Current artificial intelligence (AI) models for medical imaging predominantly focus on a single imaging modality and a single disease. Attempts to create multimodal and multi-disease models have resulted in inconsistent clinical accuracy. Furthermore, training these models typically requires large, well labelled datasets, which are costly and labour intensive to prepare. We aimed to train and evaluate an AI model that can interpret diverse imaging modalities across specialties while maintaining robust performance within each modality. We developed Multimodal, Multi-Disease Medical Imaging Foundation Model (MerMED-FM), a multi-specialty model trained using self-supervised learning and a memory module. MerMED-FM was pretrained on publicly sourced, unlabelled medical images from 12 specialties and seven imaging modalities: chest x-rays, CT, ultrasound, histopathology, colour fundus photography (CFP), optical coherence tomography (OCT), and dermatoscopy. After pretraining, the model was fine-tuned, validated, and evaluated for the diagnosis of a range of diseases on 26 public datasets and five private datasets comprising radiology, histopathology, and ophthalmology images. MerMED-FM was compared against a general-domain vision foundation model, various specialist single-modality foundation models, and a multispecialty foundation model. Models were fine-tuned using 10%, 30%, 50%, and 100% of data, with primary comparative analyses conducted using a 10% label fraction. The primary outcome was the area under the receiver operating characteristic curve (AUROC), which was summarised by imaging modality. MerMED-FM was trained on around 3·3 million images from 53 publicly available, unlabelled datasets, comprising 713 931 chest x-rays, 292 353 CT slices, 389 885 ultrasound frames, 1 017 712 pathology patches, 333 099 CFP images, 176 719 OCT slices, and 401 059 dermatoscopy images. Strong performance was achieved across all modalities at a label fraction of only 10%, with mean AUROC values of 0·844 for chest x-rays, 0·906 for CT, 0·818 for ultrasound, 0·908 for histopathology, 0·810 for CFP, 0·962 for OCT, and 0·827 for dermatoscopy. MerMED-FM has the potential to be a highly adaptable, versatile, cross-specialty foundation model that enables robust interpretation of medical imaging across diverse medical disciplines. National Medical Research Council, Singapore and the Agency for Science, Technology and Research, Singapore.
Secondary autoimmune and inflammatory diseases (SAIDs) are underrecognized and poorly described after allogeneic hematopoietic stem cell transplantation (allo-HSCT). The standardization of management is hindered by the scarcity of data on epidemiology, pathogenesis, and clinical outcomes. To improve the evidence base, we performed a retrospective multicenter EBMT database study of patients who underwent allo-HSCT between 2005 and 2019 for either hematological malignancy or severe aplastic anemia with available information on SAIDs. We included 129 cases of SAIDs and 14,617 controls. The 5-year incidence of SAIDs was 0.9% (95% confidence interval (CI): 0.7-1.1), and the 10-year incidence was 1.1% (95% CI: 0.9-1.3). The median time from allo-HSCT to the diagnosis of SAIDs was 442 days [interquartile range (IQR): 242-1082]. Overall survival after the onset of SAIDs was 83.4% (95% CI 74.6-89.3) at 2 years and 73.4% (95% CI: 62-82) at 5 years. In multivariate analysis, risk factors significantly associated with SAIDs were bone marrow failure (Hazard Ratio (HR): 3.41 [95% CI, 1.55-7.52], p = 0.002), female donor to male patient (HR 1.77 [95% CI, 1.15-2.73], p = 0.009), and the presence of chronic GvHD (HR 1.61 [95% CI, 1.04-2.51], p = 0.034). SAIDs after allo-HSCT are rare but clinically significant entities requiring further investigations, exploring treatment and prevention strategies, as well as improving diagnostic approaches.
Sialic acid-binding immunoglobulin-like lectin 9 (Siglec-9) functions as a critical hub connecting immune homeostasis and disease pathology. By recognizing "self"-associated sialylated modifications, it plays a complex and central role in maintaining physiological balance and driving disease progression. This review systematically integrates empirical research to comprehensively elucidate the biological basis of Siglec-9 and its mechanisms in various pathological states, including cancer, infection, chronic inflammation, neurodegenerative diseases, and metabolic disorders. Critically, dysregulated Siglec-9 signaling constitutes a shared pathogenic hub across diseases. Within the tumor microenvironment, it mediates immunosuppression and immune escape. During infections, pathogens exploit this pathway through molecular mimicry to evade host defenses, and its expression and genetic variants are significantly associated with susceptibility and severity in chronic diseases. Based on these mechanisms, Siglec-9 has emerged as a highly promising diagnostic biomarker, molecular imaging target, and therapeutic candidate. Intervention strategies, such as blocking antibodies, small-molecule inhibitors, and cell therapies, show significant preclinical promise. This review ultimately establishes the importance of targeting Siglec-9 and the related sialic acid-Siglec axis to develop innovative treatment strategies for various refractory diseases and identifies key directions and challenges for future translational research.
Cancer management remains fragmented across its continuum, from late-stage diagnosis and salvage therapies to non-personalized surveillance. Here, we present Oncoformer, a unified multimodal transformer model trained on the China Oncology Multimodal Prediction and Surveillance Study (COMPASS) cohort (3.67 million individuals, 17.7 million clinical visits) and validated on independent external cohorts, including the UK Biobank. Oncoformer integrates longitudinal electronic health records with chest X-ray imaging to address multiple clinical tasks: pan-cancer diagnosis (area under the receiver operating characteristic curve [AUROC] = 0.956), future cancer prediction up to 1 year before diagnosis (AUROC = 0.869), tumor stage inference (mean AUROC > 0.90), patient-specific treatment-response forecasting, and recurrence-free survival stratification across ten cancer types (all p < 0.01). Staging predictions were independently validated against postoperative pathological endpoints and shown to converge on core cancer genomic pathways. By translating routine clinical data into a dynamic view of cancer evolution, Oncoformer provides a framework for risk-informed cancer prediction and treatment stratification using routine clinical data.
To address the absence of nationally accessible proficiency testing (PT) for biospecimens in Thailand, we developed and evaluated a pilot PT programme consisting of four schemes covering DNA extraction, quantification, purity and integrity assessment. The pilot PT programme comprised two processing schemes: DNA extraction from whole blood (P-DNABLD) and frozen tissue (P-DNAFRT) and two testing schemes: DNA quantification and purity (T-DNAQ) and DNA integrity (T-DNAI). All schemes were conducted in accordance with ISO/IEC 17043:2023 and ISO 13528:2022. PT items were assessed for homogeneity and stability, and participant performance was evaluated based on consensus values using z- or z'-scores. All PT items satisfied the ISO 13528:2022 requirements for homogeneity and stability after accounting for instability-related uncertainty in T-DNAQ. 30 laboratories participated nationwide in at least one scheme (P-DNABLD, n=28; P-DNAFRT, n=17; T-DNAQ, n=29; T-DNAI, n=13). All PT items were received within 3 days of dispatch, with ≥90% arriving within the acceptable temperature range. DNA extraction yield-related measurands demonstrated substantial interlaboratory variability (CV 78.2‒95.3%), whereas purity and integrity measurands were more consistent (CV 1.3‒11.2%). In T-DNAQ, A260/A230 showed the greatest interlaboratory variability (CV 8.2‒16.4%); within-platform variability was low in T-DNAI. Across all schemes, ≥75% of participants achieved at least Satisfactory performance (|z or z'| ≤ 2.0). Direct costs were approximately 12% of those of comparable international PT schemes. This pilot demonstrates the feasibility and accessibility of a nationally coordinated PT programme for biospecimens and supports the establishment of a sustainable PT framework to harmonise biospecimen quality practices.
Functional imaging provides a decision-oriented framework for evaluating osteoarthritis (OA) beyond late structural morphology. OA is a heterogeneous, aging-associated whole-joint disease in which cartilage matrix disruption, synovial inflammation, subchondral bone remodelling, altered mechanics, and pain-circuitry changes may evolve before radiographic end-stage features dominate. Conventional radiography and symptom scores remain indispensable, but they provide limited information about active biological processes and are often poorly aligned with pain, progression risk, or treatment response. In this Perspective, we argue that functional imaging should be selected according to a defined decision context rather than technical novelty alone. Quantitative and compositional magnetic resonance imaging (MRI), dynamic contrast-enhanced MRI, selected positron emission tomography (PET)-based approaches, procedure-linked optical coherence tomography, neurofunctional imaging, mass spectrometry imaging, and isotope-informed spatial methods occupy different positions along the translational pathway. Near-term clinical and trial value is most plausible for risk enrichment, mechanistic phenotyping, longitudinal monitoring, and target-engagement assessment; discovery platforms remain essential for target nomination and endotype refinement. We therefore propose a restrained roadmap that separates candidate signals from qualified biomarkers and prioritises reproducibility, biological specificity, and clinical utility.
Federated Learning (FL) enables collaborative training across institutions without sharing sensitive data, a solution for privacy-preserving AI in medical imaging. However, hospital deployment remains challenging due to strict data protection regulations, heterogeneous infrastructures, and limited network accessibility behind firewalls. We introduce TheODen, an open-source framework for Federated training on histopathology Whole Slide Imaging (WSI). It requires no open client-side ports, enabling training through firewalls via a secure reverse-proxy architecture. We conducted, to our knowledge, the first nationwide FL study for histopathology segmentation of colorectal cancer across three German university hospitals, using breast and colorectal cancer datasets without opening firewall ports. TheODen achieves robust segmentation, with global average dice scores of 0.764 on BCSS and 0.754 on SemiCOL despite data heterogeneity and network constraints. These findings underline TheODen's potential to facilitate secure, large-scale collaborations between medical institutions and to accelerate clinical translation of AI models under real-world infrastructure constraints, providing a privacy-preserving-by-design architecture for future collaborations.
The approximately 40-week gestational period is central to human reproduction, yet the genetic architecture of diverse gestational phenotypes and their links to maternal late-life health remain unclear. In 111 phenotypes from up to 121,579 Chinese pregnancies (median n = 78,535 per phenotype), we identified 4,688 independent genome-wide significant signals, including 1,703 new associations. Gestation-specific effects were observed for 7.8% of variants across 30 phenotypes; 18.7% of signals for 24 longitudinal hematological traits exhibited genotype-by-gestational-timing interactions across five antenatal and postpartum periods. Dynamic genetic effects were enriched in growth-regulatory and hormone-regulatory pathways, reflecting maternal-fetal interactions. Genetic correlation and Mendelian randomization analyses with 80 diseases and medication traits in BioBank Japan females revealed shared genetic overlaps and potential causal links between gestational phenotypes and maternal mid-life and late-life health. These results establish a dynamic genetic atlas of human gestation, providing a framework for precision maternal health.
The development of single-chain variable fragment (scFv) antibodies constitutes a transformative advance in veterinary biologics. However, translation into veterinary practice faces unique constraints that distinguish it from human medicine: stringent cost limitations in food animal production, requirements for long-term stability under variable field conditions, cross-species immunogenicity risks, and a lack of established regulatory frameworks for veterinary antibody products. This review systematically summarizes recent progress in scFv research through three integrated pillars: AI-assisted design tools, efficient expression systems, and precision veterinary applications. We first critically evaluate the paradigm shift enabled by artificial intelligence and computational tools, distinguishing approaches that have been directly validated on scFvs from those established in broader antibody engineering contexts. We then evaluate and compare mainstream expression platforms, including bacterial, yeast, mammalian, plant, and insect cell systems, focusing on their scalability, cost, and suitability for veterinary production. Finally, we map technological advances to concrete veterinary needs, detailing their roles in clinical diagnostics, therapeutics, immunomodulation, and biosecurity. By synthesizing current innovations, identifying translational gaps, and providing an application-oriented framework for platform selection, this review aims to inform the rational development of next-generation scFv-based biologics tailored to animal health requirements and field conditions.
Ovarian cancer remains a highly lethal malignancy, primarily due to the lack of rapid and reliable early screening tools. Cancer antigen 125 (CA-125) is a key serum biomarker, but conventional immunoassays are either too complex for point-of-care use or lack sufficient sensitivity. Lateral flow immunoassays (LFIAs) offer speed and simplicity, but their sensitivity and reliability in complex matrices are often inadequate. Therefore, there is a clear need for a sensitive, robust, dual-readout LFIA platform for accurate CA-125 detection at the point of care. We developed a novel dual-mode LFIA by integrating gold nanoparticles (AuNPs) with fluorescent microspheres (FMs) into a single nanocomposite (AuNP-FMs). This platform leverages surface plasmon resonance to amplify fluorescence while retaining a colorimetric readout. The assay achieves detection limits of 0.5 U/mL (fluorescence) and 1 U/mL (colorimetric) for CA-125 within 15 min, representing 10-fold and 20-fold improvements over single FM-based and AuNP-based LFIAs, respectively. It demonstrates high specificity, excellent reproducibility (RSD <6.6%), and reliable performance in human serum as well as clinical samples, thereby successfully distinguishing cancer patients from healthy individuals. This work presents a sensitive and reliable dual-mode biosensor that overcomes the key limitations of traditional LFIAs. The synergy between the colorimetric and fluorescent signals enhances both the robustness and accuracy of CA-125 detection in clinical settings. The platform is cost-effective, portable, and readily adaptable for point-of-care ovarian cancer screening, demonstrating significant potential for clinical translation.
The ecological pollution and hidden food safety risks induced by non-degradable plastic packaging have spurred the development of sustainable alternatives for food packaging. Chitosan (CS) stands out as a favorable candidate due to its abundant natural sources, intrinsic antibacterial activity, superior biocompatibility and excellent biodegradability. Although scattered investigations on chitosan-based electrospun nanofibers (CENs) have been reported, few systematic reviews have comprehensively compared the spinnability of different electrospinning routes and distinguished the synergistic antibacterial mechanisms between chitosan substrates and diverse functional fillers. This review systematically summarizes mainstream electrospinning technologies for CENs, and thoroughly compares their respective merits, drawbacks and applicable scenarios for food packaging. We further elaborate the performance evolution of nanofibers incorporated with antibacterial additives in mechanical, thermal, barrier, antibacterial and antioxidant aspects, alongside the biosafety assessment of functional fillers and controlled-release behaviors of bioactive components. Moreover, this work deeply explores the preservation mechanisms and practical application performances of chitosan nanofibers on perishable fruits, vegetables, meat and aquatic products, and analyzes the research progress and development potential of such membranes in integrated active and intelligent food packaging. Finally, critical bottlenecks restricting industrialization are discussed, including food-contact safety, substance migration risks and low mass-production capacity. Corresponding forward-looking research directions are proposed to promote the commercial translation of CENs as eco-friendly, and multifunctional packaging materials.
The Illinois Perinatal Quality Collaborative (ILPQC) aims to enhance perinatal care in Illinois. However, up to 50% of women do not attend postpartum care. This study assessed the impact of universal screening for social determinants of health (SDoH) during labor and delivery on postpartum care attendance. A retrospective chart review was conducted for women who delivered at a large academic medical center. The study received exempt status from Loyola University Institutional Review Board. Universal SDoH screening was implemented in March 2022. Data were collected from a control group (six months pre-implementation) and an intervention group (six months post-implementation). The primary outcome was attendance at both scheduled postpartum visits. Comparisons were made using Pearson chi-square for categorical variables and ANOVA for continuous variables. Finally, a multinomial regression model, adjusted for patient characteristics, was used to estimate the effect of universal screening on postpartum appointment attendance. Significance was set at p < .05. The sample included 1,098 women, with 541 in the control group and 557 in the intervention group. The SDoH screening completion rate was 75.0% (n = 415). Basic demographics were similar, with minor differences addressed in regression models. Women in the intervention group were 1.69 times more likely to attend both postpartum visits (RRR = 1.69, 95% CI [1.14, 2.51], p = .009) compared to those in the control group. Implementing SDoH screening at our hospital improved postpartum visit attendance.
Behçet's syndrome (BS) is rare in the West. Using UK-linked primary and secondary care data we examine incidence, prevalence and clinical characteristics of this multisystem syndrome. Clinical Practice Research Datalink (CPRD) Aurum and linked Hospital Episode Statistics (HES) enabled estimates of BS epidemiology using the earliest diagnosis code in CPRD or HES between 2001 and 2021. Median time from the first clinical code of a manifestation of interest to confirmed BS diagnosis was determined. A matched case-control design (CPRD Aurum) enabled ORs for BS for a variety of preselected covariates of interest to be calculated. 1565 incident BS cases arose in the study interval, translating to a final incidence rate of 0.75/100 000 person years at risk in 2021. Prevalence rose from 8.62 to 19.80 per 100 000 person years. In 1628 cases of BS in CPRD Aurum, matched with 11 190 controls, the triad of genital ulceration (adjusted OR (aOR) 155.08), uveitis (aOR 43.00) and oral ulceration (aOR 24.73) was significantly increased among participants with BS. ORs for Crohn's disease (5.69) were also higher in the BS cohort compared with controls. First appearance of the clinical code for genital ulceration was associated with the shortest interval to diagnosis (median 240 days; IQR 63-745). BS prevalence appears to increase over time in the UK, potentially linked to changing population demographics and migration, and to performance characteristics of the 2014 International Criteria for Behçet's Disease classification criteria in low-prevalence settings. Recognised clinical characteristics and inflammatory conditions with phenotypical overlap were associated with BS. Diagnostic delay could be reduced by recognition of risk factors and differentiating clinical manifestations. Case ascertainment relied on diagnostic coding without direct validation against classification criteria, which is an important limitation.
The rigorous conduct of well designed clinical trials is critical to improving outcomes for patients with gynaecological cancers. The Gynecologic Cancer InterGroup (GCIG) has promoted collaboration and harmonisation to guide clinical trials in ovarian and endometrial cancer through a series of consensus meetings with published statements. GCIG convened the inaugural Cervical Cancer Consensus Conference in Clinical Research in Dublin, Ireland, on Oct 14-15, 2024, to develop a consensus on design elements, essential questions, and unmet needs in cervical cancer research. All 33 GCIG member groups and additional representatives from patient advocates, young investigators, investigators from low-income and middle-income countries, and the Cervical Cancer Research Network participated in developing, refining, and adopting 16 statements within four topic groups: diagnosis and staging, early-stage disease, locally advanced disease, and persistent or recurrent disease. Consensus was obtained for all statements. Areas of unmet need were identified and tabulated. The high agreement rates generated through a process of debates, discussions, and consensus represent a positive commitment that incorporates harmonisation, inclusion, and pragmatism in the conduct of academic clinical trials globally.
Futile cycles (FCs), also known as substrate cycles, are a pair of opposing biochemical reactions that continually convert a substrate into a product and back. In doing so, FCs waste ATP without producing a tangible metabolic output (thus termed 'futile'). Because ATP hydrolysis is exothermic, recent studies have extensively focused on the thermogenic function of various FCs, particularly in adipose tissue. However, the function of FCs on other target organs and their primary biological functions remain poorly defined. In this forward-looking minireview/perspective, we discuss a few underexplored functions of FCs that underpin metabolic flexibility and systemic metabolic health. We propose an integrative model in which discrete FCs across metabolic organs act in concert to regulate cellular energetics and organismal metabolic physiology. We postulate that FCs sense and integrate metabolic status, redox balance, and metabolite signaling, with mitochondria serving as the central hub where energetic and signaling cues converge to generate a calibrated cellular response. Given the broad regulatory role of FCs, including in metabolic flexibility, future studies should aim to define the wider functions vis-à-vis metabolic homeostasis in health and disease.
Estrogen deficiency during menopause may impair gut barrier integrity, increasing permeability and leading to low-grade inflammation. This study aims to assess blood I-FABP levels as a biomarker of gut barrier integrity in menopausal women and to examine their association with menopausal symptoms and lifestyle factors. A cross-sectional, case-control study was conducted with a total of 40 women, including 20 menopausal and 20 non-menopausal participants, recruited from Çukurova University Balcali Hospital. Serum I-FABP levels were measured using a standardized sandwich ELISA technique. The statistical analyses included the independent samples t-test, Mann-Whitney U test, chi-squared test, logistic regression, and receiver operating characteristic curve analysis. The average serum I-FABP levels were significantly higher in menopausal women than in non-menopausal controls (1.17 ± 0.28 vs. 0.73 ± 0.30 ng/mL, p < 0.001). The ROC analysis yielded an AUC of 0.869, indicating 85% sensitivity and 80% specificity at the optimal threshold of 1.0 ng/mL. Elevated I-FABP levels were associated with a 22.7-fold increased likelihood of menopause (p < 0.001). Age was an independent predictor of elevated I-FABP levels (OR = 1.72, p = 0.024). In Conclusion, Serum I-FABP levels are significantly increased in menopausal women, suggesting compromised gut barrier integrity likely related to estrogen deficiency. Although the sample size provides sufficient statistical power, its small magnitude remains a notable limitation. More long-term studies are needed to confirm these results and look into the reasons behind them.
Pigment dispersion syndrome and pigmentary glaucoma are important causes of ocular hypertension and glaucomatous optic neuropathy, yet their genetic determinants remain incompletely defined, particularly across diverse ancestries. This study aimed to use a large multi-ancestry cohort from the All of Us Research Program to investigate the genetic basis of pigment dispersion syndrome and pigmentary glaucoma. Case-control study. 572 cases and 37,808 controls with array genotyping and 537 cases and 35,493 controls with whole genome sequencing. Using electronic health record phenotyping in the All of Us Research Program, we performed multi-ancestry genome-wide association analyses using both array-based data and whole genome sequencing-based data comparing patients with pigment dispersion syndrome or pigmentary glaucoma to those without either condition. We also performed Firth penalized regression and Fisher analyses, and we performed principal component analyses to assess effect sizes across genetic ancestries. We applied statistical fine-mapping, examined for cross-trait overlap, and assessed expression quantitative trait locus associations for lead variants. P-values and odds ratios of lead loci from genome-wide association analyses; size of credible sets determined from fine-mapping; allele frequency of lead variants in cases, controls, and the general population; expression quantitative trait locus effect size and p-values linking lead variants to gene expression. We identified four loci reaching genome-wide significance across analyses, including signals near EPHA7 (which mediates cell-cell signaling), within TYR (involved in melanin synthesis and replicated from prior studies), within LINC01138, and near OTX2. Statistical fine-mapping refined three of these loci to single-variant 95% credible sets and narrowed the TYR locus to small credible sets, prioritizing possible causal variants. Effect estimates were broadly consistent across genetic ancestry clusters. Lead variants showed regulatory evidence in expression quantitative trait locus, including reduced EPHA7 expression. These findings implicate both melanogenesis and cell-cell adhesion and signaling pathways in pigment dispersion syndrome and pigmentary glaucoma.
Acute pulmonary exacerbations (APE) in persons with cystic fibrosis (PwCF) are associated with a reduction in long-term pulmonary function. Vancomycin is recommended as first-line treatment of methicillin-resistant Staphylococcus aureus -APE in PwCF. While vancomycin poses a significant risk of nephrotoxicity, the relationship between vancomycin exposure and nephrotoxicity in PwCF has not been characterized. We aimed to establish the relationship between vancomycin exposure and onset of nephrotoxicity in PwCF, and to identify predictors of nephrotoxicity. This was a retrospective study among 52 adults PwCF. Nephrotoxicity events were defined based on the RIFLE criteria, as a serum creatinine increase of 1.5 times from baseline or a ≥ 25% decrease in glomerular filtration rate. A population pharmacokinetic (popPK) model was used to derive exposure measures (i.e., area-under the concentration-time curve (AUC)). A total of 590 vancomycin serum concentrations were available for analysis. A one-compartment model best describes vancomycin pharmacokinetics. Creatinine clearance was included as a significant covariate in the clearance. The popPK model allowed quantification of cumulative AUC (AUCcum) and AUC24 values, and AUC normalized (AUCnorm) and cumulative dose were further calculated. Each variable of interest was used to fit a generalized linear mixed-effects (GLM) model. The GLM model identified AUCcum as a significant predictor of vancomycin nephrotoxicity (p < 0.01). Additionally, patients with more than one dosing occasion were significantly more likely to experience nephrotoxicity (p < 0.01). Based on the final AUCcum model, the odds of experiencing a nephrotoxicity event increased by 32.6% (95% CI: 11.0%-58.4%) for each doubling of AUCcum. Our results demonstrate an association between increased vancomycin exposures and risk of nephrotoxicity. This study provides a foundation for optimizing personalized vancomycin dosing strategies in PwCF to balance efficacy and nephrotoxicity risk.
Endometrial injury-related disorders, including intrauterine adhesions, thin endometrium, and chronic endometritis, are a major cause of female infertility. Conventional therapeutic approaches, primarily hormone therapy and surgical interventions, show limited effectiveness in patients with moderate to severe endometrial damage. In this context, regenerative medicine has emerged as a promising direction to overcome current therapeutic limitations and promote functional reconstruction of the endometrium. A comprehensive literature search was conducted across PubMed, Web of Science, Embase, and Scopus to identify relevant studies on endometrial repair and regenerative medicine. The search covered publications from January 2000 to August 2025. The following keyword combinations and MeSH terms were used: ("endometrial repair" OR "endometrial regeneration" OR "intrauterine adhesion" OR "Asherman syndrome" OR "thin endometrium" OR "chronic endometritis") AND ("mesenchymal stem cells" OR "platelet-rich plasma" OR "exosomes" OR "extracellular vesicles" OR "biomaterials" OR "hydrogel" OR "scaffold" OR "regenerative medicine"). Both basic science investigations and clinical studies were included to provide a comprehensive overview of current developments in the field. To minimize the risk of omission, the reference lists of included articles and relevant reviews were manually screened, and potentially pertinent studies were further evaluated. Inclusion criteria: (1) Original research articles or meta-analyses; (2) Focus on therapeutic interventions for endometrial regeneration; (3) Human clinical studies or mammalian animal models. Exclusion criteria: (1) Non-English articles, letters, conference proceedings, etc.; (2) Purely descriptive studies of endometrial physiology without therapeutic intervention, articles lacking quantitative data or sufficient methodological details. The screening process followed the PRISMA 2020 guidelines (Fig. 1). After removing duplicates, 1936 records were screened by title and abstract, and 605 full-text articles were assessed for eligibility. Finally, 102 studies were included in this qualitative synthesis. This review provides a systematic synthesis of recent advances in cell-based therapies, cell-free approaches, and bioengineering strategies for endometrial repair. Evidence derived from different sources of mesenchymal stem cells, including bone marrow, umbilical cord, and endometrium, is comparatively evaluated alongside platelet-rich plasma and extracellular vesicles, with an emphasis on hierarchical assessment of evidence levels. Several critical issues are further examined. Current stem cell-based interventions are largely characterized by a broad reparative profile, yet precise targeting of key molecular mechanisms remains insufficient. To date, no studies have demonstrated the capacity to directly reverse suppression of signaling pathways such as the GZMA-PARD3 axis or to restore depleted regenerative stromal subpopulations, including IGFBP3⁺ cells. Clinical investigations of platelet-rich plasma exhibit substantial heterogeneity, which appears to stem mainly from the absence of standardized preparation protocols and the frequent reliance on surrogate endpoints, particularly endometrial thickness. However, the predictive value of endometrial thickness for live birth is limited, and its use as a primary endpoint may overestimate clinical benefit. Therefore, emphasis should shift toward patient-centered outcomes such as live birth rate. Concerns also arise regarding the degradation kinetics of biomaterials, as mismatches between material resorption and the cyclical regenerative dynamics of the endometrium may increase the risk of secondary adhesions. In addition, the actual delivery efficiency of microneedle-based systems within the enclosed and humid uterine environment has not yet been fully clarified. Based on this critical appraisal, we propose a translational framework that links mechanistic discovery with precision intervention. The importance of long-term follow-up is emphasized, with live birth rate regarded as a more clinically meaningful primary endpoint for evaluating therapeutic efficacy in endometrial regenerative medicine. Endometrial regenerative medicine is transitioning from empirical repair toward mechanism-informed reconstruction. Current evidence is dominated by short-term (< 12 months), small-sample (n < 50) exploratory studies with high interventional heterogeneity. Clinical translation will require unified preparation standards, individualized treatment strategies, and well-designed multicenter randomized controlled trials with live birth rate as the primary endpoint and minimum 2-year follow-up.
Lipidomes are highly complex and variable, and high-throughput analytical strategies are essential for comprehensive lipidome profiling in large-scale analysis. The use of hydrophilic interaction liquid chromatography (HILIC) coupled with conventional untargeted acquisition strategies remains unpopular due to the challenges in interpreting complex spectral data. To address this gap, we present a detailed lipidomic workflow based on a fast 6.7 min-long HILIC separation method coupled with data-dependent acquisition/sequential window acquisition of all theoretical fragment ions analysis (DDA/SWATH) and improve its applicability with a novel framework for untargeted data processing. In parallel, we also established a targeted high-resolution multiple reaction monitoring (MRMHR) method for comprehensive multi-class lipidomic profiling. To evaluate the applicability of these two approaches in large-scale clinical lipidomics, we used both workflows for the analysis of 240 case-control matched plasma samples from a population-based prospective cohort. After method development and optimization, application of the two complementary workflows to 240 plasma samples generated lipid profiles of over 400 and 500 features for targeted and untargeted analysis, respectively. As expected, the coverage of lipids by the untargeted workflow was notably different from that of the targeted workflow at the lipid species level. In terms of quantitation of lipids detected in both approaches, we observed differences in the quantitation of lipids in individual samples, but statistical comparisons of sample groups remained largely concordant. Furthermore, we demonstrated that the choice of internal standards for normalization can influence downstream statistical outcomes. Two fast HILIC-based workflows were demonstrated to be applicable for large-scale clinical lipidomics and represent reliable high-throughput alternatives to existing longer separation methods for both untargeted and targeted analyses using high-resolution mass spectrometry. Furthermore, the influence of targeted and untargeted approaches, together with the assessment of internal standard selection on downstream statistical outcomes, provides important methodological considerations for large-scale lipidomic studies.