding these attitudes is important to inform the best approaches to exploring their use in medicine.

Objective

Our aim was to evaluate the health care professionals' awareness and perceptions regarding potential applications of ChatGPT in the medical field, including potential benefits and challenges of adoption.

Methods

We designed a 33-question online survey that was distributed among health care professionals via targeted emails and professional Twitter and LinkedIn accounts. The survey included a range of questions to define respondents' demographic characteristics, familiarity with ChatGPT, perceptions of this tool's usefulness and reliability, and opinions on its potential to improve patient care, research, and education efforts.

Results

One hundred and fifteen health care professionals from 21 countries responded to the survey, including physicians, nurses, researchers, and educators. Of these, 101 (87.8%) had heard of ChatGPT, mainly from peers, social media, and news, and 77 (76.2%) had used ChatGPT at least once. Participants found ChatGPT to be helpful for writing manuscripts (n=31, 45.6%), emails (n=25, 36.8%), and grants (n=12, 17.6%); accessing the latest research and evidence-based guidelines (n=21, 30.9%); providing suggestions on diagnosis or treatment (n=15, 22.1%); and improving patient communication (n=12, 17.6%). Respondents also felt that the ability of ChatGPT to access and summarize research articles (n=22, 46.8%), provide quick answers to clinical questions (n=15, 31.9%), and generate patient education materials (n=10, 21.3%) was helpful. However, there are concerns regarding the use of ChatGPT, for example, the accuracy of responses (n=14, 29.8%), limited applicability in specific practices (n=18, 38.3%), and legal and ethical considerations (n=6, 12.8%), mainly related to plagiarism or copyright violations. Participants stated that safety protocols such as data encryption (n=63, 62.4%) and access control (n=52, 51.5%) could assist in ensuring patient privacy and data security.

Conclusions

Our findings show that ChatGPT use is widespread among health care professionals in daily clinical, research, and educational activities. The majority of our participants found ChatGPT to be useful; however, there are concerns about patient privacy, data security, and its legal and ethical issues as well as the accuracy of its information. Further studies are required to understand the impact of ChatGPT and other large language models on clinical, educational, and research outcomes, and the concerns regarding its use must be addressed systematically and through appropriate methods.

Trends and Scientific Production on Isometric Training: A Bibliometric Analysis

Isometric training is a method focused on muscle strengthening without joint movement and has gained attention in recent years due to its applicability in rehabilitation and sports medicine. However, no comprehensive bibliometric analysis focused exclusively on adult populations has been performed. This study aimed to analyze the scientific production related to isometric training in adults; identify prominent authors, journals, and thematic trends; and evaluate the evolution of interest in this field over time. A bibliometric review was performed using the Web of Science Core Collection (SCI-E, SSCI, and ESCI). A specific search strategy was applied to identify articles and reviews focused on isometric training in adult populations. A total of 238 records met the inclusion criteria. Data were analyzed using Excel 2016 and VOSviewer software1.6.20. Bibliometric indicators such as Price's Law, Bradford's Law, Lotka's Law, h-index, and co-occurrence and co-authorship network analysis were applied. The results showed a steady increase in publications in the last decade, highlighting the categories of Sports Science, Physiology, and Cardiovascular. The Journal of Applied Physiology was the most frequent source, and Springer Nature was the most prolific publisher. The h-index identified 21 highly cited papers, and Lotka's Law confirmed the existence of a small group of prolific authors. VOSviewer analysis revealed clear thematic clusters, mainly around blood pressure regulation, rehabilitation, and aging. International collaboration was evident, with the United States, Canada, and the United Kingdom leading the co-authorship networks. Scientific interest in isometric training for adult populations is growing, particularly in relation to cardiovascular health and rehabilitation. Despite this, gaps remain in terms of methodological consistency and standardized protocols. Addressing these issues could improve the applicability and scientific impact of this training modality.

Transforming beef quality through healthy breeding: a strategy to reduce carcinogenic compounds and enhance human health: a review

The presence of carcinogenic substances in beef poses a significant risk to public health, with far-reaching implications for consumer safety and the meat production industry. Despite advancements in food safety measures, traditional breeding methods have proven inadequate in addressing these risks, revealing a substantial gap in knowledge. This review aims to fill this gap by evaluating the potential of healthy breeding techniques to significantly reduce the levels of carcinogenic compounds in beef. We focus on elucidating the molecular pathways that contribute to the formation of key carcinogens, such as heterocyclic amines (HCAs) and polycyclic aromatic hydrocarbons (PAHs), while exploring the transformative capabilities of advanced genomic technologies. These technologies include genomic selection, CRISPR/Cas9, base editing, prime editing, and artificial intelligence-driven predictive models. Additionally, we examine multi-omics approaches to gain new insights into the genetic and environmental factors influencing carcinogen formation. Our findings suggest that healthy breeding strategies could markedly enhance meat quality, thereby offering a unique opportunity to improve public health outcomes. The integration of these innovative technologies into breeding programs not only provides a pathway to safer beef production but also fosters sustainable livestock management practices. The improvement of these strategies, along with careful consideration of ethical and regulatory challenges, will be crucial for their effective implementation and broader impact.

Cannabidiol-loaded-injectable depot formulation for the treatment of triple-negative breast cancer: design, development, in-vitro and in-ovo evaluation of its anticancer activity

Triple-negative breast cancer (TNBC) is an invasive and difficult-to-treat carcinoma that represents 15-20 % of breast malignancies and is frequently diagnosed in younger women. Chemotherapy is the mainstay treatment approach. Cannabidiol (CBD), the main non-psychoactive cannabinoid, has shown a potential anticancer activity in TNBC, enhancing the effect of conventional antineoplastics. This research aims to develop in situ forming implants (ISFIs) as a long-acting depot formulation of CBD with potential application in TNBC. This formulation is intended to be administered in the tumor site during neoadjuvant chemotherapeutic regimens, allowing a controlled CBD release. ISFIs were elaborated with 100 mg of polycaprolactone (PCL) and 2.5 mg (2.5-CB-ISFI), 5 mg (5-CB-ISFI) or 10 mg (10-CB-ISFI) of CBD dissolved in 400 µL of NMP. All the formulations exhibited a controlled drug release for around two months. 10-CB-ISFI formulation with the highest CBD content and the most suitable CBD release profile was selected for biological studies. This formulation inhibited the proliferation and migration of MDA-MB-231 and 4T1 cells and exerted an antiangiogenic effect in ovo. Interestingly, the antiangiogenic activity of 10-CB-ISFI was higher compared with CBD in solution administered at the same concentration, showing vascular inhibition percentages of around 80 % and 60 %, respectively. Finally, this formulation reduced the growth of MDA-MB-231-derived tumors developed in the chorioallantoic membrane (CAM) model. The single administration of 10-CB-ISFI exhibited a similar antitumor efficacy to the daily administration of CBD in solution (≈60 % tumor growth inhibition). Therefore, the injectable depot formulation of CBD developed in this work showed a promising utility in TNBC treatment.

Machine Learning Clinical Decision Support for Interdisciplinary Multimodal Chronic Musculoskeletal Pain Treatment: Prospective Pilot Study of Patient Assessment and Prognostic Profile Validation

Background

Chronic musculoskeletal pain (CMP) impacts around 20% of people globally, resulting in patients living with pain, fatigue, restricted social and employment capacity, and reduced quality of life. Interdisciplinary multimodal pain treatment (IMPT) programs have been shown to provide positive and sustained outcomes where all other interventions have failed. IMPT programs combined with multidimensional machine learning predictive patient profiles aim to improve clinical decision support and personalized patient assessments, potentially leading to better treatment outcomes.

Objective

We aimed to investigate integrating machine learning with IMPT programs and its potential contribution to clinical decision support and treatment outcomes for patients with CMP.

Methods

This prospective pilot study used a machine learning prognostic patient profile of 7 outcome measures across 4 clinically relevant domains, including activity or disability, pain, fatigue, and quality of life. Prognostic profiles were created for new IMPT patients in the Netherlands in November 2023 (N=17). New summary indicators were developed, including defined categories for positive, negative, and mixed prognostic profiles; an accuracy indicator with high, medium, and low levels based on weighted true- or false-positive values; and an indicator for consistently positive or negative outcomes. The consolidated reporting guidelines checklist for prognostic machine learning modeling studies was completed to provide transparency of data quality, model development methodology, and validation.

Results

The machine learning IMPT prognostic patient profiles demonstrated high accuracy and consistency in predicting patient outcomes. The profile, combined with extended new prognostic summary indicators, provided improved identification of patients with predicted positive, negative, and mixed outcomes, supporting more comprehensive assessment. Overall, 82.4% (14/17) of prognostic patient profiles were consistent with clinician assessments. Notably, clinician case notes indicated the stratified prognostic profiles were directly discussed with around half (8/17, 47.1%) of patients. Clinicians found the prognostic patient profiles helpful in 88.2% (15/17) of initial IMPT assessments to support shared clinician and patient decision-making and discussion of individualized treatment planning.

Conclusions

Machine learning prognostic patient profiles showed promising contributions for IMPT clinical decision support and improving treatment outcomes for patients with CMP. Further research is needed to validate these findings in larger, more diverse populations.

Assessing the Impact on Electronic Health Record Burden After Five Years of Physician Engagement in a Canadian Mental Health Organization: Mixed-Methods Study

Background

The burden caused by the use of electronic health record (EHR) systems continues to be an important issue for health care organizations, especially given human resource shortages in health care systems globally. As physicians report spending 2 hours documenting for every hour of patient care, there has been strong interest from many organizations to understand and address the root causes of physician burnout due to EHR burden.

Objective

This study focuses on evaluating physician burnout related to EHR usage and the impact of a physician engagement strategy at a Canadian mental health organization 5 years after implementation.

Methods

A cross-sectional survey was conducted to assess the perceived impact of the physician engagement strategy on burnout associated with EHR use. Physicians were invited to participate in a web-based survey that included the Mini-Z Burnout questionnaire, along with questions about their perceptions of the EHR and the effectiveness of the initiatives within the physician engagement strategy. Descriptive statistics were applied to analyze the quantitative data, while thematic analysis was used for the qualitative data.

Results

Of the 254 physicians invited, 128 completed the survey, resulting in a 50% response rate. Among the respondents, 26% (33/128) met the criteria for burnout according to the Mini-Z questionnaire, with 61% (20/33) of these attributing their burnout to EHR use. About 52% of participants indicated that the EHR improves communication (67/128) and 38% agreed that the EHR enables high-quality care (49/128). Regarding the physician engagement strategy initiatives, 39% (50/128) agreed that communication through the strategy is efficient, and 75% (96/128) felt more proficient in using the EHR. However, additional areas for improvement within the EHR were identified, including (1) medication reconciliation and prescription processes; (2) chart navigation and information retrieval; (3) longitudinal medication history; and (4) technology infrastructure challenges.

Conclusions

This study highlights the potential impact of EHRs on physician burnout and the effectiveness of a unique physician engagement strategy in fostering positive perceptions and improving EHR usability among physicians. By evaluating this initiative in a real-world setting, the study contributes to the broader literature on strategies aimed at enhancing physician experience following large-scale EHR implementation. However, the findings indicate a continued need for system-level improvements to maximize the value and usage of EHRs. The physician engagement strategy demonstrates the potential to enhance physicians' EHR experience. Future efforts should prioritize system-level advancements to increase the EHR's impact on quality of care and develop standardized approaches for engaging physicians on a broader Canadian scale.

Neuroprotective potential of epigallocatechin gallate in Neurodegenerative Diseases: Insights into molecular mechanisms and clinical Relevance

Neurodegenerative diseases (NDs) such as Alzheimer's disease, Parkinson's disease, Huntington's disease, and amyotrophic lateral sclerosis pose significant challenges due to their complex pathophysiology and lack of effective treatments. Green tea, rich in the epigallocatechin gallate (EGCG) polyphenolic component, has demonstrated potential as a neuroprotective agent with numerous medicinal applications. EGCG effectively reduces tau and Aβ aggregation in ND models, promotes autophagy, and targets key signaling pathways like Nrf2-ARE, NF-κB, and MAPK. This review explores the molecular processes that underlie EGCG's neuroprotective properties, including its ability to regulate mitochondrial dysfunction, oxidative stress, neuroinflammation, and protein misfolding. Clinical research indicates that EGCG may enhance cognitive and motor abilities, potentially inhibiting disease progression despite absorption and dose optimization limitations. The substance has been proven to slow the amyloidogenic process, prevent protein aggregation, decrease amyloid cytotoxicity, inhibit fibrillogenesis, and restructure fibrils for synergistic therapeutic effects. The review highlights the potential of EGCG as a natural, multi-targeted strategy for NDs but emphasizes the need for further clinical trials to enhance its therapeutic efficacy.

The role of nanomedicine and artificial intelligence in cancer health care: individual applications and emerging integrations-a narrative review

Cancer remains one of the deadliest diseases globally, significantly impacting patients' quality of life. Addressing the rising incidence of cancer deaths necessitates innovative approaches such as nanomedicine and artificial intelligence (AI). The convergence of nanomedicine and AI represents a transformative frontier in cancer healthcare, promising unprecedented advancements in diagnosis, treatment, and patient management. This narrative review explores the distinct applications of nanomedicine and AI in oncology, alongside their synergistic potential. Nanomedicine leverages nanoparticles for targeted drug delivery, enhancing therapeutic efficacy while minimizing adverse effects. Concurrently, AI algorithms facilitate early cancer detection, personalized treatment planning, and predictive analytics, thereby optimizing clinical outcomes. Emerging integrations of these technologies could transform cancer care by facilitating precise, personalized, and adaptive treatment strategies. This review synthesizes current research, highlights innovative individual applications, and discusses the emerging integrations of nanomedicine and AI in oncology. The goal is to provide a comprehensive understanding of how these cutting-edge technologies can collaboratively improve cancer diagnosis, treatment, and patient prognosis.

Using Natural Language Processing and Machine Learning to classify the status of kidney allograft in Electronic Medical Records written in Spanish

INTRODUCTION

Accurate identification of graft loss in Electronic Medical Records of kidney transplant recipients is essential but challenging due to inconsistent and not mandatory International Classification of Diseases (ICD) codes. We developed and validated Natural Language Processing (NLP) and machine learning models to classify the status of kidney allografts in unstructured text in EMRs written in Spanish.

METHODS

We conducted a retrospective cohort of 2712 patients transplanted between July 2008 and January 2023, analyzing 117,566 unstructured medical records. NLP involved text normalization, tokenization, stopwords removal, spell-checking, elimination of low-frequency words and stemming. Data was split in training, validation and test sets. Data balance was performed using undersampling technique. Feature selection was performed using LASSO regression. We developed, validated and tested Logistic Regression, Random Forest, and Neural Networks models using 10-fold cross-validation. Performance metrics included area under the curve, F1 Score, accuracy, sensitivity, specificity, Negative Predictive Value, and Positive Predictive Value.

RESULTS

The test performance results showed that the Random Forest model achieved the highest AUC (0.98) and F1 score (0.65). However, it had a modest sensitivity (0.76) and a relatively low PPV (0.56), implying a significant number of false positives. The Neural Network model also performed well with a high AUC (0.98) and reasonable F1 score (0.61), but its PPV (0.49) was lower, indicating more false positives. The Logistic Regression model, while having the lowest AUC (0.91) and F1 score (0.49), showed the highest sensitivity (0.83) with the lowest PPV (0.35).

CONCLUSION

We developed and validated three machine learning models combined with NLP techniques for unstructured texts written in Spanish. The models performed well on the validation set but showed modest performance on the test set due to data imbalance. These models could be adapted for clinical practice, though they may require additional manual work due to high false positive rates.

Preoperative neoadjuvant chemotherapy or immunotherapy in head and neck cancer: A systematic review and meta-analysis of surgical risk and pathologic response

OBJECTIVE

This systematic review and meta-analysis aimed to evaluate surgical complications, pathologic responses, and disease progression in patients with head and neck squamous cell carcinoma treated with neoadjuvant chemotherapy or immunotherapy.

METHODS

A comprehensive literature search was conduct across four databases (PubMed, Embase, Cochrane Library, and Scopus) and grey literature sources to identify neoadjuvant therapies in head and neck cancer patients. Only prospective clinical trials were included. The certainty of evidence was appraised using GRADE tool.

RESULTS

A total of 12 clinical trials me the inclusion criteria, comprising, six studies on neoadjuvant chemotherapy (Cisplatin and 5-FU) and six on immunotherapy (Nivolumab, Nivolumab plus Ipilimumab, Pembrolizumab) were analyzed. The mean time from drug administration to surgery ranged from 18 to 29 days. The overall surgical complication rate was 32.8 %, with the lowest observed in the Pembrolizumab group (9 %) and the highest in the Nivolumab plus Ipilimumab group (36.7 %). However, risk ratios for surgical complications were not statistically significant for Nivolumab (RR = 1.68, p = 0.078) or chemotherapy (RR = 1.1, p = 0.70). The complete pathologic response (pCR) rate was low (4 %), highest in the Cisplatin and 5-FU group (11 %). In contrast, the partial pathologic response (pPR) rate reached 58 % with Nivolumab plus Ipilimumab. Disease progression after surgery occurred in 19.4 %, with the lowest progression rate observed in the Nivolumab plus Ipilimumab group (7.7 %). The certainty of evidence was rated as very low for chemotherapy and low for immunotherapy.

CONCLUSION

Combination immunotherapy, particularly Nivolumab with Ipilimumab, demonstrated favorable pPR rates and reduced disease progression but was increased surgical complications. The overall low pCR across all regimen treatments highlight the need for improved therapeutic strategies.

ChatOCT: Embedded Clinical Decision Support Systems for Optical Coherence Tomography in Offline and Resource-Limited Settings

Optical Coherence Tomography (OCT) is a critical imaging modality for diagnosing ocular and systemic conditions, yet its accessibility is hindered by the need for specialized expertise and high computational demands. To address these challenges, we introduce ChatOCT, an offline-capable, domain-adaptive clinical decision support system (CDSS) that integrates structured expert Q&A generation, OCT-specific knowledge injection, and activation-aware model compression. Unlike existing systems, ChatOCT functions without internet access, making it suitable for low-resource environments. ChatOCT is built upon LLaMA-2-7B, incorporating domain-specific knowledge from PubMed and OCT News through a two-stage training process: (1) knowledge injection for OCT-specific expertise and (2) Q&A instruction tuning for structured, interactive diagnostic reasoning. To ensure feasibility in offline environments, we apply activation-aware weight quantization, reducing GPU memory usage to ~ 4.74 GB, enabling deployment on standard OCT hardware. A novel expert answer generation framework mitigates hallucinations by structuring responses in a multi-step process, ensuring accuracy and interpretability. ChatOCT outperforms state-of-the-art baselines such as LLaMA-2, PMC-LLaMA-13B, and ChatDoctor by 10-15 points in coherence, relevance, and clinical utility, while reducing GPU memory requirements by 79%, while maintaining real-time responsiveness (~ 20 ms inference time). Expert ophthalmologists rated ChatOCT's outputs as clinically actionable and aligned with real-world decision-making needs, confirming its potential to assist frontline healthcare providers. ChatOCT represents an innovative offline clinical decision support system for optical coherence tomography (OCT) that runs entirely on local embedded hardware, enabling real-time analysis in resource-limited settings without internet connectivity. By offering a scalable, generalizable pipeline that integrates knowledge injection, instruction tuning, and model compression, ChatOCT provides a blueprint for next-generation, resource-efficient clinical AI solutions across multiple medical domains.

Drivers of the emergence and dissemination of high-risk resistance genes in cattle farm

Extended spectrum β-lactamase (ESBL)- and carbapenemase-producing Enterobacterales (CPE) are recognized by WHO as critical concerns. The high cephalosporin resistance rate in a cattle farm in 2018 prompted us to conduct long-term (2019-2023) and extensive monitoring to explore risk factors for the import and transmission of ESBLs and CPE in this farm. Among 1288 samples from cattle, the environment, milk, and biological vectors, 48.8 % carried blaCTX-M-positive Enterobacterales with blaCTX-M-55 being dominant (76.4 %), and blaNDM-5-positive strains emerged in 2022 with a 1.9 % detection rate. blaCTX-M-55 and blaNDM-5 were likely introduced through various routes, especially wild birds, and have persisted due to overuse of cephalosporins in the farm. The spread of these genes was driven by the horizontal transmission of IncHI2 and IncX3 plasmids and clonal dissemination of certain clones. Cross-regional and cross-border transmission of blaCTX-M-55- and/or blaNDM-5-bearing bacteria and plasmids possibly occurred via wild birds, animal trade, and other means. Our findings suggest that the import, persistence, and dissemination of these genes within and beyond this farm, were fueled by suboptimal biosecurity practices and inadequate antibiotic stewardship, highlighting the urgency for integrated public and ecosystem health policies to prevent the spread of resistance genes as part of a holistic One Health strategy. ENVIRONMENTAL IMPLICATION: The high prevalence and long-term persistence of extended-spectrum β-lactamases and the emergence of carbapenemases in cattle and the environment signify a critical risk of transmitting high-risk resistance genes, posing a significant threat to human health. Consequently, bacteria carrying these genes in animal farms should be regarded as "hazardous materials". Import, persistence, and dissemination of these genes within and beyond this farm were exacerbated by suboptimal biosecurity practices and inadequate antibiotic stewardship, highlighting the urgency for integrated public and ecosystem health policies to mitigate the environmental risks associated with gene transmission as part of a comprehensive One Health strategy.

Student perceptions of artificial intelligence in Doctor of Physical Therapy Education

PURPOSE

Health profession education is impacted by developments in artificial intelligence (AI). While AI provides useful supplemental material, it is unknown how it will affect learning complex clinical topics. Artificial intelligence has the potential to be a powerful tool when used in tandem with traditional human studying methods. This study aims to understand Doctor of Physical Therapy students' perceptions of AI as a study tool.

METHODS

Students participated in an online pre/post survey evaluating their perceptions of AI's ability to diagnose and treat clinical cases, and two corresponding research activities where students were asked to complete a neurologic case study using classroom notes and a separate case using AI. Descriptive statistics, Wilcoxon signed-rank tests, and effect sizes were calculated.

RESULTS

Twenty-three first year Doctor of Physical Therapy students volunteered to participate. In the post-survey, students' perceptions of AI included increased trust to provide a clinical diagnosis (p = .001; r = 0.66) and a treatment plan (p = .005; r = 0.58), and to consider patient-centered plans of care (p = .036; r = 0.44). There was no significant difference following the research activities about concern handling sensitive information (p = .499), replacing a human clinician (p = .255), or the need for human oversight (p = .833).

CONCLUSION

The use of AI as a study tool is rapidly developing. This study showed that students' perceptions of AI were significantly more positive after an exposure of clinical course application. These results are beneficial to educators to ensure that students are shown the practical applications of AI throughout learning strategies.

The Role of Xenobiotics and Anelloviruses in Colorectal Cancer: Mechanisms and Perspectives

Xenobiotics are non-natural chemical compounds to which the human population is exposed. Chronic exposure to certain xenobiotics is associated with various diseases, including cancer development. Anelloviruses (AVs), including Torque Teno Virus (TTV), Torque Teno Mini Virus (TTMV), and Torque Teno Midi Virus (TTMDV), are ubiquitous viruses found in the general population. As no disease has been definitively associated with AVs, they are sometimes referred to as "viruses awaiting a disease". This review explores the potential roles of xenobiotics and AVs in colorectal cancer (CRC) development and suggests a potential interplay between them. Evidence suggests an association between certain xenobiotics (like pesticides, cigarette smoke components, and dietary factors) and CRC, while such an association is less clear for AVs. The high prevalence of AVs suggests these infections alone may be insufficient to disrupt homeostasis; thus, additional factors might be required to promote disease, potentially including cancer.

Intravenous Administration of Remdesivir at the Acute Phase of SARS-CoV-2 Infection Is Associated with a Lower Prevalence of Post-COVID-19 Pain

Background/Objective: Evidence suggests that the administration of antivirals at the acute phase of SARS-CoV-2 infection is associated with lower COVID-19 severity, accordingly, the administration of antivirals at the acute phase of the infection could prevent post-COVID-19 symptoms. The current study investigated the effects of the intravenous administration of Remdesivir at hospitalization (acute phase of SARS-CoV-2 infection) in COVID-19 survivors on the development of post-COVID-19 pain symptoms. Methods: A cohort of previously hospitalized COVID-19 survivors who received intravenous administration of Remdesivir at the acute COVID-19 phase (n = 216) were matched with a cohort of previously hospitalized COVID-19 survivors who did not receive any antiviral treatment at the acute phase of the infection (n = 216). In a face-to-face interview, they were asked for the development of pain symptoms attributed to SARS-CoV-2 infection and whether the symptom persisted at the time of the study (mean follow-up: 18.4, SD: 0.8 months). Clinical/hospitalization data were collected from medical records. Anxiety/depressive symptoms and sleep quality were also assessed with validated self-reported questionnaires. Results: No differences in hospitalization data and the presence of previous chronic conditions were seen between patients receiving or not receiving intravenous administration of Remdesivir during hospitalization. The multivariate analysis revealed that the intravenous administration of Remdesivir at the acute COVID-19 phase was a protective factor for the development of overall post-COVID-19 pain (OR 0.444, 95% CI 0.292-0.674, p < 0.001). A protective effect of administrating intravenous Remdesivir was specifically seen for thorax/chest (OR 0.277, 95% CI 0.100-0.766, p = 0.01) and lumbar spine (OR 0.347, 95% CI 0.143-0.844, p = 0.02) pain. Conclusions: Current results support a potential protective role of the intravenous administration of Remdesivir at the acute phase of SARS-CoV-2 infection for developing long-term post-COVID-19 pain in previously hospitalized COVID-19 survivors. Studies investigating the effects of the oral administration of antivirals in non-hospitalized populations are needed to generalize these findings.

The Role of Claude 3.5 Sonet and ChatGPT-4 in Posterior Cervical Fusion Patient Guidance

BACKGROUND

This study evaluates the role of ChatGPT-4 and Claude 3.5 Sonet in postoperative management for patients undergoing posterior cervical fusion. It focuses on their ability to provide accurate, clear, and relevant responses to patient concerns, highlighting their potential as supplementary tools in surgical aftercare.

METHODS

Ten common postoperative questions were selected and posed to ChatGPT-4 and Claude 3.5 Sonet. Ten independent neurosurgeons evaluated responses using a structured framework that assessed accuracy, response time, clarity, and relevance. A 5-point Likert scale also measured satisfaction, quality, performance, and importance. Advanced statistical analyses were used to compare the 2 artificial intelligence platforms, including sensitivity, specificity, P values, confidence intervals, and Cohen's d.

RESULTS

Claude 3.5 Sonet outperformed ChatGPT-4 across all metrics, particularly in accuracy (96.5% vs. 80.6%), response time (92.9% vs. 76.4%), clarity (94.6% vs. 75.4%), and relevance (95.5% vs. 74.0%). Likert scale evaluations showed significant differences (P < 0.001) in satisfaction, quality, and performance, with Claude achieving higher ratings. Statistical analyses confirmed large effect sizes, high inter-rater reliability (kappa: 0.85-0.92 for Claude), and narrower confidence intervals, reinforcing Claude's consistency and superior performance.

CONCLUSIONS

Claude 3.5 Sonet demonstrated exceptional capability in addressing postoperative concerns for posterior cervical fusion patients, surpassing ChatGPT-4 in accuracy, clarity, and practical relevance. These findings underscore its potential as a reliable artificial intelligence tool for enhancing patient care and satisfaction in surgical aftercare.

SARS-CoV-2-induced sensory perturbations: A narrative review of clinical phenotypes, molecular pathologies, and possible interventions

Background

The acute and post-acute sequelae of SARS-CoV-2 infection have been of great clinical interest since the inception of the COVID-19 pandemic. Despite a high prevalence of individuals with persistent symptoms, a wholistic view of the effects of SARS-CoV-2 on special sensory systems is lacking. Considering the significant impact of normal sensory function on quality of life, the goal of this review is to highlight unresolved issues related to SARS-CoV-2-associated insults to the sensory nervous system.

Major findings

In this narrative review, we discuss the epidemiology of SARS-CoV-2-induced sensory perturbations, underlying pathological mechanisms, and possible therapeutic strategies across the olfactory, gustatory, somatosensory, visual, and auditory systems. Examined literature included studies with human biospecimens, human-derived cell lines, and naturally susceptible animal models, which highlighted evidence of persistent functional disruption in all sensory systems. SARS-CoV-2 infection was associated with persistent inflammation in the olfactory epithelium/bulb, somatosensory ganglia, and gustatory systems, long-term transcriptional perturbations in the sensory central nervous system and peripheral nervous system, and detectable degeneration/apoptosis in the gustatory and visual systems. Few studies have proposed evidence-based therapeutic strategies for attenuating specific sensory abnormalities after SARS-CoV-2 infection.

Conclusion

While the olfactory system, and to some extent the visual and somatosensory systems, have been more thoroughly investigated from symptomatology, behavioral and molecular perspectives, there is still an unmet need for the development of therapeutics to treat COVID-induced impairment of these systems. Further, additional attention must be placed on COVID-associated impairment of the gustatory, visual, and auditory systems, which lack detailed mechanistic investigations into their pathogenesis.

In silico docking, ADMET profiling, and bio-accessibility experimentation on Breynia retusa phytocompounds and in vitro validation for anti-proliferative potencies against ovarian carcinoma

This study aimed to assess the medicinal properties of Breynia retusa, a plant rich in phytocompounds predominantly used as an ethnomedicinal agent in Western Ghats, which appeared to be promising for therapeutic use, especially in the treatment of ovarian cancer. Herein, its cytotoxic potential on ovarian cancer cell lines SKOV-3, neurotoxicity, antioxidant activity, and molecular docking was determined to aid in explaining the mechanisms of interactions with proteins related to ovarian cancer. B . retusa methanolic extract demonstrated exuberant antioxidant activity, with 81.91% scavenging ability of DPPH radicals and efficient reduction of phosphomolybdenum (22.98 mg ascorbic acid equivalents antioxidant capacity/g extract). The extract proved to be an important anti-inflammatory agent through membrane stabilization inhibition of 83%. The cytotoxicity study against the SKOV-3 cell line indicated an IC50 value of 34.01 µg/mL and a very negligible neurotoxicity in SH-SY5Y cell lines. The GC-MS and HPLC profiling indicated many anticancer compounds in the extract such as secalciferol, methyl gallate, ricinoleic acid, gallic acid, and naringenin. The docking study showed significant interactions of secalciferol molecules with the key ovarian cancer proteins, which include IGF1 (-6.758 kcal/mol) and c-ERBB2 (-4.281 kcal/mol). Fatty acid derivatives and methyl gallate showed efficient dock scores (< -5.0 kcal/mol) with antioxidant (catalase and superoxide dismutase) enzymes and inflammatory cytokines (IL-6 and COX-1), respectively, as evidences of antioxidant and anti-inflammatory potentials. The bio-accessibility of phenolics and their antioxidant activity ranged above 90%, indicating the promising bioavailability of phytochemicals expected in vivo. Hence the current study emphasizes the anticancer potential of B. retusa phytocompounds that appeared to interact very strongly with ovarian cancer targets and confirms the dose-dependent cytotoxic and antioxidant activities of B. retusa methanolic extract.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13205-025-04276-8.

Salivary biomarkers: a promising approach for predicting immunotherapy response in head and neck cancers

Head and neck cancers, including cancers of the mouth, throat, voice box, salivary glands, and nose, are a significant global health issue. Radiotherapy and surgery are commonly used treatments. However, due to treatment resistance and disease recurrence, new approaches such as immunotherapy are being explored. Immune checkpoint inhibitors (ICIs) have shown promise, but patient responses vary, necessitating predictive markers to guide appropriate treatment selection. This study investigates the potential of non-invasive biomarkers found in saliva, oral rinses, and tumor-derived exosomes to predict ICI response in head and neck cancer patients. The tumor microenvironment significantly impacts immunotherapy efficacy. Oral biomarkers can provide valuable information on composition, such as immune cell presence and checkpoint expression. Elevated tumor mutation load is also associated with heightened immunogenicity and ICI responsiveness. Furthermore, the oral microbiota may influence treatment outcomes. Current research aims to identify predictive salivary biomarkers. Initial studies indicate that tumor-derived exosomes and miRNAs present in saliva could identify immunosuppressive pathways and predict ICI response. While tissue-based markers like PD-L1 have limitations, combining multiple oral fluid biomarkers could create a robust panel to guide treatment decisions and advance personalized immunotherapy for head and neck cancer patients.

Dihydroartemisinin Targets the NFIC/FBN1 Cascade to Enhance Wound Healing in Chronic Skin Ulcer by Inhibiting Fibroblast Ferroptosis

Dysfunction of fibroblasts contributes to a pathological state to delay wound repair in chronic skin ulcer (CSU). Dihydroartemisinin (DHA), a derivative of artemisinin, has a therapeutic potential in diverse diseases owing to multiple pharmacological effects. However, no attempt was made to evaluate the function of DHA in CSU. Human dermal fibroblasts were isolated from the peripheral ulcerative tissues in CSU patients (uHFBs) and normal skins (nHFBs). Cell migration, proliferation, apoptosis, and ability were detected. Ferroptosis was evaluated by detecting Fe2+, iron and ROS contents. Immunoblot and quantitative PCR analyses were performed to quantify expression. The NFIC/FBN1 binding relationship was verified by luciferase reporter assay. The CSU mouse model was established, and histology and Masson's staining was used to analyze DHA efficacy. DHA increased NFIC expression in uHFBs. DHA accelerated cell proliferation and migration and impeded ferroptosis in uHFBs, which could be partially counteracted by NFIC reduction. Mechanistically, NFIC transcriptionally elevated FBN1 expression, and DHA increased FBN1 expression by NFIC. NFIC increase enhanced uHFB proliferation and migration and suppressed ferroptosis, which could be abrogated by FBN1 downregulation. Moreover, DHA improved wound repair in CSU mice by upregulating NFIC and FBN1. Additionally, NFIC and FBN1 were underexpressed in uHFBs versus nHFBs. Our findings indicate that DHA has the efficacy to improve wound repair in CSU mice and upgrades skin fibroblast function via the NFIC/FBN1 cascade. DHA may be a novel drug for CSU treatment.

Exploring immune checkpoint inhibitors: Focus on PD-1/PD-L1 axis and beyond

Immunotherapy emerges as a promising approach, marked by recent substantial progress in elucidating how the host immune response impacts tumor development and its sensitivity to various treatments. Immune checkpoint inhibitors have revolutionized cancer therapy by unleashing the power of the immune system to recognize and eradicate tumor cells. Among these, inhibitors targeting the programmed cell death protein 1 (PD-1) and its ligand (PD-L1) have garnered significant attention due to their remarkable clinical efficacy across various malignancies. This review delves into the mechanisms of action, clinical applications, and emerging therapeutic strategies surrounding PD-1/PD-L1 blockade. We explore the intricate interactions between PD-1/PD-L1 and other immune checkpoints, shedding light on combinatorial approaches to enhance treatment outcomes and overcome resistance mechanisms. Furthermore, we discuss the expanding landscape of immune checkpoint inhibitors beyond PD-1/PD-L1, including novel targets such as CTLA-4, LAG-3, TIM-3, and TIGIT. Through a comprehensive analysis of preclinical and clinical studies, we highlight the promise and challenges of immune checkpoint blockade in cancer immunotherapy, paving the way for future advancements in the field.

The Large Language Model ChatGPT-4 Exhibits Excellent Triage Capabilities and Diagnostic Performance for Patients Presenting With Various Causes of Knee Pain

PURPOSE

To provide a proof-of-concept analysis of the appropriateness and performance of ChatGPT-4 to triage, synthesize differential diagnoses, and generate treatment plans concerning common presentations of knee pain.

METHODS

Twenty knee complaints warranting triage and expanded scenarios were input into ChatGPT-4, with memory cleared prior to each new input to mitigate bias. For the 10 triage complaints, ChatGPT-4 was asked to generate a differential diagnosis that was graded for accuracy and suitability in comparison to a differential created by 2 orthopaedic sports medicine physicians. For the 10 clinical scenarios, ChatGPT-4 was prompted to provide treatment guidance for the patient, which was again graded. To test the higher-order capabilities of ChatGPT-4, further inquiry into these specific management recommendations was performed and graded.

RESULTS

All ChatGPT-4 diagnoses were deemed appropriate within the spectrum of potential pathologies on a differential. The top diagnosis on the differential was identical between surgeons and ChatGPT-4 for 70% of scenarios, and the top diagnosis provided by the surgeon appeared as either the first or second diagnosis in 90% of scenarios. Overall, 16 of 30 diagnoses (53.3%) in the differential were identical. When provided with 10 expanded vignettes with a single diagnosis, the accuracy of ChatGPT-4 increased to 100%, with the suitability of management graded as appropriate in 90% of cases. Specific information pertaining to conservative management, surgical approaches, and related treatments was appropriate and accurate in 100% of cases.

CONCLUSIONS

ChatGPT-4 provided clinically reasonable diagnoses to triage patient complaints of knee pain due to various underlying conditions that were generally consistent with differentials provided by sports medicine physicians. Diagnostic performance was enhanced when providing additional information, allowing ChatGPT-4 to reach high predictive accuracy for recommendations concerning management and treatment options. However, ChatGPT-4 may show clinically important error rates for diagnosis depending on prompting strategy and information provided; therefore, further refinements are necessary prior to implementation into clinical workflows.

CLINICAL RELEVANCE

Although ChatGPT-4 is increasingly being used by patients for health information, the potential for ChatGPT-4 to serve as a clinical support tool is unclear. In this study, we found that ChatGPT-4 was frequently able to diagnose and triage knee complaints appropriately as rated by sports medicine surgeons, suggesting that it may eventually be a useful clinical support tool.

Enhanced recovery after surgery (ERAS) protocol with multimodal analgesia incorporating transverse abdominis plane block after elective caesarean delivery: A randomised controlled trial

OBJECTIVE

To evaluate ERAS protocol with multimodal analgesia incorporating transverse abdominis plane(TAP) block compared to routine postoperative care for pain control in women undergoing elective caesarean delivery.

METHODS

This randomized study was conducted after ethical approval and CTRI registration (2020/10/028652). Total 140 women scheduled for elective caesarean were randomised into two groups. ERAS protocol was implemented in Group-1(n=70), and routine practice was followed in Group-2(n=70). Multimodal analgesia included TAP block after skin closure and intravenous paracetamol at shifting to recovery, followed by oral drugs as per analgesic clock. Outcomes were Pain Numerical Rating Scale(NRS) at rest and on movement at 6, 12, 24, 48 hours and at readiness for discharge, and need for rescue analgesics.

RESULTS

NRS was significantly less in group-1 at all time points at rest and movement. Mean±SD NRS at rest was 30.57±1.52 vs. 64.71±1.52 at 6 hours, 28.14±1.39 vs. 56.57±1.39 at 12 hours in group-1 vs. group-2 respectively; (P=0.001). NRS on mobilisation was 37.71±1.57 vs. 74.00±1.57 and 36.42±1.46 vs. 65.28±1.46 at 6 and 12 hours respectively, (P=0.001). Patients needing rescue analgesia were 30 % vs. 87.1 % within 6 hours(65 % reduction), in 24.3 % vs. 77.1 % between 6-12 hours, 4.3 % vs. 40 % between 12-24 hours in group-1 and 2 respectively; with no requirement after 48 hours in either group. The mean Likert score for satisfaction was 4.4±0.60 and 3.4±0.78 in group-1 and group-2, respectively, p<0.001.

CONCLUSION

ERAS protocol incorporating TAP block was effective in reducing postoperative pain, with 65 % reduction in requirement of rescue analgesia within six hours.

Sequential Treatment of Metastatic Colorectal Cancer in Taiwan: Real-World Evidence From Regorafenib and Trifluridine/Tipiracil Use

OBJECTIVE

This study aims to evaluate the real-world effectiveness and safety of sequential treatment with regorafenib and trifluridine/tipiracil (FTD-TPI) in patients with metastatic colorectal cancer (mCRC) in Taiwan.

METHODS

Data were obtained from Taiwan's National Health Insurance Research Database (NHIRD) to assess clinical outcomes in mCRC patients who were treated with both drugs in either sequential order from 2016 to 2019. Overall survival (OS) was analyzed using Kaplan-Meier curves and Cox's proportional hazard models, with adjustments made for age, gender, Quan-CCI score, presence of liver metastases, number of metastatic sites, and the use of anti-epidermal growth factor receptor medications. Additionally, age-stratified subgroups and sensitivity analyses were conducted to examine the robustness of our findings.

RESULTS

Five hundred and twenty-eight patients receiving both study drugs were included. The regorafenib/FTD-TPI group demonstrated a longer median OS of 14.1 months compared with 10.2 months in the FTD-TPI/regorafenib group (p = 0.007). The survival benefit for the regorafenib/FTD-TPI sequence remained significant after adjustment (adjusted HR, 1.49; p = 0.002). The mean treatment duration was also longer for regorafenib/FTD-TPI than FTD-TPI/regorafenib (337 vs. 214 days; p < 0.01). No significant difference between the sequential treatment groups was observed in any adverse event of interest. Both subgroup and sensitivity analyses yielded outcomes consistent with the main analysis.

CONCLUSION

The findings indicated that initiating treatment with regorafenib followed by FTD-TPI had superior clinical outcomes compared with the reverse sequence among mCRC patients. This study offers real-world evidence for clinical decision-making and treatment optimization.

Geriatric Nutritional Risk Index as a Predictor of Delirium and Pressure Injuries in Critically Ill Older Patients With Ischaemic Stroke: An Observational Cohort Study

BACKGROUND

Malnutrition is a prevalent phenomenon among patients with ischaemic stroke, and it is associated with a multitude of adverse health outcomes.

AIM

To evaluate the Geriatric Nutritional Risk Index (GNRI) as a predictor of both delirium and pressure injuries in critically ill older adults with ischaemic stroke.

STUDY DESIGN

An observational cohort study of 969 patients with ischaemic stroke conducted at a tertiary academic medical centre in the United States, divided into two groups based on GNRI scores: at risk of malnutrition (GNRI ≤ 98) and not at risk (GNRI > 98). Delirium was assessed via the Confusion Assessment Method for the Intensive Care Unit and nursing notes; pressure injuries were identified through direct clinical observation using the International Pressure Injury Staging System. Multivariable logistic regression, propensity score matching, and inverse probability of treatment weighting were used for analysis.

RESULTS

Patients at risk for malnutrition had a significantly higher prevalence of delirium and pressure injuries compared with those not at risk (66.4% vs. 46.4% for delirium and 30.3% vs. 9.7% for pressure injuries, both p < 0.001). Multivariable analysis showed that lower GNRI scores were significantly associated with increased risks of both delirium (OR: 1.75, 95% CI: 1.28-2.40, p < 0.001) and pressure injuries (OR: 2.70, 95% CI: 1.79-4.09, p < 0.001). The results remained consistent even after propensity score matching and inverse probability of treatment weighting analyses.

CONCLUSIONS

The study shows that the GNRI is an effective predictor of the risk of pressure injury and delirium in older adults with ischaemic stroke.

RELEVANCE TO CLINICAL PRACTICE

The assessment and management of nutritional status using GNRI in clinical practice has the potential to facilitate the early detection of high-risk patients and the implementation of targeted nutritional interventions.

Managing Preoperative Anxiety and Thromboprophylaxis in Children Undergoing Urological Procedures: An Update of the European Association of Urology/European Society for Paediatric Urology Guidelines on Paediatric Urology

Context

The literature on preventative measures against anxiety and antithrombotic management in children undergoing urological procedures is still limited, resulting in a generally low level of evidence. These guidelines aim to provide a practical approach based on a consensus from the European Association of Urology (EAU)/European Society for Paediatric Urology (ESPU) Paediatric Urology Guidelines Panel.

Objective

The authors aim to provide the 2024 EAU/ESPU Paediatric Urology Guidelines Panel update of the chapter on perioperative management of urological procedures in children.

Evidence acquisition

A structured literature review was performed for all relevant publications published from the last update until April 03, 2023.

Evidence synthesis

The most important updates include the following: anxiety and distress should be prevented or relieved by combining measures such as premedication, distraction techniques, and presence of parents or caregivers. Clinicians should select the appropriate premedication depending on the patient's age, underlying conditions, and psychological status. A particular focus must be placed on paediatric patients with "special needs", including children with psychophysical disorders that impact their relational and cognitive abilities. This unique population requires carefully tailored perioperative management. The incidence of perioperative thromboembolic events in the paediatric population is generally low. Controversies still exist on whether to perform a preoperative coagulation panel test on a routine basis. Neonates and adolescents are at a higher risk of perioperative thromboembolic events than the other children. Standard perioperative antithrombotic prophylaxis is not recommended due to a lack of high-quality evidence-based data.

Conclusions

This paper is a summary of evidence on preventative measures against anxiety and antithrombotic management in children undergoing urological procedures.

Patient summary

In this summary and update of the European Association of Urology/European Society for Paediatric Urology guidelines on paediatric urology, we provide practical considerations for preventative measures against anxiety and antithrombotic management in children undergoing urological procedures.

Search for Mutations Connected With Non-Response to Anti-EGFR Therapy in mCRC in the Morphologically Defined Regions of Primary Tumours

BACKGROUND

Emerging evidence suggests that tumour morphological heterogeneity may influence mutational profiles relevant to therapy response. In this pilot study, we aimed to assess whether mutations identified within specific morphological patterns or at the invasion front correlate with shorter time to progression after anti-EGFR therapy, as compared to whole-tissue analysis.

METHODS

We investigated genetic mutations in 142 samples from primary tumours of 39 KRAS wild-type metastatic colorectal cancer (CRC) patients receiving anti-EGFR therapy. Deep next-generation sequencing was performed on whole-tumour sections and six morphology-defined tumour regions.

RESULTS

Mutations in genes linked to anti-EGFR therapy response (KRAS, BRAF, NRAS, PTEN and PI3KCA) were found uniquely in the non-responder group, with substantial variability across morphological sub-regions. BRAF mutations were aligned with serrated and mucinous morphologies, while KRAS mutations (p.Lys147Glu and p.Ala146Thr) were associated with mucinous and desmoplastic morphologies. In all cases, the cumulative mutational profile from sub-regions provided more details than that of the whole-tumour profile.

CONCLUSION

Our findings highlight that comprehensive analysis, considering morphological heterogeneity, is crucial for personalised CRC treatment strategies.

Obesity, white adipose tissue and cancer

White adipose tissue (WAT) is crucial for whole-body energy homeostasis and plays an important role in metabolic and hormonal regulation. While healthy WAT undergoes controlled expansion and contraction to meet the body's requirements, dysfunctional WAT in conditions like obesity is characterized by excessive tissue expansion, alterations in lipid homeostasis, inflammation, hypoxia, and fibrosis. Obesity is strongly associated with an increased risk of numerous cancers, with obesity-induced WAT dysfunction influencing cancer development through various mechanisms involving both systemic and local interactions between adipose tissue and tumors. Unhealthy obese WAT affects circulating levels of free fatty acids and factors like leptin, adiponectin, and insulin, altering systemic lipid metabolism and inducing inflammation that supports tumor growth. Similar mechanisms are observed locally in an adipose-rich tumor microenvironment (TME), where WAT cells can also trigger extracellular matrix remodeling, thereby enhancing the TME's ability to promote tumor growth. Moreover, tumors reciprocally interact with WAT, creating a bidirectional communication that further enhances tumorigenesis. This review focuses on the complex interplay between obesity, WAT dysfunction, and primary tumor growth, highlighting potential targets for therapeutic intervention.

Development and Comparative Evaluation of a Reinstructed GPT-4o Model Specialized in Periodontology

BACKGROUND

Artificial intelligence (AI) has the potential to enhance healthcare practices, including periodontology, by improving diagnostics, treatment planning and patient care. This study introduces 'PerioGPT', a specialized AI model designed to provide up-to-date periodontal knowledge using GPT-4o and a novel retrieval-augmented generation (RAG) system.

METHODS

PerioGPT was evaluated in two phases. First, its performance was compared against those of five other chatbots using 50 periodontal questions from specialists, followed by a validation with 71 questions from the 2023-2024 'In-Service Examination' of the American Academy of Periodontology (AAP). The second phase focused on assessing PerioGPT's generative capacity, specifically its ability to create complex and accurate periodontal questions.

RESULTS

PerioGPT outperformed other chatbots, achieving a higher accuracy rate (81.16%) and generating more complex and precise questions with a mean complexity score of 3.81 ± 0.965 and an accuracy score of 4.35 ± 0.898. These results demonstrate PerioGPT's potential as a leading tool for creating reliable clinical queries in periodontology.

CONCLUSIONS

This study underscores the transformative potential of AI in periodontology, illustrating that specialized models can offer significant advantages over general language models for both educational and clinical applications. The findings highlight that tailoring AI technologies to specific medical fields may improve performance and relevance.

Textual Proficiency and Visual Deficiency: A Comparative Study of Large Language Models and Radiologists in MRI Artifact Detection and Correction

RATIONALE AND OBJECTIVES

To assess the performance of Large Language Models (LLMs) in detecting and correcting MRI artifacts compared to radiologists using text-based and visual questions.

MATERIALS AND METHODS

This cross-sectional observational study included three phases. Phase 1 involved six LLMs (ChatGPT o1-preview, ChatGPT-4o, ChatGPT-4V, Google Gemini 1.5 Pro, Claude 3.5 Sonnet, Claude 3 Opus) and five radiologists (two residents, two junior radiologists, one senior radiologist) answering 42 text-based questions on MRI artifacts. In Phase 2, the same radiologists and five multimodal LLMs evaluated 100 MRI images, each containing a single artifact. Phase 3 reassessed the identical tasks 1.5 months later to evaluate temporal consistency. Responses were graded using 4-point Likert scales for "Management Score" (text-based) and "Correction Score" (visual). McNemar's test compared response accuracy, and the Wilcoxon test assessed score differences.

RESULTS

LLMs outperformed radiologists in text-based tasks, with ChatGPT o1-preview scoring the highest (3.71±0.60 in Round 1; 3.76±0.84 in Round 2) (p<0.05). In visual tasks, radiologists performed significantly better, with the Senior Radiologist achieving 92% and 94% accuracy in Rounds 1 and 2, respectively (p<0.05). The top-performing LLM (ChatGPT-4o) achieved only 20% and 18% accuracy. Correction Scores mirrored this difference, with radiologists consistently scoring higher than LLMs (p<0.05).

CONCLUSION

LLMs excel in text-based tasks but have notable limitations in visual artifact interpretation, making them unsuitable for independent diagnostics. They are promising as educational tools or adjuncts in "human-in-the-loop" systems, with multimodal AI improvements necessary to bridge these gaps.

Perspectives and Experiences With Large Language Models in Health Care: Survey Study

BACKGROUND

Large language models (LLMs) are transforming how data is used, including within the health care sector. However, frameworks including the Unified Theory of Acceptance and Use of Technology highlight the importance of understanding the factors that influence technology use for successful implementation.

OBJECTIVE

This study aimed to (1) investigate users' uptake, perceptions, and experiences regarding LLMs in health care and (2) contextualize survey responses by demographics and professional profiles.

METHODS

An electronic survey was administered to elicit stakeholder perspectives of LLMs (health care providers and support functions), their experiences with LLMs, and their potential impact on functional roles. Survey domains included: demographics (6 questions), user experiences of LLMs (8 questions), motivations for using LLMs (6 questions), and perceived impact on functional roles (4 questions). The survey was launched electronically, targeting health care providers or support staff, health care students, and academics in health-related fields. Respondents were adults (>18 years) aware of LLMs.

RESULTS

Responses were received from 1083 individuals, of which 845 were analyzable. Of the 845 respondents, 221 had yet to use an LLM. Nonusers were more likely to be health care workers (P<.001), older (P<.001), and female (P<.01). Users primarily adopted LLMs for speed, convenience, and productivity. While 75% (470/624) agreed that the user experience was positive, 46% (294/624) found the generated content unhelpful. Regression analysis showed that the experience with LLMs is more likely to be positive if the user is male (odds ratio [OR] 1.62, CI 1.06-2.48), and increasing age was associated with a reduced likelihood of reporting LLM output as useful (OR 0.98, CI 0.96-0.99). Nonusers compared to LLM users were less likely to report LLMs meeting unmet needs (45%, 99/221 vs 65%, 407/624; OR 0.48, CI 0.35-0.65), and males were more likely to report that LLMs do address unmet needs (OR 1.64, CI 1.18-2.28). Furthermore, nonusers compared to LLM users were less likely to agree that LLMs will improve functional roles (63%, 140/221 vs 75%, 469/624; OR 0.60, CI 0.43-0.85). Free-text opinions highlighted concerns regarding autonomy, outperformance, and reduced demand for care. Respondents also predicted changes to human interactions, including fewer but higher quality interactions and a change in consumer needs as LLMs become more common, which would require provider adaptation.

CONCLUSIONS

Despite the reported benefits of LLMs, nonusers-primarily health care workers, older individuals, and females-appeared more hesitant to adopt these tools. These findings underscore the need for targeted education and support to address adoption barriers and ensure the successful integration of LLMs in health care. Anticipated role changes, evolving human interactions, and the risk of the digital divide further emphasize the need for careful implementation and ongoing evaluation of LLMs in health care to ensure equity and sustainability.