Healthcare services and data ecosystems remain fragmented, inequitable and misaligned with patient, clinician and public health needs. Three pathways to patient-centred Artificial Intelligence (AI) have emerged to address this: data collaboration, Whole System Approaches and digital inclusion. Each has advanced the field. However, none delivers the continuous, participatory and outcome-grounded learning that equitable AI demands. The populations whose data drives AI learning must also be the populations who benefit from it. This principle has not been structurally embedded in any existing pathway. This narrative review defines a framework for operationalising Adaptive Machine Learning (AML) as a fourth, integrative pathway to patient-centred AI. AML is defined as continuous learning from real-world individual-level data to generate personalised, population-level insights returning direct benefit to the communities who produce them. A narrative review of PubMed and Google Scholar (2018-2025) was conducted across five thematic domains: AI and machine learning in healthcare; real-world data and evidence; patient-centred care and digital health; data governance and AI regulation; and adaptive clinical trial design. Search terms covered healthcare AI applications and adaptive, continuously learning systems, emphasising prediction-oriented rather than static diagnostic approaches. Eligibility was determined through iterative thematic analysis. Eight systemic barriers to patient-centred AI are identified and the three existing pathways critically examined for limitations in addressing bias, validation and equitable implementation. AML is proposed as the mechanism that closes this gap, connecting systems medicine, personalised medicine and precision medicine through continuous, outcome-based learning, identifying three original conceptual contributions as enablers: Integrative Data Governance (IDG), Adaptive Clinical Trial (ACT) designs and Evidence Sandbox Facilities (ESFs). Progress requires globally standardised Healthcare Data Recycling frameworks, adaptive evaluation methodologies and ESFs enabling multi-stakeholder validation. Future research should address federated AML validation, collective consent mechanisms and alignment of regulatory frameworks with international AI governance standards.

Photocatalytic synthesis, as a green and efficient approach to organic synthesis, has been widely adopted in various organic transformations; however, its application in the synthesis of covalent organic frameworks (COFs) remains in the early stages of development. Moreover, to date, no photocatalytic approach has been employed for the synthesis of amine-linked COFs, which feature a higher density of basic sites and demonstrate enhanced stability. Herein, we report a novel one-pot photoredox-catalyzed reductive amination strategy for the synthesis of amine-linked COFs under ambient conditions for the first time. Based on this strategy, a series of amine-linked COFs were synthesized. Among these materials, DT-Am-COF exhibits excellent photocatalytic activity and demonstrates high efficiency as a catalyst for the light-driven oxidative hydrogenation of arylboronic acids and oxidation of sulfides. This study not only presents an effective method for the synthesis of amine-linked COFs, but also opens new insights into their preparation.

The ability to encode well-defined physical properties and spatiotemporally controlled responses into multifunctional nanocomposites is enabling a new generation of advanced responsive systems. We describe the synthesis of magnetic nanoflowers (NFs) and their full dispersion in PEG-based trimethylolpropane ethoxylate triacrylate (TET) to generate photo-curable nanocomposites. NF-loaded magnetic photocurable TET (Mol. Wt. 912 g mol-1, NF-TET912) is shown to provide full particle dispersion and hence crosslinked polymers with full retention of the AC-field heating response. We demonstrate that these characteristics enable use of NF-TET912 for biomedical applications as a versatile coating on glass, metal and plastics and for multiple 2D and 3D manufacturing techniques including templating, direct light processing (DLP) and direct laser writing using 2 photon polymerization (2PP). Nanocomposite structures, achieved using DLP, have demonstrated good fidelity to design, and provide targeted (5°C) AC-field induced local temperature increases. Complex structures prepared using 2PP, have submicron feature size and resolution. The NF concentration range within which homogeneity and dispersion are possible is shown match the range needed for bio- applications and the materials exhibit no gross cellular toxicity. Finally, the potential of NF-TET912 for AC-field stimulated release of the antibiotic ciprofloxacin is evaluated.

Ternary copper halides, such as Cs3Cu2X5 (X = Cl, Br, or I), are a new class of lead-free nanocrystals with exceptional optical properties, making them promising for photo-down-conversion display applications. However, recent studies on blue-emitting ternary copper halide nanorods (NRs) have reported low photoluminescence quantum yields (PLQY), rendering them inadequate for displays. Using a dithiol ligand-assisted growth method, here we have synthesized Cs3Cu2Br5 NRs with a uniform rod morphology that exhibit an improved PLQY of 63% and a photoluminescence excitation (PLE) anisotropy of 0.201. We used transient absorption spectroscopy to explore the exciton recombination mechanisms and provide insights into the optical performance differences between 1,2-ethanedithiol (EDT) and 1,6-hexanedithiol (HDT)-treated NRs. Density functional theory calculations are conducted to evaluate ligand binding modes, binding energies, and lattice distortion. Our findings indicate that short-chain dithiols (EDT/HDT) enable bidentate ligand-assisted shape control of Cs3Cu2Br5 NRs by displacing long-chain oleate ligands and forming Cu-S- bonds orthogonal to the preferential growth direction. Control experiments with monothiol ligands show that these ligands demonstrate effective surface passivation but do not promote nanorod elongation. These findings offer new perspectives on the mechanism of nanorod formation, therefore highlighting the potential of developing lead-free semiconductor NRs for optoelectronic applications.

Artificial intelligence (AI) and machine learning (ML) are transforming immunohematology, offering significant enhancements in precision, accuracy, and personalized transfusion care. This expert overview summarizes current applications of AI/ML in phenotypic and genotypic red blood cell (RBC) antigen determination, antibody detection, crossmatching, and quality control. Recent studies demonstrate the efficacy of image-based and deep learning methods, with accuracies reaching 99 percent for automated phenotyping and genotyping. AI-assisted models for antibody detection and compatibility assessment have also emerged, providing rapid, standardized decision-making that reduces human error and improves transf usion outcomes. Notably, AI-driven platforms like the Haem-Match Consortium's bloodMatcher potentially reduce alloimmunization risks by as much as 80 percent, highlighting the technology's clinical relevance. AI also bolsters hemovigilance and quality control by analyzing vast clinical data for early detection of adverse events. Challenges remain, however, including algorithmic bias, data privacy, and the need for extensive real-world validation to ensure safe and ethical adoption. Integration with existing hospital information systems and development of robust regulatory frameworks are crucial to fully harness these tools. Future research should prioritize validation in diverse populations and transparent AI development practices. Overall, AI and ML hold promise to revolutionize immunohematology by enhancing accuracy, minimizing adverse transfusion reactions, and supporting patient-centric transfusion care.

Berberine, an alkaloid with broad bioactive properties, exerts its pleiotropic effects by modulating SIRT3, a mitochondrial NAD[+]-dependent deacetylase. This review examines the dual role of SIRT3 as a target of berberine in the management of various disorders while also addressing its associated toxicological profiles. The antihyperglycemic effects of berberine are mediated through SIRT3 inhibition, which promotes hyperacetylation of complex I subunits, degradation of MPC1, an increased AMP/ATP ratio, AMPK activation, and suppression of PEPCK1, collectively resulting in the inhibition of gluconeogenesis. In contrast, the hypolipidemic, cardioprotective, and hepatoprotective effects of berberine are associated with SIRT3 activation, which facilitating adipose tissue remodeling, enhancing β-oxidation in NAFLD, promoting lipophagy in diabetic cardiomyopathy, and mitigates doxorubicin-induced cardiotoxicity. Furthermore, berberine exhibits antitumor duality, exerting contrasting effects in colorectal cancer and hepatoma models, and demonstrates versatile mitochondrial metabolic reprogramming in polycystic ovarian syndrome. Notably, ocular toxicity associated with berberine has been linked to SIRT3 inhibition in photoreceptors cells. Overall, the therapeutic versatility of berberine is fundamentally associated with its context-dependent modulation of SIRT3, positioning it as a promising mitochondrial epigenetics modulator for conditions such as metabolic syndrome, NAFLD, and DCM. However, its therapeutic application must be carefully balanced against organ-specific toxicity risks.

Gynecologic cancer during pregnancy requires treatment decisions that preserve maternal oncologic efficacy while minimizing fetal harm. This focused narrative review critically evaluates systemic therapy for cervical, ovarian, and other selected gynecologic malignancies during pregnancy, with surgery, radiotherapy, obstetric outcomes, and emerging technologies discussed only when they directly inform systemic-treatment decisions. Evidence is interpreted using a transparent four-category framework based on study design, consistency, directness, and applicability to pregnancy. The most consistent evidence supports avoiding cytotoxic chemotherapy during organogenesis and permits selected standard-dose regimens after the first trimester when maternal benefit justifies treatment. Anthracycline-, taxane-, and platinum-based regimens are supported predominantly by international cohorts, registries, and case series rather than randomized trials. Evidence for targeted agents and immune checkpoint inhibitors remains sparse, indirect, and vulnerable to reporting bias; therefore, these treatments should generally be deferred during pregnancy unless an individualized multidisciplinary assessment identifies a compelling maternal indication and no effective alternative. Nanomedicine remains preclinical and is summarized solely as a future research direction. This evidence-calibrated approach emphasizes indication-specific treatment, avoidance of iatrogenic prematurity, and prospective maternal and offspring follow-up.

Large-scale population sequencing requires reference genomes that accurately represent the genetic background of target populations. Although high-quality linear references such as T2T-CHM13 have improved the completeness of human genome representation, a single reference genome remains insufficient to capture genetic diversity across populations. Pangenome graphs provide a richer representation of multi-source genomic variation, but translating pangenome diversity into practical reference genomes remains challenging. First, the direct application of pangenome graphs in routine large-scale sequencing analysis is limited by graph complexity and poor compatibility with existing linear-reference workflows. Second, existing pangenome-driven personalized references mainly target individual samples rather than populations, and current path recommendation strategies may overlook local source consistency during branch-node selection. In this study, we propose PGpath, a pangenome-based method for recommending population-adapted linear reference genomes. PGpath generates a reference from a selected pangenome subpath, preserving pangenome diversity while retaining a linear format compatible with standard downstream pipelines. In this study, we characterize the shared genetic background of samples using population-level k-mer frequency features and reformulate reference recommendation as branch-node prediction in the pangenome graph. To encourage local source consistency, we develop a topology-constrained multi-task neural network incorporating source-consistency relationships between adjacent branch nodes. Experiments on Human Pangenome Reference Consortium (HPRC) population datasets and HG002 variant detection benchmarks showed that PGpath-derived references improved read mapping performance and reduced downstream variant detection errors compared with commonly used references. PGpath provides a practical strategy for translating pangenome diversity into population-adapted linear references for large-scale genomic analysis.

The interpretation of genetic variants remains central to precision medicine, and machine-learning-based variant effect prediction tools have become indispensable for supporting decision-making processes. In this review we trace the progression from hand-crafted features and traditional machine learning classifiers, through supervised ensembles, to deep-learning architectures and foundation models that learn representations directly from biological sequences. We classify methods for genetic variant effect prediction by underlying machine learning model and variant type handled, from missense to non-coding and structural, as well as across feature categories, describing their biological rationale, adoption rate, and known limitations. We further examine three cross-cutting methodological challenges-feature orthogonality, data circularity, and training data bias-that affect reliability and clinical applicability, providing a critical overview of the current landscape and the challenges that lie ahead.

Narrative medicine has emerged as an influential framework in contemporary healthcare, arguing that attending to patients' stories improves the quality of clinical practice and ethical reasoning. Yet the conversation surrounding narrative practice has developed largely within medicine, leaving nursing positioned as a secondary site of application rather than a primary source. This paper challenges that positioning. The paper argues that narrative nursing is not an extension of narrative medicine but a primary ethical practice in its own right, grounded in nursing's unique structural capacity for sustained, embodied presence at the site where illness stories unfold. The argument develops across three interlocking dimensions. The first establishes the conceptual foundations of narrative practice in healthcare, tracing a genealogy that extends beyond the most cited frameworks to include the contributions of Kleinman, Frank, and Benner. This genealogy suggests that nursing's relationship to narrative practice is not derivative but constitutive. The second examines nursing's structural position. Continuity, embodied presence, and relational depth are not features that nursing needs to acquire; they are its defining characteristics, and they are precisely the conditions under which illness stories can be told and heard. The third develops the ethical argument. Hearing a patient's story is an act of epistemic justice, moral witnessing, and relational care. Nursing's structural position makes it the discipline positioned to perform that act, drawing on the traditions of care ethics and the relational foundations already embedded in nursing practice. The paper concludes by arguing that narrative nursing has the potential to make a distinctive contribution to clinical ethics, one grounded in sustained, relational presence rather than in episodic encounter. The implications for nursing education, clinical practice, and professional identity are considered.

Due to their characteristics, sonicators are widely used in the pharmaceutical and food industries for the preparation of stable suspensions and oil-in-water emulsions. Ultrasound-assisted processing enables the production of lipid nanoparticles, which are of considerable interest as drug delivery systems. However, despite their widespread use, the influence of the numerous parameters governing ultrasound-mediated lipid nanoparticle formation remains insufficiently characterized. In this study, a systematic and comprehensive investigation was conducted to evaluate the effects of key experimental variables on the size distribution of lipid nanoparticles produced via microprobe sonication. Liposomes composed of a natural phospholipid were employed as model systems. Instrument-related parameters, including nominal and effective power, treatment time, and probe positioning, were examined alongside sample characteristics such as temperature, lipid concentration, dispersion volume, and lipid composition. In parallel, the extent of phospholipid degradation induced by sonication was assessed in relation to the formation of reactive species during cavitation. The results provide quantitative relationships describing particle size evolution as a function of the most relevant factors and suggest that sonication exhibits a behavior analogous to that of a continuous stirred-tank reactor and high-pressure homogenizers. They also indicate that, under the studied conditions, phospholipids remain chemically unaltered. Overall, the aim of this work is to develop a coherent predictive framework for ultrasound-mediated lipid nanoparticle synthesis and to offer practical guidance for process optimization. The proposed framework is expected to be broadly applicable to related colloidal systems, including emulsions and suspensions, in pharmaceutical and food applications across both academic and industrial settings.