共找到 20 条结果
暂无摘要（点击查看详情）
Timely access to pathology reports has increased the need for clear patient-facing explanations. We evaluated whether large language model (LLM)-generated responses to pathology report questions from patients are comparable in quality to explanations written by pathologists and assessed how LLM configuration influences performance. Sixty-five anonymized real-world patient questions from an online pathology education platform were answered using 5 LLM configurations varying by model architecture, prompting strategy, and retrieval-augmented generation. Responses were evaluated using a structured rubric that assessed accuracy, relevance, clarity, empathy, and safety; they were compared using pairwise arena testing with Bradley-Terry modeling to rank performance. Across rubric domains, LLM responses demonstrated performance comparable to pathologist explanations, with 1 configuration meeting noninferiority criteria. Pairwise arena comparisons indicated that the configuration parameters strongly influenced performance, with both model size and retrieval augmentation associated with improved response preference. The highest-performing configuration combined a larger model with retrieval from a curated pathology knowledge base and was strongly preferred over pathologist-written responses. Carefully configured LLM systems can generate patient-facing explanations of pathology reports comparable in quality to pathologist-written explanations. Prompting strategy, model size, and retrieval integration were associated with performance differences, underscoring the importance of system configuration in developing LLM-based tools to expand access to understandable pathology information.
Background Clinicians frequently face questions that require rapid, evidence-based answers. Artificial intelligence (AI) tools are increasingly used for this purpose, yet their reliability for clinical decision-making remains uncertain. This study compared two generative large language model (LLM) systems (ChatGPT and Gemini) and a retrieval-supported clinical platform (OpenEvidence) to determine which provides the most reliable, clear, and clinically applicable information in obstetrics, gynecology, and urogynecology. Methods A cross-sectional comparative design was used to evaluate ChatGPT (GPT-5), Gemini (Gemini 2.5), and the retrieval-supported platform OpenEvidence. Twenty-four clinical questions across three subspecialties were independently assessed by two blinded specialists using the Expert-Adapted DISCERN (EA-DISCERN) tool, which rates 12 quality domains on a five-point scale. Mean ± SD scores were compared across systems and clinical domains using repeated-measures analysis. Results OpenEvidence achieved the highest mean total score (54.0 ± 2.3), outperforming Gemini (50.3 ± 2.4) and ChatGPT (48.7 ± 2.4) (p < 0.001). OpenEvidence scored significantly higher in evidence-based domains; clinical accuracy, guideline consistency, completeness, transparency, and reliability across all fields. As of this writing, Gemini ranked between the two, showing a modest advantage over ChatGPT in rationale explanation and evidence transparency, while both generative models scored higher in language fluency and readability. Overall, total EA-DISCERN scores ranked OpenEvidence highest, followed by Gemini, then ChatGPT. Inter-rater reliability for the total score was ICC[2,1] (absolute agreement = 0.391). Conclusions OpenEvidence provided more guideline-aligned and transparent responses, whereas ChatGPT and Gemini were generally more fluent and readable. For OB/GYN clinicians, retrieval-supported platforms may be more suitable for point-of-care verification, while generative models should be used more cautiously and with clinician oversight.
Chemical reaction databases typically report the molecular structures of reactant and product compounds, as well as their stoichiometry, but lack information, in particular, on the correspondence of reactant and product atoms. These atom-to-atom maps (AAM), however, are crucial for applications including chemical synthesis planning in organic chemistry and the analysis of isotope labeling experiments in modern metabolomics. AAMs therefore need to be reconstructed computationally. This situation is aggravated, furthermore, by the fact that chemically correct AAMs are, fundamentally, determined by quantum-mechanical phenomena and thus cannot be reliably computed by solving graph-theoretical optimization problems defined by the reactant and product structures. A viable solution for this problem is to shift the focus into first identifying a partial AAM containing the reaction center, i.e., covering the atoms incident with all bonds that change during a reaction. This then leads to the problem of extending the partial map to the full reaction. The AAM of a reaction is faithfully represented by the Imaginary Transition State (ITS) graph, providing a convenient graph-theoretic framework to address the questions of when and how a partial AAM can be extended. We show that an unique extension exists whenever, and only if, these partial AAMs cover the reaction center. Moreover, uniqueness results are generalized to partial AAMs in situations where hydrogen atoms are not represented explicitly. In this case their extension can be computed by solving a constrained graph-isomorphism search between specific subgraphs of ITS graphs. We close by benchmarking different tools for this task.
The psychological impact of stillbirth on parents is profound, increasing the need for respectful care. Despite the existence of international guidelines, there has been no clinical confirmation of their efficacy in improving parental mental health outcomes. This study is a web-based cross-sectional study and part of the OPALE (Observatory on PerinatAL hEalth) project. Participants were selected if they suffered a pregnancy loss after 20 weeks (including termination of pregnancy for medical reasons) in the last 10 years. The survey includes: the CiaoLapo Stillbirth Support (CLASS) checklist, the Perinatal Grief Scale (PGS), the National Stressful Events Survey PTSD Short Scale (NSESSS), and questions on satisfaction with care. 261 participants completed the survey. In a multivariate analysis, higher CLASS scores were correlated with lower PGS and NSESSS scores, suggesting a direct relationship between guideline adherence and better psychological outcomes. Specifically, satisfying over 40 of the 60 checklist items independently predicted greater care satisfaction (OR 2.0, CI 1.1-3.8), higher experiences of respectful care (OR 3.6, CI 1.9-7.0), lower grief (OR 0.08, CI 0.1-0.2), and reduced PTSD symptoms (OR 0.21, CI 0.1-0.5). This is the first study which identifies a correlation between adherence to stillbirth care guidelines and better psychological outcomes, indicating their importance in enhancing parents' mental health.
Bone is among the most apoptotically active tissues in the body. During remodeling, repair, and disease, dying osteoclasts, osteoblast-lineage cells, osteocytes, mesenchymal stem/stromal cells, and injury-associated cells release apoptotic bodies (ABs) that retain parent-cell-derived cargo and surface ligands. These vesicles are increasingly viewed not only as debris for efferocytic clearance but also as source-specific signalling units that shape skeletal cell fate, immune activity, mineralization, and repair. This review integrates current evidence for ABs across skeletal homeostasis and disease. We first define ABs within the broader extracellular vesicle landscape, emphasizing vesicle heterogeneity, isolation and characterization challenges, and terminology boundaries. We then examine AB sources and recipient interfaces, including osteoclasts, osteoblast-lineage cells, osteocytes, mesenchymal stem/stromal cells, platelet-derived ABs in injury repair, macrophages, osteoclast phagocytes, chondrocytes, and bone lining cells. We further discuss how dysregulated AB signalling contributes to osteoporosis, osteoarthritis, and bone metastasis, as well as alveolar bone destruction, aging-related bone loss, osteochondral mineralization, and bone injury repair. We highlight therapeutic implications, including AB-based or AB-inspired strategies, the cathepsin K (CTSK)-responsive self-assembling peptide nanoparticle OsteoSAVE for in vivo generation of osteoclast-derived ABs, and the hypothesis that antiresorptive therapies may reshape osteoclast-derived AB (OC-AB) production. We also identify unresolved translational questions, including AB lifespan, circulation, source attribution, and direct human validation.
Diabetic foot ulcers are a major complication of diabetes and contribute significantly to morbidity and healthcare burden in India. Physiotherapists play a vital role in prevention and management; however, their knowledge, attitude, and clinical practices in the Indian context remain unexplored. The study is aimed at assessing the knowledge, attitudes, and practices (KAP) of physiotherapists practicing across India regarding diabetic foot care through a national online cross-sectional survey. A cross-sectional national online survey was conducted from January 2024 to March 2025 among physiotherapists practicing across India. A total of 249 participants were included using a convenience sampling technique. Data were collected using a self-developed, content-validated questionnaire consisting of 20 questions across knowledge, attitude, and clinical practice domains. Descriptive statistical analysis was performed using Jamovi (2.6.23). A total of 249 respondents from 28 Indian states were included. While 94.8% acknowledged their role in diabetic care, only 41% had received formal training in diabetic foot management. Although most participants identified clinical signs of complications, inconsistencies were observed in interpreting ABI and HbA1c values. Although 98.8% reported providing patient education practices, only 47.4% performed routine foot assessments. Greater clinical experience was associated with better application of knowledge in practice. Indian physiotherapists demonstrate positive attitudes toward diabetic foot care; however, gaps remain in formal training and consistent clinical practice persist. Strengthening structured education and developing standardized clinical guidelines are essential to improve practice.
To identify and compare barriers and enablers influencing physical activity in adults and parent-proxy responses for children with type 1 diabetes in the UK. A cross sectional survey was distributed via BreakthroughT1D (formerly JDRF) networks between December 2022 and January 2023. The survey included closed and open-ended questions on physical activity behaviours and influencing factors. Responses were collected from adults with type 1 diabetes and parent-proxy responses for children with type 1 diabetes. Quantitative data were analysed using descriptive statistics and two-proportion z-tests. Free-text responses were analysed thematically. Barriers and enablers were synthesised using the COM-B model and socioecological framework to identify behavioural determinants and intervention targets. 311 responses were analysed (182 adults; 129 parent proxies for children). Stigma and negative comments were reported by nearly three-quarters of both groups. Adults more frequently reported clinical and motivational barriers: fear of hypoglycaemia (53%) and lack of motivation (39%). Adults reported significantly less healthcare provider support for physical activity (24%) compared with parent-proxy responses for children (57%). Parent-proxy responses for children more frequently reported external barriers, including lack of education among coaches (31%). This is one of the first UK-based studies to comprehensively compare barriers and enablers across age groups. Stigma in physical activity and sports settings emerges as a major, under-recognised barrier requiring urgent attention. Multilevel interventions are required, targeting individual knowledge and motivation alongside systemic changes. These should include mandatory diabetes education for coaches and teachers, policy enforcement to address discrimination and enhanced healthcare provider support for physical activity.
Physiotherapists value evidence-based practice (EBP) but applying it when selecting diagnostic tests remains challenging. Existing evidence mainly addresses general EBP implementation, while factors specifically influencing evidence-based diagnostic decision-making are less clear. To achieve consensus on key barriers and facilitators affecting the selection of diagnostic tests within EBP among physiotherapists and manual therapists in primary care in the Netherlands and Belgium, and to identify broadly relevant targets for education, guideline development, and implementation. A qualitative study using a modified Delphi method was conducted. A preliminary literature search informed a focus group discussion, which was used to refine relevant barriers and facilitators. In round 1, participants answered open-ended questions structured around the three domains of EBP and their integration. In rounds 2 and 3, they rated derived statements to determine consensus. Consensus was defined as ≥70% agreement on a 5-point Likert scale. Eighty-five participants commenced the study, and 42 completed all rounds. Out of 33 identified barriers, consensus was reached on 12, including limited generalisability of research, time constraints, and financial limitations. Commonly cited obstacles such as difficulties accessing or interpreting literature and language issues were considered less relevant. Consensus was reached on 22 facilitators, including improved access to scientific resources, clear reporting of diagnostic accuracy, and enhanced translation of research findings into practice. Consensus on barriers to EBP was limited, but broad agreement existed on practical facilitators, notably improving access to scientific information, education, and tools supporting evidence-based diagnostic decision-making.
Erupting evidence for immune checkpoint inhibitors (ICIs) in treating non-small cell lung cancer (NSCLC) has changed the paradigm for resectable NSCLC. Many questions remain regarding how ICIs can affect the perioperative trajectory of patients beyond successful resection and survival benefits, however. The objective of this narrative review is to discuss the evolving perioperative care of ICI-treated patients with NSCLC. This review focuses on the role of anesthesiologists as perioperative experts within a multidisciplinary team and identifies key areas requiring further investigation. Patients presenting for lung cancer resection after immunotherapy must be thoroughly screened for immune-related adverse reactions (irAEs). The most fatal irAEs perioperatively include myocarditis and pneumonitis, which may require immunosuppression and surgical delay up to 6 weeks. A streamlined multidisciplinary approach to preoperative irAE detection and management is ideal. Most recent trials report no increased surgical complexity following immunochemotherapy versus chemotherapy alone. This interpretation is limited by a lack of standardization in the number of immunotherapy cycles, timing of surgery, and grading tools for surgeon-reported difficulty. Development and validation of surgical complexity predictors may help the anesthesiologist in perioperative planning, especially for analgesia. Key areas warranting further research include the negative interaction between opioids and ICI efficacy and the positive effects of prehabilitation, particularly nutrition, on ICI response. Anesthesiologists should be aware of potential irAEs and predictors of surgical complexity, ideally as part of a multidisciplinary, team-based approach in surgical planning for chemoimmunotherapy-treated patients with NSCLC. Emerging areas of interest include the interaction of immunotherapy with opioids and prehabilitation.
Menin inhibition has emerged as a promising therapeutic strategy for acute myeloid leukemia (AML) driven by KMT2A rearrangements or NPM1 mutations. Novel menin inhibitors, including revumenib, bleximenib, ziftomenib, and enzomenib, are currently under clinical evaluation, and selected updated clinical results were presented at the 2025 American Society of Hematology (ASH) Annual Meeting. This brief review summarizes the key findings and discusses the emerging clinical questions regarding combination strategies, treatment sequencing, and molecularly defined use of menin inhibitors.
Ovarian cancer (OC) remains one of the most lethal gynecological malignancies despite advances in conventional therapeutic approaches. The dismal approximately 30% five-year survival rate for advanced disease highlight the urgent need for innovative treatment strategies. Chimeric antigen receptor T (CAR-T) cell therapy, having revolutionized treatment paradigms in hematological malignancies, faces significant challenges when applied to the complex immunosuppressive tumor microenvironment (TME) of ovarian cancer. This comprehensive review investigates critical research questions regarding the extent to which cellular and molecular components of the ovarian cancer TME inhibit CAR-T cell cytotoxic function, and the signaling patterns associated with reduced CAR-T cell infiltration or persistence in tumor masses. We systematically examine the multifaceted immunosuppressive mechanisms within the ovarian cancer TME and evaluate breakthrough strategies designed to overcome these barriers, including next-generation CAR engineering, combinatorial therapeutic approaches, and innovative TME-modulating technologies. The insights presented here provide a crucial roadmap for translating CAR-T therapy from promising concept to clinical reality in ovarian cancer treatment, potentially transforming outcomes for patients with this recalcitrant malignancy where innovative therapeutic options are urgently needed.
X-linked adrenoleukodystrophy (ALD) is a progressive neurometabolic disorder caused by pathogenic variants in the ABCD1 gene, resulting in the systemic accumulation of very-long-chain fatty acids (VLCFAs). C26:0-lysophosphatidylcholine (LPC(26:0)) is the primary biochemical marker of ALD and the basis for newborn screening programs worldwide. LPC(26:0) arises from the accumulation of VLCFA-CoA species that are incorporated into phosphatidylcholine via the lysophospholipid acyltransferase LPLAT10, followed by phospholipase A2-mediated hydrolysis. Compared to conventional plasma VLCFA analysis, measurement of LPC(26:0) by liquid chromatography or flow injection tandem mass spectrometry in plasma or dried blood spots offers superior sensitivity and specificity, even in female patients, for whom VLCFA analysis yields false-negative results in 15-20% of cases. Beyond diagnosis, accumulating evidence indicates that LPC(26:0) and the broader landscape of VLCFA-containing lipids correlate with disease severity. Higher levels are associated with cerebral ALD, adrenal insufficiency, and severe spinal cord disease, and data from newborn screening cohorts suggest that neonatal LPC(26:0) levels may also stratify risk for early-onset disease manifestations. LPC(26:0) accumulates in CNS lipoproteins and exhibits direct neurotoxic and proinflammatory properties in experimental models. This implicates LPC(26:0) not only as a biomarker, but also as a potential mediator of ALD pathology. Furthermore, LPC(26:0) shows promise as a pharmacodynamic biomarker of biochemical correction across therapeutic strategies, from hematopoietic stem cell transplantation to emerging approaches including ELOVL1 inhibition. However, key questions remain regarding the cell type and tissue origin of circulating LPC(26:0), its precise contribution to neuroinflammation, and the threshold levels that reliably predict clinical outcomes in individual patients.
Coagulopathy after cardiopulmonary bypass substantially contributes to postoperative bleeding, transfusion requirements, morbidity, and mortality. The traditional treatment option is fresh frozen plasma (FFP), but prothrombin complex concentrate (PCC) has become an alternative in the past years. This study aimed to systematically map the current evidence and identify critical knowledge gaps regarding the efficacy, optimal timing, dosing regimens, point-of-care testing guidance, administration order, cost-effectiveness, and safety of 4-factor PCC in adult patients undergoing on-pump cardiac surgery, thereby informing future research priorities. A scoping review was conducted following the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) guidelines. A search of several databases was conducted on 8 April 2025 without restrictions on the language or date of the publications. Study selection, data extraction, and charting were performed independently by two mutually blinded reviewers. A total of 57 primary studies (seven randomised controlled trials, 36 retrospective cohorts, five prospective cohorts, and nine experimental investigations) and 15 reviews with or without meta-analyses were identified. Evidence consistently indicates that PCC reduces postoperative chest tube output and transfusion requirements compared with FFP, although the effect sizes vary. Reported dosing ranged from fixed 500-4000 IU to weight-based 15-30 IU kg-1. Viscoelastic testing guides PCC administration; however, standardised thresholds are lacking. Limited data suggest that PCC may be more cost-effective than FFP, and serious adverse events appear to be rare. No study has directly compared different PCC dosing regimens or prophylactic use in high-risk patients. Substantial evidence gaps remain concerning the optimal dose, timing, viscoelastic targets, sequencing with fibrinogen concentrate, and health economic analyses of PCC use in cardiac surgery. High-quality randomised trials addressing these unanswered questions are warranted to refine the clinical guidelines.
Ionised hypocalcaemia can worsen haemorrhage due to its role in clot formation, vascular tone and cardiac contractility. Current guidelines recommend maintaining normal ionised calcium levels in the bleeding trauma patient. However, guidance does not specify when to replace calcium, and for paediatric trauma, it is unclear when clinicians should measure or decide to replace calcium. This study aimed to survey the practices, protocols and opinions regarding the measurement and replacement of calcium in paediatric major trauma. A cross-sectional survey with single-stage voluntary and snowball sampling. Individual emergency departments (EDs) responded to questions on departmental practices, while individual clinicians provided opinions on the measurement and replacement of calcium based on a clinical vignette with two variations (a haemodynamically stable and a haemodynamically unstable). The survey was administered using Online Surveys V3 (JISC, 2025). Respondents were invited to participate through collaborative research networks. There were responses from 67 individual EDs and 99 individual respondents from 19 different countries. Most EDs had a paediatric major haemorrhage protocol (61/67, 91.0%), but this did not include calcium for 16/67 (23.9%) of the EDs. The timing of calcium replacement was either not included or not specified in most protocols (44/67, 65.7%), with variations in when it should be replaced when specified.Opinions on considering calcium replacement before blood products were significantly different based on whether the patient in the vignette was clinically stable or unstable. Opinions on replacement doses of exogenous calcium were significantly different when the calcium was moderately low (<1.12 mmol/L), but were not significantly different if the calcium was severely low (<1.0 mmol/L). There is considerable variation among EDs regarding ionised calcium measurement and calcium replacement, both before blood product transfusion and in major haemorrhage protocols. Clinicians' opinions on calcium measurement and replacement also vary widely. More research is needed to reach a consensus. III.
Persistent pain is a complex condition that places substantial demands on physical therapists and healthcare systems. Understanding how clinicians experience and navigate these complexities is essential for improving care. To investigate the challenges, priorities, and opportunities from the perspective physical therapists who work with people experiencing persistent pain. We conducted focus groups with physical therapists working in pain management settings. Participants responded to four open-ended discussion questions in written and verbal formats. Data were analyzed using systematic text condensation to identify themes. Sixty-six physical therapists across eight focus groups participated. An overarching theme of "complexity of pain" emerged, reflecting its influence across all aspects of care. Eleven interrelated subthemes were identified: clinical skills and education; funding models; professional networking and referral pathways; health professional awareness of evidence-based pain management; community awareness of evidence-based pain management; access to care; physical therapy advocacy and leadership; patient engagement and expectations; clinical time; clinician and patient emotional wellbeing; and clinical and social complexity. From the perspectives of Australian physical therapists, the complexities of managing persistent pain are deeply interrelated and extend across the healthcare system. These challenges are not attributable to pain complexity alone but reflect systemic issues spanning education, funding, access, and advocacy. Addressing these challenges requires coordinated action at individual, organizational, and societal levels.
The thermodynamics of methane hydrate in porous media remains poorly understood with important questions left unanswered regarding its nucleation when confined at the nanometer scale. In particular, key aspects such as the shape and size of the critical nucleus in confinement challenge existing theories which fail to capture experimental observations. Here, a combination of molecular modeling tools and classical nucleation theory is used to predict the critical nucleus of methane hydrate confined within a slit-shaped nanopore. Owing to capillary/curvature effects and the associated Laplace pressure, the hemispherical nucleus at a single surface or bridge nucleus between two parallel surfaces are found to form at thermodynamic conditions that differ from that for a spherical nucleus in bulk solution. Using such thermodynamic modeling, we predict the temperature dependence of the critical nucleus size in nanoconfinement. Moreover, by estimating the energy barriers along different nucleation paths, we identify a crossover between surface nucleation through a hemispherical nucleus to confined nucleation through a bridge nucleus for methane hydrate in nanopores.
From a historical perspective, long non-coding RNAs (lncRNAs) represent a relatively short story. However, this story has many plot lines, thrills, twists, and turns that altogether form quite a long saga of its own. lncRNAs stay at the forefront of a recent paradigm shift from a protein-only world to the mysterious RNA world, the enormous complexity of which we are only beginning to appreciate. Here, we review the most enigmatic aspect of lncRNAs, their coding ability in the context of whole-cell regulation. What peptides do lncRNAs encode as true translons? What is the mechanism of their translation? What roles do these ncRNA-encoded peptides (ncPEPs) play during differentiation and development and in distinct pathologies? Do these ncPEPs contribute to the already known regulatory roles of lncRNAs? Shouldn't we coin the coding subset of lncRNAs its apt name: cryptic long non-coding RNAs (crpt-lncRNAs) for their cryptic coding capacity? These and other questions concerning these current winners of the spotlight in molecular biology, which await resolution, are discussed so that molecular history can be rewritten once again.
Health professions educators widely agree that inquiry must be fostered in trainees, yet fundamental questions remain about how to cultivate it as a habit. This scoping review examined instructional approaches to teaching inquiry in medicine and nursing through the lens of John Dewey's philosophical framework, which conceptualizes inquiry as comprising both a five-phase method and a psychological attitude toward inquiry methods. Our analysis of 103 articles describing inquiry-focused educational interventions in medical and nursing education shows that, beyond clinical reasoning and evidence-based practice, current approaches predominantly emphasize research participation, focusing on building knowledge and skills in inquiry methods while largely neglecting two critical components of Deweyan inquiry. First, interventions rarely engage learners in the initial phase of recognizing and experiencing uncertainty-the crucial starting point that drives authentic inquiry. Second, despite well-established links between attitudes and behavior, promoting positive attitudes toward inquiry methods was rarely an explicit educational objective. We conclude that teaching inquiry purely as a method, without attending to the formative experiences of uncertainty and attitude development, falls short of cultivating inquiry as a sustained professional habit. Future educational interventions can be strengthened by explicitly designing for awareness of uncertainty, including clear attitudinal objectives, and measuring not only the skills acquired but also the attitude formed toward inquiry methods.
Patients use artificial intelligence-based large language models (AI-LLMs) to research minimally invasive surgery (MIS) for hallux valgus; however, their reliability remains uninvestigated. To compare the quality and readability of ChatGPT and Gemini responses regarding MIS bunion surgery. Ten frequently asked questions reflecting diverse clinical and procedural inquiries were submitted to ChatGPT and Gemini. Quality was assessed via DISCERN and 5-point Likert scales. Readability and actionability were evaluated using the Patient Education Materials Assessment Tool (PEMAT) and Flesch-Kincaid Reading Ease (FKRE). No significant differences existed between models (p > 0.05). Both exceeded sufficiency thresholds for DISCERN (49.7; 49) and Likert (5.0; 4.9). While understandability (85.9%; 83.3%) and FKRE (32.9; 33.5) met requirements, both failed actionability (38.3%; 34%). Both AI models offer reliable, high-quality theoretical information regarding MIS hallux valgus surgery. However, they are insufficient in providing actionable guidance and exceed ideal reading complexity for general patient populations.