Abstract Diabetic retinopathy (DR), a leading cause of vision loss among working‐age adults, has traditionally been viewed through the lens of hyperglycaemia‐driven metabolic dysregulation. Yet mounting evidence positions inflammation, not as a side effect but as a central pathogenic driver across all stages of DR. This review synthesises the emerging paradigm of the gut‒retina axis: the proposition that gut microbiota dysbiosis actively contributes to DR pathogenesis through inflammatory mechanisms. We examine how diabetes‐induced alterations in microbial composition, characterised by reduced taxonomic diversity, depletion of short‐chain fatty acid (SCFA)‐producing bacteria and expansion of pro‐inflammatory taxa, compromise intestinal barrier integrity and permit translocation of microbial products such as lipopolysaccharide into the systemic circulation. The resulting metabolic endotoxaemia, compounded by a shift in the microbial metabolite landscape (elevated trimethylamine N‐oxide, p‐cresyl sulphate and branched‐chain amino acids; diminished SCFAs and indole‐3‐propionic acid), establishes a state of chronic low‐grade systemic inflammation or metaflammation. These gut‐derived signals converge with local metabolic stressors in the retina to activate microglial Toll‐like receptor 4 (TLR4) and NACHT, LRR, and PYD domains‐containing protein 3 (NLRP3) inflammasome signalling, disrupt the blood‒retinal barrier and amplify pro‐inflammatory cytokine cascades that drive progression from non‐proliferative to proliferative DR. We further evaluate emerging therapeutic strategies that target the gut microbiota, including probiotics, prebiotics, dietary interventions and faecal microbiota transplantation, and discuss the translational challenges that must be overcome before these approaches can complement conventional DR management.
Abstract In precision medicine, the rapid expansion of large‐scale databases and routinely collected clinical data has enabled the development of increasingly accurate predictive models utilizing artificial intelligence. However, predictive accuracy does not guarantee the provision of actionable guidance for clinical interventions. In this review, we examine causal AI, a sophisticated tool that integrates causal inference principles with contemporary machine learning methodologies. This integration signifies a paradigm shift from association‐centric modeling toward decision‐relevant, intervention‐oriented evidence. We commence with an exposition of the methodology for constructing causal models from cohort data, accompanied by a thorough elucidation of the technical concepts underpinning this process. Furthermore, how to leverage causal models to mine the causal relationships is provided, as well as real examples of translational applications of causal AI results in clinical practice. Finally, we outline existing challenges and possible future research directions and applications of causal AI.
Abstract Hydrogel‐based microneedles (HMNs) have emerged as an excellent platform for transdermal drug delivery (TDD). They not only offer minimally invasive, and painless administration but also have exceptional biocompatibility, and allow for sustained and controlled release of therapeutic agents. Recent advances have enabled the design of smart HMNs, which respond to stimuli such as temperature, pH, light, and mechanical stress, significantly enhancing the efficiency of transdermal systems. This review explores the recent advancements in HMNs, focusing on their design principles, biomaterial selection, drug loading strategies, and stimuli‐responsive functionalities. The integration of polymers, nanomaterials, and bioactive molecules is emphasized to overcome the skin penetration barrier with enhancement in permeability, strength, and drug loading. This review also outlines the potential of hydrogel systems in personalized medicine and highlights their recent uses in delivering various therapeutics. Application of HMNs in localized cancer treatment, sustained hormone delivery, smart patches for diabetes and hypoglycemia management, ocular therapies, cardiovascular repair, and enhanced vaccination strategies are also discussed. This shows potential in complex wound healing, offering antibacterial, anti‐inflammatory, and tissue regeneration benefits. The challenges of limited drug diffusion, skin irritation, and scaling‐up clinical translation are analyzed. The text also offers perspectives on future directions for hydrogel‐based transdermal delivery, emphasizing the importance of interdisciplinary approaches that combine material sciences, nanotechnology, and biomedical engineering. This review thus aims to guide ongoing research toward clinically viable and patient‐friendly transdermal HMNs.
Abstract Nanobodies (Nbs), derived from the variable domains of camelid heavy‐chain‐only antibodies, have emerged as a transformative tools in modern biomedicine owning to their nanoscale size, exceptional stability, and unique capacity to recognize cryptic epitopes inaccessible to conventional antibodies. This review provides a comprehensive and up‐to‐date overview of the Nb field, spanning fundamental discovery to advanced translational and clinical applications. We first describe the structural and biochemical features that distinguish Nbs from classical immunoglobulins and underpin their favorable biophysical properties. We then systematically summarize current strategies for Nb discovery and production, including the construction of immune, naive, and synthetic/semi‐synthetic libraries, state‐of‐the‐art display and screening technologies, and diverse expression platforms. A major focus is devoted to Nb applications in diagnostics and therapeutics. In diagnostics, we highlight their growing roles in in vitro biosensing and in vivo molecular imaging. In therapeutics, we critically discuss recent advances in cancer immunotherapies, infectious diseases intervention, and emerging applications targeting neurological disorders. Finally, we review the current landscape of clinical translation and approved Nb‐based medicines, and discuss remaining challenges and future perspectives. Collectively, this review aims to provide a rigorous and integrative resource for understanding the evolution of Nbs from diverse discovery platforms to next‐generation biomedical solutions.
Abstract Recent advances in antibody discovery technologies have profoundly reshaped in vitro diagnostics (IVD). Over the past century, diagnostic antibody development has evolved from serum‐derived polyclonal and hybridoma‐derived monoclonal antibodies to phage display, single‐B‐cell screening, reverse sequencing, and now AI‐assisted design, firmly establishing antibodies as indispensable components of modern IVD systems. Yet as antibody generation grows increasingly scalable, reproducible, and standardized, conventional metrics like affinity and specificity can no longer fully account for assay‐level diagnostic performance, widening the critical gap between binder acquisition and assay‐ready functionality. This review reveals a clear paradigm shift in diagnostic antibody research: from discovery‐centered strategies to assay‐aware, design‐oriented engineering. We organize current progress around four core themes: inherent tradeoffs among major discovery routes; context‐dependent affinity interpretation; key advances in functional antibody engineering; and AI's emerging role linking sequence optimization with structure, developability, and assay translation. A consistent trend emerges: diagnostic antibodies are evolving from passive molecular recognition reagents into engineered, structurally programmable, system‐integrated analytical units. Future IVD advances will thus depend not only on improved molecular recognition, but also on rational design of antibody behavior within complex diagnostic systems.
Abstract Neurological sequelae, ranging from acute brain dysfunction (e.g., delirium) to long‐term cognitive impairment, are increasingly recognized as significant complications in survivors of critical care units, particularly in patients with lung injury caused by severe pulmonary inflammation mechanical ventilation. However, the underlying mechanisms linking lung injury to neurological dysfunction remain poorly understood. Traditional in vitro and animal models face significant limitations in replicating complex, bidirectional interactions between lungs and the brain. This review narratively summarizes the current understanding of how severe pulmonary infections contribute to brain dysfunction, focusing on key pathophysiological pathways such as systemic inflammation, infection‐associated hypoxia, mechanical ventilation‐induced injury, and alterations in the pulmonary microenvironment. We highlight recent advances in the development and application of lung and brain organoids, as well as multiorgan‐on‐a‐chip technologies, which enable modeling of complex inter‐organ crosstalk that traditional models cannot capture. Furthermore, we discuss how these cutting‐edge platforms can simulate clinically relevant processes—such as ventilator‐associated pneumonia, acute respiratory distress syndrome, and infection‐induced cognitive impairment‐offering promising avenues for preclinical mechanistic studies and informing future therapeutic development and patient stratification in critical illness. By critically analyzing the current limitations of organoid and organ‐on‐a‐chip systems, we propose future directions to enhance their mechanistic fidelity, standardization, and translational potential. This comprehensive approach ultimately improve our understanding of lung‒brain interactions targeting multiorgan dysfunction in intensive care setting.
Abstract Lateral flow assays (LFAs) are widely used for point‐of‐care diagnostics, but their detection capability remains constrained by limited molecule immobilization, weak signal output from conventional labels, and the absence of new amplification mechanisms. This review first examines the fundamental principles and current key operational modes of LFA briefly, and then systematically analyzes how two‐dimensional (2D) materials address these limitations through a function‐centric framework. Under the LFA context, we categorize 2D material contributions into several mechanistic roles: as high‐contrast signal reporters exploiting intrinsic optical properties, as signal scaffolds enabling dense biomolecule loading, as catalytic amplifiers achieving significant sensitivity enhancement, and as substrates to convert signals. Beyond existing applications, we identify the key challenges of 2D materials for LFA and underexplored opportunities, including 2D material‐functionalized fibers and membranes for structural optimization, and emerging 2D members whose unique functional properties could enable next‐generation LFA modalities. This work provides a critical synthesis of material‐biosensing integration, offering practical design rules for translating 2D‐material advantages into clinically viable LFA platforms.
Abstract Bacterial resistance to antibiotic therapies, coupled with the propensity of microbes to develop highly recalcitrant biofilms on medical devices, which are primary cause of implant failures. These infections necessitate complex revision surgeries and contribute to morbidity and mortality rates in healthcare. Despite the limitations inherent in various targeted sterilization approaches, nonthermal plasma (NTP) has emerged as a highly potent and versatile technology for the decontamination of medical device surfaces. In our review, initially, we discussed comprehensively the implant surface physicochemical attributes that cause the attachment of bacteria, which advances to biofilms on the medical device surfaces. Key findings discuss various antimicrobial approaches, distinctly categorizing plasma interventions. Plasma surface modification (PSM) is highlighted for altering surface wettability and immobilizing antimicrobial agents. Furthermore, the review evaluates plasma‐activated liquids (PALs) for their role in delivering reactive species to complex geometries. In contrast, direct NTP approaches emphasize the immediate gaseous‐phase interaction of reactive oxygen and nitrogen species (RONS) with the device surface to achieve acute infection control. The review analyses the overall NTPs in vitro studies, mostly using plasma‐treated liquids, and in vivo by infiltration of ROS, generating oxidative stress and DNA leakage from the cell walls, as demonstrated efficiently against a broad range of bacteria. The review concludes with the challenges, such as standardization of plasma type and its process parameters.
Abstract The high heterogeneity of skin cutaneous melanoma (SKCM) remains a major obstacle to effective personalized treatment. Existing classification methods may not fully capture the molecular characteristics dictated by the extensive metabolic reprogramming, limiting treatment options. While existing classifications focus on driver mutations, they often overlook the metabolic plasticity due to changes in physicochemical conditions in the cancer microenvironments that facilitate therapeutic resistance. This study aimed to identify novel, metabolism‐centric molecular subtypes of SKCM samples to explore potential clinical vulnerabilities and investigate possible underlying mechanisms. We identified two distinct subtypes through unsupervised clustering based on multi‐omic data. The two subtypes were found to have distinct prognostic and biological drivers: the pro‐oxidative metabolic subtype (POMS) and the inflammation‐driven subtype (IMS). The POMS tends to have significantly worse survival than the IMS patients. Our integrative analyses support a computational prediction that POMS is associated with elevated oxidative‐stress‐related signatures, increased TMB, reduced immune‐related pathway activity, and hyperactive mTORC1 signaling, suggesting a stress‐adaptive metabolic state that warrants future experimental validation. The IMS relied on inflammatory signaling, such as the JAK‐STAT pathway. In silico predictions suggest that POMS exhibit increased sensitivity to Trametinib and Sorafenib, while IMS is sensitive to PLX4720, Rapamycin, Ruxolitinib, and Cisplatin. In retrospective public data, the classifier showed strong prognostic relevance beyond American Joint Committee on Cancer staging, while prospective validation on independent clinical samples is needed before clinical applications. Our study may inform future investigations into the treatment resistance of the disease.
Abstract Lateral flow assays (LFAs) have become the foundation of point‐of‐care (POC) diagnostics, valued for their user‐friendly format, cost‐effectiveness, and rapid turnaround time. However, despite their widespread success, most notably in pregnancy testing and infectious disease monitoring, conventional colourimetric LFAs often suffer from insufficient sensitivity, typically failing to reach sub‐picomolar detection limits, limiting their utility for the early‐stage detection of low‐abundance biomarkers. In recent years, aptamer‐based LFAs have emerged as a promising alternative to conventional antibody‐based formats owing to their superior stability, reproducibility and programmability. This review provides a critical analysis of recent strategies that have been developed to improve the sensitivity of aptamer‐based LFAs. Enhancement strategies are systematically classified into six distinct domains: (i) flow modulation techniques that optimise reaction kinetics; (ii) sample preconcentration methods; (iii) advanced signal transduction reporters beyond traditional gold nanoparticles; (iv) chemical signal amplification, including nanozyme and enzymatic catalysis; (v) structural strategies for maximising label accumulation; and (vi) the engineering of aptamers as programmable recognition elements. Particular emphasis is placed on amplification strategies that are compatible with or uniquely enabled by aptamer‐based detection systems. Finally, the emerging transition from passive biological selection to active molecular engineering is discussed, outlining the future trajectory of ultra‐sensitive next‐generation LFAs.
Abstract Bone defect repair remains a major clinical challenge in orthopedics. Over 2 million cases caused by trauma, tumors, and other factors occur annually, with large‐scale defects posing a particular bottleneck due to limited self‐healing capacity. The inherent limitations of traditional bone grafting techniques, such as donor site scarcity and immune rejection, have driven the rapid advancement of bone tissue engineering and the development of novel bone repair materials. This review summarizes the pathological mechanisms of bone repair, encompassing three stages: inflammation, regeneration, and remodeling. It elaborates on the regulatory roles of immune cells, stem cells, and cytokines within each stage. Key material categories and advantages are highlighted: bioceramics offer excellent osteoconductivity; polymers provide adaptability; metallic materials meet load‐bearing demands, with degradable metals avoiding second surgery; composites achieve synergistic performance. This review outlines evaluation systems, analyzes clinical challenges including complex microenvironments, material matching, multicell regeneration, and translation barriers. Finally, it highlights frontier directions such as bionic design, intelligent regulation, and immune modulation, offering theoretical references for developing novel bone repair materials.
Abstract Diabetic Foot Ulcers (DFUs) are among the most feared complications of diabetes mellitus (DM). The management of DFUs emphasizes limb salvage, and to achieve this, clinical tools are utilized to identify patients who may require a more aggressive initial approach. Current clinical prediction models fail to account for variability in ulcer characteristics and patient‐specific comorbidities, limiting their precision in individualizing outcome prediction. This review explores the emerging role of molecular biomarkers in personalizing DFU outcome prediction. The pathophysiology of DFUs is examined with an emphasis on disruptions in wound healing specific to DM, focusing on biomarkers involved at different stages of wound healing. This review highlights studies that have shown predictive potential of several biomarkers in a variety of biological samples from patients with DFUs. Despite promising findings, challenges remain in their clinical adoption. Larger studies and the development of accessible, biomarker‐based diagnostics are essential to translate this approach into clinical settings and ultimately reduce the global burden of DFUs through personalized therapy, which would considerably increase the quality of life of people with DM.
