消化内科领域文献阅读汇编

To evaluate the clinical applicability of previously established transcriptomic signatures (molecular subtypes, components and GemPred status) in metastatic pancreatic cancer, we conducted a retrospective pooled analysis of 178 patients from three phase 2 trials (PRODIGE35/37, AFUGEM; 2013-2016) testing first-line regimens (FOLFIRINOX, GemNab, FuNab, FOLFIRI3). RNA sequencing was performed on primary/metastatic tumors across French centers, with blinded assessment of subtypes (immune classical, pure basal-like, stroma-activated), quantitative components, and GemPred status. Primary endpoint: progression-free survival (PFS). Immune classical subtype showed superior median PFS (9.03 months) and OS (11.27 months) versus basal-like and stroma-activated subtypes (PFS: p=0.015; OS: p=0.010). Higher classical component correlated with improved PFS (HR=0.83, p=0.048) and OS (HR=0.71, p=0.001). Inactive stroma predicted better PFS (HR=0.64, p=0.001) and OS (HR=0.71, p=0.015). GemPred-negative patients treated with FOLFIRINOX versus GemNab had higher ORR (46.9% vs. 19.1%, p=0.046), longer PFS (8.2 vs. 2.3 months; HR=2.28, p=0.008), and OS (11.6 vs. 5.0 months; HR=2.04, p=0.021). No differences occurred in GemPred-positive patients. In a formal treatment-by-GemPred interaction analysis (FOLFIRINOX vs. GemNab), the interaction was significant for OS (adjusted p-interaction=0.050) but not for PFS (adjusted p-interaction=0.51). Transcriptomic signatures retain prognostic and predictive utility in metastatic pancreatic cancer, with GemPred representing a hypothesis-generating predictive signal for OS (e.g., a FOLFIRINOX OS benefit in GemPred-negative). Prospective validation is warranted for clinical implementation.

New-onset diabetes (NOD) increases the risk of pancreatic cancer. Previous studies have mainly focused on the presence or absence of diabetes, with limited attention to NOD severity at diagnosis and its clinical course. This 15-year longitudinal nationwide cohort study aimed to analyse the pancreatic cancer risk based on the initial severity and progression pattern of NOD. We included 402 663 individuals with or without NOD from the Korean National Health Insurance Service (2005-2019). Initial NOD severity was defined by baseline antidiabetic treatment intensity: no antidiabetics, oral antidiabetics and insulin. Pancreatic cancer risk increased stepwise with higher initial treatment intensity in NOD compared with individuals without diabetes (adjusted HRs (aHRs) 3.01, 4.32 and 5.60 for no antidiabetics, oral antidiabetics and insulin, respectively). Within the same baseline treatment-intensity category, individuals with rapid escalation within 6 months had a higher risk of pancreatic cancer than those with stable or decreased treatment intensity. In some comparisons between groups with different baseline treatment intensities, the risk of pancreatic cancer was higher with less intensive baseline therapy but subsequent intensification than with more intensive baseline therapy without progression. Notably, drug-naïve individuals who initiated antidiabetic treatment within 6 months had a greater risk even than those who were initially on oral medication without worsening (aHR 6.50 vs 3.67). These patterns were more prominent in pancreatic cancer cases diagnosed within 3 years after NOD. Greater initial severity of NOD and rapid early aggravation were both associated with an increased risk of pancreatic cancer.

Pancreatic cancer continues to be among the most lethal malignancies. The distinct tumor microenvironment (TME) not only facilitates tumor cell proliferation, invasion, and metastasis but also establishes a formidable barrier to the diffusion of chemotherapeutic agents and immunotherapies. Consequently, it is imperative to identify novel compounds that target the tumor microenvironment (TME). This study aims to identify a small molecule as a therapeutic agent for pancreatic cancer through dual mechanisms, specifically DDR1 inhibition and methuosis induction. We characterized DDR1 as a primary target of N-(4-Methyl-3-(4-(5-(4-methylpiperazin-1-yl)pyridin-3-yl)-1H-pyrazol-1-yl)phenyl)-6-(trifluoromethyl)picolinamide (IHMT-140) using ADP-Glo assays, KINOMEscan, Western blotting, and molecular docking. Colony formation, invasion, wound-healing migration, and immunohistochemical staining assays demonstrated that IHMT-140 suppresses epithelial-mesenchymal transition (EMT) by regulating key EMT markers. Cell viability assays, including Cell Titer-Glo and live-cell imaging, transmission electron microscopy (TEM), and immunofluorescence staining, revealed that IHMT-140 triggers methuosis in pancreatic cancer cells. RNA-seq was also employed to confirm the underlying mechanisms. Finally, Cell Titer-Glo, Western blot, and flow cytometry were applied to demonstrate the combined antitumor effects of IHMT-140 and gemcitabine. Gemcitabine uptake and tumor accumulation were quantified by high-resolution mass spectrometry. The in vivo efficacy was tested in a xenograft model, with tumor sections analyzed by H&E and pan-collagen staining. We discovered N-(4-Methyl-3-(4-(5-(4-methylpiperazin-1-yl)pyridin-3-yl)-1H-pyrazol-1-yl)phenyl)-6-(trifluoromethyl)picolinamide (IHMT-140) as a novel dual-function agent that acts as a DDR1 inhibitor and a methuosis inducer. It simultaneously remodels the tumor stromal ECM and activates excessive macropinocytosis. Pharmacological evaluation shows that IHMT-140 inhibits tumor invasion and metastasis, enhances gemcitabine accumulation in tumors, and significantly improves therapeutic outcomes. Our findings present a promising therapeutic strategy for pancreatic cancer, integrating DDR1 kinase inhibition with methuosis induction to augment chemotherapy efficacy. This approach provides a potential novel approach for pancreatic cancer therapy.

Pancreatic adenocarcinoma (PAAD) is a highly lethal malignancy with limited prognostic biomarkers and therapeutic targets. Lactate-driven lactylation has recently emerged as an important regulator of cancer progression, but its role in PAAD remains unclear. In this study, integrative analysis of TCGA and GEO datasets, combined with experimental validation, identified a five-gene lactylation-associated signature (LRP3, TTLL6, TSGA13, PRKCG, and SDK2) that effectively stratified PAAD patients by survival risk. High-risk tumors displayed an immunosuppressive phenotype with reduced immune infiltration, Th2-skewed remodeling, checkpoint activation, and distinct mutational and drug-sensitivity features. Among the signature genes, PRKCG was significantly downregulated in PAAD and associated with advanced disease and worse prognosis. PRKCG overexpression inhibited tumor cell proliferation, migration, invasion, and xenograft growth, while enhancing apoptosis. Mechanistically, lactate-induced lactylation impaired PRKCG-dependent activation of the p53 pathway without altering PRKCG expression, and mutation of predicted lactylation sites partially rescued this effect. These findings define a lactylation-associated prognostic model for PAAD and highlight the lactate-PRKCG-p53 axis as a potential therapeutic vulnerability.

Pancreatic ductal adenocarcinoma (PDAC) presents a substantial challenge due to its resistance to cancer treatments. This limited efficacy is, in part, attributed to the immunosuppressive tumor microenvironment (TME), which impairs effector T (Teff) cell activity. Interleukin-2 (IL-2) is a key cytokine for T cell activation, but its therapeutic use is limited by a short half-life, systemic toxicity, and regulatory T (Treg) activation. To address this limitation, we engineered Bifidobacterium longum, a probiotic obligate anaerobe that selectively colonizes the TME, to continuously secrete Super-mutant IL-2 (SumIL-2), an engineered IL-2 variant that preferentially activates Teff cells over Treg cells, thereby delivering SumIL-2 selectively to the tumor (BifidoSumIL-2). Systemic administration of BifidoSumIL-2 significantly suppressed tumor growth in both subcutaneous tumors and orthotopic PDAC in mice, inducing an improved Teff/Treg ratio. Combining BifidoSumIL-2 with chemotherapy, radiation, and immunotherapy further restrained orthotopic PDAC growth, highlighting its therapeutic potential for difficult-to-treat cancers like PDAC.

Defining immune dysregulation during the asymptomatic prodrome of immune-mediated diseases offers opportunities for early disease detection and interception. In inflammatory bowel disease (IBD), prodromal immune changes remain poorly characterised. To define preclinical immunological alterations by characterising longitudinal serum antibody repertoires using high-throughput phage-display immunoprecipitation sequencing (PhIP-Seq). We applied PhIP-Seq to profile antibody responses in 2000 longitudinal serum samples from 200 individuals who developed Crohn's disease (CD), 200 who developed ulcerative colitis (UC) and 100 matched healthy controls within the US military Proteomic Evaluation and Discovery in an IBD Cohort of Tri-service Subjects cohort, collected up to 10 years before diagnosis. Antibody repertoires were profiled against 357 000 microbial-associated, viral-associated, food-associated and immune-associated peptides. Antibody repertoire variability was increased up to ~4 years prediagnosis in pre-CD and pre-UC individuals. Differential analyses revealed elevated herpesvirus-directed responses (notably Epstein-Barr virus) and anti-flagellin antibodies up to 10 years prediagnosis in CD, particularly in individuals who later developed complicated or ileal disease. In contrast, responses to encapsulated bacteria (eg, Streptococcus pneumoniae, Haemophilus, Neisseria) progressively declined towards diagnosis. Pre-UC was characterised by combined antimicrobial, antiviral and autoantibody signatures, including antibodies against the MAP kinase-activating death domain protein. Large-scale serological profiling of archived prediagnostic samples identified disease-specific immune trajectories years before IBD onset, providing novel insights into disease pathogenesis in its prodromal phase.

In the progression from inflammatory bowel disease to associated cancer, the clonal mutational landscape shifts from selection of mutations in inflammatory genes to selection for cancer-driver mutations. How prevalence and expansion of either type of mutant clones could be impacted by the cellular environments in which they arise and how this affects the neoplastic outcome of colitis remains unknown. Here we combine in vivo lineage tracing, in silico modeling, mutational profiling and spatial transcriptomics in a mouse model of colitis-associated tumorigenesis to capture clone fates associated with chronic inflammation. We identify epithelial- and immune-enriched neighborhoods and propose a model in which establishment of a reparative tissue environment facilitates tumor initiation by promoting the selection and expansion of pro-oncogenic clones, reducing the span of inflammation-resistant neighborhoods containing nononcogenic clones.

Excessive foraging of colonic mucin glycans by gut bacteria is associated with diseases such as inflammatory bowel disease. Although Akkermansia muciniphila is an important mucin degrader, the role of carbohydrate sulfatases that facilitate digestion of these heavily sulfated glycans remains unclear. Combining in vitro digestion assays, proteomics and structural biology, we show that A. muciniphila sulfatases, such as Amuc1755 and Amuc0953, have rare adaptations targeted towards known sulfated mucin structures. They show larger degrees of modularity, including a previously unknown mucin-binding domain. When grown on colonic mucin substrates, glycoproteins of reduced size were important for the growth of A. muciniphila. Further mutational analysis and localization studies revealed that desulfation of N-acetyl-D-glucosamine was periplasmic, while desulfation of D-galactose occurred extracellularly and in the periplasm. These data improve our understanding of contexts for the positive health correlations of A. muciniphila while metabolizing colonic mucin as its sole carbon source.

Background Direct endoscopic necrosectomy (DEN) is commonly performed for walled-off necrosis (WON), yet optimal timing remains uncertain. Step-up strategies defer necrosectomy until clinical deterioration, whereas immediate DEN may provide earlier source control but risks unnecessary intervention. We conducted a systematic review and meta-analysis of randomized controlled trials (RCTs) comparing immediate versus step-up DEN. Methods Major databases were searched through April 26, 2026, for RCTs comparing immediate versus step-up DEN. The primary outcome was clinical success. Secondary outcomes included technical success, necrosectomy sessions, total interventions, length of stay, adverse events, bleeding, organ failure, mortality, recurrence, and readmission. Random-effects meta-analysis was performed using Hartung-Knapp adjustment. Certainty of evidence was assessed using Grading of Recommendations, Assessment, Development, and Evaluation (GRADE). Results Five randomized trials, including 269 patients (134 immediate DEN; 135 step-up strategy), were analyzed. In four trials reporting standalone study-defined clinical success, immediate DEN was associated with higher success than step-up strategy (OR 2.56, 95% CI 1.03-6.37; p=0.046; I²=0%). A post hoc sensitivity analysis including ACCELERATE yielded a similar estimate (OR 2.59, 95% CI 1.33-5.04; p=0.017; I² = 0%). Immediate DEN was associated with shorter length of stay (MD -8.02 days, 95% CI -15.00 to -1.05; p=0.033; I²=42%). Technical success, necrosectomy sessions, total interventions, adverse events, bleeding, organ failure, mortality, and readmission did not differ significantly. Conclusions Immediate DEN was associated with higher clinical success and shorter hospital stay compared with a step-up strategy. These findings support selective consideration of immediate DEN in appropriate patients with WON, while timing remains individualized.

Ferroptosis links cellular metabolism to immune regulation. Beyond its role as an iron-dependent form of regulated cell death, ferroptosis generates signals, including oxidized lipids, iron metabolites, and damage-associated molecular patterns, that influence inflammatory and immune responses. The pancreas is particularly susceptible to ferroptotic stress because of its high metabolic demand and close integration with immune and stromal networks. In pancreatitis, ferroptosis translates metabolic injury into innate immune activation, contributing to sterile inflammation and tissue damage. In pancreatic cancer, ferroptotic vulnerabilities can be exploited therapeutically, yet ferroptosis-associated signals may also support immune suppression, immune evasion, and treatment resistance. These findings suggest that ferroptosis functions as an immunometabolic checkpoint rather than simply a cell death program. Here, we discuss how ferroptosis shapes immune responses in pancreatitis and pancreatic cancer and examine the factors that determine whether it promotes inflammation, antitumor immunity, or immune tolerance. We also review ferroptosis-targeted therapies and the challenges associated with their clinical application.

The impact of computer-aided detection (CADe) systems on lesion detection remains uncertain. We evaluated whether a CADe system improves adenoma detection rate (ADR) in a colorectal cancer (CRC) screening program. This multicenter randomized clinical trial included patients aged 40-79 years who underwent colonoscopy after a positive fecal immunochemical test (FIT) or for surveillance. Patients were randomized to either conventional or CADe-assisted colonoscopy. The primary end point was ADR. Secondary end points were detection rates for serrated lesions (SLDR), polyps (PDR), advanced adenomas (AADR), and advanced serrated lesions (ASLDR). We performed subgroup analysis according to indication, endoscopist ADR, and bowel cleansing. Effect estimates were reported as adjusted risk ratios (aRRs) and 95%CIs. 857 patients were randomized and 827 were included in the analysis. Baseline characteristics were comparable between groups (mean age: 61.7 [SD 6.5]; male 490; positive FIT: 552). We did not find statistically significant differences between CADe-assisted and conventional colonoscopy for ADR (60.8% vs. 57.7%; aRR 1.05, 95%CI 0.94-1.17). No significant differences were observed for SLDR (8.8% vs. 7.2%; aRR 1.04, 0.65-1.64), PDR (69.6% vs. 68.8%; aRR 1.00, 0.91-1.10), AADR (23.6% vs. 24.4%; aRR 0.90, 0.71-1.14), or ASLDR (4.9% vs. 4.8%; aRR 0.93, 0.51-1.68). Subgroup analysis showed no significant differences. In colonoscopies performed in a CRC screening program, CADe-assisted colonoscopy did not increase ADR compared with conventional colonoscopy.

Early detection of colorectal polyps is crucial to reduce the morbidity and mortality associated with colorectal cancer. However, during endoscopy, continuous camera motion and the complex clinical environment often degrade key visual cues (e.g., polyp morphology, texture, and boundaries). This leads to cross-frame view shifts, which are characterized by heterogeneous appearances of the same lesion over time, thereby introducing spurious correlations that hinder reliable assessment. To address this, we propose a causality-inspired representation learning framework with spatiotemporal memory for polyp detection in colonoscopy videos (CIRL-Polyp). Specifically, we develop a novel View-Shift-Aware Causal Intervention Module (VACIM) to remove non-causal influences by enforcing prediction invariance under view shift perturbations. To further constrain non-causal factors, we introduce a dual-branch detection framework that processes the original and intervention frame sequences in parallel and enforces prediction consistency across branches, thereby promoting invariance to non-causal variations. In addition, we propose Causal Temporal Consistency Memory (CTCM) to leverage long sequence-dependency features and stabilize causal representations by constructing memory banks across branches and performing temporal consistency-enhanced cross-attention. Comprehensive experiments on two public video datasets and a private dataset demonstrate that CIRL-Polyp outperforms existing methods, which validates the effectiveness and suggests the clinical potential of the proposed framework from a causal perspective.

The Model for End-Stage Liver Disease (MELD) score is widely used to prioritize patients for liver transplantation and to estimate short-term mortality in end-stage liver disease. Inaccuracies in serum creatinine measurements, particularly interference from bilirubin, both key components of the MELD, may influence clinical decision-making. However, standardized approaches to address such analytical bias are lacking. Here we show that bilirubin-related interference leads to clinically relevant MELD distortions and associated outcomes. We developed a correction model using controlled in vitro matrices and validated it against representative patient samples. The model was applied to a large cohort from registry data and a separate clinical population. After correction, clinically meaningful score shifts occurred in a substantial proportion of patients, with lower corrected scores associated with altered transplantation probability and mortality estimates. These findings highlight the importance of harmonized, interference-resistant creatinine assays to improve fairness and accuracy in liver transplant allocation.

Liver sinusoidal microthrombosis (LST) is recognized as an initiating event in fibrogenesis and portal hypertension in chronic liver diseases. Liver sinusoidal endothelial cell (LSEC)-derived chemoattractants recruit neutrophils and macrophages, promoting LST in congestive hepatopathy (CH). However, the driving molecules from LSECs for LST remain unclear. This study aims to elucidate that LSECs inflammatory via cyclooxygenase-2 (COX-2) upregulation triggers metabolic reprogramming promoting thrombospondin-1 (TSP-1)-mediated LST and portal hypertension. A murine model of LST and portal hypertension was established by partial ligation of inferior vena cava (pIVCL). LST and portal hypertension were suppressed in both LSEC-specific COX-2 knockout mice (Ptgs2ΔLSEC) and celecoxib-treated wild-type mice induced by pIVCL. RNA sequencing of mouse liver tissue and untargeted metabolomics of human hepatic sinusoidal endothelial cells (HHSECs) revealed that COX-2 inhibition was concurrent with downregulation of the AKT/mTOR pathway, reduced lactate, and decreased TSP-1. In vitro, COX-2-derived prostaglandin E2 (PGE2) activated the AKT/mTOR pathway, driving glycolytic reprogramming and lactate production. In turn, the accumulation of lactate induced by COX-2 upregulation enhanced histone H3K9 lactylation, which transcriptionally upregulated Thbs1 (encoding TSP-1), thereby instigating a prothrombotic phenotype in LSECs. Collectively, this study uncovers a novel pathogenic axis in LST formation, where COX-2 drives a prothrombotic switch in LSECs via AKT/mTOR-mediated metabolic reprogramming and lactate-dependent epigenetic upregulation of TSP-1. Targeting LSEC COX-2 may represent a promising therapeutic strategy for mitigating LST and portal hypertension.

Fibromyalgia is a common and debilitating chronic pain syndrome of poorly understood etiology. Here we conduct a multi-ancestry genome-wide association study meta-analysis across 2,563,755 individuals (54,629 cases and 2,509,126 controls) from 11 cohorts, identifying 26 risk loci for fibromyalgia. The strongest association was with a coding variant in HTT, the causal gene for Huntington's disease. Gene prioritization implicated the HTT regulator GPR52, as well as diverse genes with neural roles, including DCC, DRD2/NCAM1, MDGA2 and CELF4. Fibromyalgia heritability was exclusively enriched within brain tissues and neural cell types. Fibromyalgia showed strong, positive genetic correlation with a wide range of chronic pain, psychiatric and somatic disorders, including genetic correlations above 0.7 with low back pain, post-traumatic stress disorder and irritable bowel syndrome. Despite large sex differences in fibromyalgia prevalence, the genetic architecture of fibromyalgia was nearly identical between males and females. This study provides robust genetic evidence defining fibromyalgia as a central nervous system disorder, thereby establishing a biological framework for its complex pathophysiology and extensive clinical comorbidities.

Neoplastic recurrence risk at scheduled surveillance intervals after colorectal endoscopic submucosal dissection (ESD), as well as its association with histologic risk features, is unclear, resulting in uncertainty for post-ESD surveillance recommendations. We conducted a systematic review and meta-analysis of 1-, 3-, and 5-year post-ESD neoplastic recurrence with subgroup analyses. A systematic search was performed up until October 2025. Eligible studies were assessed for neoplastic recurrence at or near the resection site at 1-, 3-, and 5-year scheduled surveillance colonoscopies. Rates of metachronous neoplastic lesions at the same time-intervals were also collected. Data were pooled using a random-effects model, and subgroup analyses were conducted for histology of the index lesion (low or high grade dysplasia, and T1 cancer) and resection quality (R0 vs. non-R0). 10 studies encompassing 5306 lesions met the inclusion criteria. The pooled R0 resection rate was 82% (95%CI 79%-85%). Overall neoplastic recurrence rates were low and stable over time: 1.2% (95%CI 0.4%-2.6%), 1.4% (95%CI 1.0%-1.9%), and 1.9% (95%CI 0.9%-3.1%) at 1, 3, and 5 years, respectively. Malignant recurrence was rare (0.2%) and confined to noncurative resections or deep submucosal invasion. The pooled rate of metachronous lesions was 1.4% (95%CI 0.4%-3.0%). Limited time-stratified data suggest progressive accrual of metachronous neoplasia during follow-up rather than early post-ESD failure. Neoplastic recurrence is rare up to the 5-year endoscopic surveillance, especially in those with R0 resection and favorable histology. In low risk patients, an extended surveillance interval of up to 5 years can be considered, mostly because of an increased risk of metachronous lesions.

Hypnotherapy is an evidence-based therapy recommended for IBS, but access to therapist-led hypnotherapy is limited by costs and availability. We studied the effectiveness of smartphone-based self-guided hypnotherapy versus in-person therapist-delivered hypnotherapy. This multicentre, three-armed non-inferiority randomised controlled trial (RCT) included patients with IBS (Rome IV, 16-75 years). Patients were randomised to 12-week in-person therapist-delivered hypnotherapy, smartphone-based self-guided hypnotherapy or online self-guided psychoeducation. The primary endpoint was abdominal pain response per Food and Drug Administration (FDA) definition (≥30% reduction from baseline in weekly average of worst daily abdominal pain in at least 2 out of 4 weeks follow-up). The non-inferiority margin was 10%. Secondary endpoints included ≥50 points IBS-symptom severity scale reduction and changes in psychological symptoms. A total of 230 patients (mean±SD age 38.2±13.8 years, 70.4% female) were randomised. FDA abdominal pain response rates were 48% for in-person therapist-delivered hypnotherapy, 33% for smartphone-based self-guided hypnotherapy and 22% for psychoeducation. Non-inferiority of online to in-person hypnotherapy could not be shown (-14.7%, 95% CI -29.3% to 0.9%). In-person hypnotherapy was more effective than psychoeducation (OR 3.48, 95% CI 1.68 to 7.20, p<0.001) at week 16, both hypnotherapy groups were more effective at 6-month follow-up (OR 3.11, 95% CI 1.37 to 7.09, p=0.007 and OR 4.68, 95% CI 2.07 to 10.57, p<0.001). There were no significant treatment effects on psychological symptoms between groups. This multicentre RCT could not demonstrate non-inferiority of smartphone-based self-guided hypnotherapy compared with in-person therapist-delivered hypnotherapy in reducing abdominal pain. However, substantial response rates for smartphone-based self-guided hypnotherapy highlight its potential as a broadly applicable and cost-effective alternative. NCT03899779.

Surgical resection is a cornerstone of curative therapy for hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (iCCA), the two most common primary liver malignancies. In most patients with HCC, underlying cirrhosis is present, whereas in iCCA an identifiable risk factor is often absent. Regardless of tumour type, surgical planning must account for the future liver remnant to minimise the risk of post-hepatectomy liver failure, evaluating not only its volume but also its functional reserve. Furthermore, a deeper knowledge of surgical oncological principles combined with integrative strategies, including systemic and locoregional therapies, has enabled downstaging and conversion of initially unresectable liver malignancies, expanding surgical eligibility across a growing spectrum of hepatic tumoural conditions. Additionally, advances in the understanding of molecular biology and new treatment strategies, particularly in iCCA, have the potential to further extend the boundaries of resectability.

The population of Vietnam remains underrepresented in global genomic databases. Here, we present VN1K, a resource of multi-omics and phenotypic information for 1011 unrelated Vietnamese individuals. We present high-depth short-read whole-genome sequencing data for all samples along with various -omics datasets. Using a high-sensitivity variant detection pipeline, which includes a pangenome graph reference and a deep-learning framework, we identify approximately 42 million variants with 7 million short insertions/deletions and 90 thousand structural variants. VN1K also features a whole-genome methylation profile based on long read sequencing. We create a genotype imputation panel with high accuracy on the Vietnamese population, allowing us to identify variants with significantly different allele frequencies in the Vietnamese population compared to other populations. We establish the functional relevance of some of these variants, particularly those in genes associated with genetic disorders, immune diseases, and drug responses, by integrating the allele frequency differences with known genotype-phenotype associations and clinical annotations. Further, we map various loci related to hepatitis B virus infection, triglyceride levels, LDL-C levels, serum glucose levels, HbA1c levels, and levels of two liver enzymes (ALT and AST). The VN1K dataset is accessible via genome.vinbigdata.org, an integrated platform with both linear and graph-based genome browsers.

Optimizing murine models for studying Metabolic Dysfunction-Associated Steatohepatitis (MASH) and fibrosis is crucial for understanding disease mechanisms and evaluating therapies. In this study, we characterized diet-induced male murine models of MASH and liver fibrosis using an AI-digital pathology (DP) pipeline for zone-specific analysis and spatial (co)-localization of key MASH features within the liver microarchitecture, providing insights beyond standard histology. Model characterization included an integrative omics-based approach, standard blood-chemistry, and traditional histology. AI-DP revealed previously unknown temporal events leading to MASH, emphasizing the interplay between inflammation and dysmetabolism in disease progression. We also noted distinct morphometric characteristics of granulomas and their correlation with fibrosis. In efficacy studies of clinically-validated treatments, Semaglutide (GLP-1RA), Resmetirom (THRβ-agonist), and MK-4074 (Acetyl-CoA-carboxylase inhibitor), AI-DP demonstrated differential effects on macrosteatosis, microsteatosis, and colocalized fibrosis. Overall, integrating AI enables identification of fit-for-purpose disease models for therapeutic testing, and facilitates robust preclinical study designs for advancing effective therapeutic strategies.

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.

Colorectal cancer (CRC) exhibits significant heterogeneity in response to immunotherapy that cannot be fully explained by microsatellite status alone. Although elevated bile acid levels are recognized as an important risk factor for CRC, their impact on immunotherapy responsiveness remains poorly understood. Here, we demonstrate that high bile acid levels profoundly impair anti-PD-1 efficacy in both CRC patient cohort and mouse models, accompanied by reduced infiltration and functional impairment of tumor-infiltrating CD8+ T cells. Bile acid profiling identified deoxycholic acid (DCA) as the key bile acid species mediating this immunosuppressive effect. In vitro and in vivo studies have shown that DCA not only suppressed CD8+ T cell effector function but also drove them toward terminal exhaustion, thereby limiting responsiveness to anti-PD-1. Mechanistically, DCA disrupted mitochondrial fitness in CD8+ T cells by suppressing oxidative phosphorylation and inducing excessive mitochondrial reactive oxygen species (mtROS) production. In parallel, DCA enhanced ubiquitination-dependent degradation of Parkin, thereby inhibiting mitophagy and causing the accumulation of damaged mitochondria. These convergent defects in mitochondrial homeostasis ultimately promoted CD8+ T cell dysfunction and terminal exhaustion. Notably, pharmacological reactivation of mitophagy via Urolithin A reversed these defects and restored the antitumor efficacy of anti-PD-1 in vivo. Collectively, our findings identified a DCA-Parkin-mitophagy axis that drives CD8+ T cell terminal exhaustion and compromises immunotherapy efficacy, providing a potential metabolic intervention strategy to improve immunotherapy responses in CRC patients with elevated bile acid levels.

Artificial intelligence (AI), particularly foundation and generative models, is reshaping the practice of hepatology through enhanced knowledge synthesis, quantitative and reproducible analysis of multimodal data, and personalized clinical decision support. This narrative review examines the transition from task-specific discrimination AI to large language models (LLMs), multimodal foundation models, and agentic AI. We synthesize evidence from original and validation studies, clinical evaluations, and benchmark studies, as well as expert reviews and regulatory frameworks across metabolic dysfunction-associated steatotic liver disease, chronic hepatitis B, cirrhosis and portal hypertension, hepatocellular carcinoma, and liver transplantation. LLMs can convert free-text notes into structured data, summarize longitudinal electronic health records, support patient education, and retrieve guideline-based information. Retrieval-augmented generation and agentic AI may improve traceability and workflow support, but current evidence is largely retrospective or proof-of-concept. In digital pathology and imaging, discriminative AI has enabled more quantitative and reproducible histologic scoring and biomarker analysis. Pathology and multimodal foundation models offer transferable representations, report generation, and cross-modal reasoning, but hepatology-specific validation remains limited. Key risks include hallucination, automation bias, domain shift across centers and devices, and inequities due to under-representation of patient subgroups. We outline the future directions for safe AI model deployment based on multimodal foundation models, prospective and federated evaluation, lifecycle governance, and continuous monitoring for performance, calibration, and equity. Most generative AI applications in hepatology remain at the proof-of-concept stage, and rigorous prospective validation with human-in-the-loop oversight is required before clinical integration.

Individuals with asymptomatic SARS-CoV-2 infection can unknowingly transmit the virus, yet identifying such subclinical infections in post-vaccinated populations remains challenging. We conducted a longitudinal study of 129 infection-naïve vaccine recipients immunized with various combinations of SARS-CoV-2 spike (S) protein vaccine platforms. Sera were collected before the first dose (v1), at 2 weeks (v7) and 6 months (v8) after the third dose. Taiwan's first major COVID-19 outbreak occurred between v7 and v8. We measured anti-nucleocapsid (anti-N) and anti-S IgG antibody titers by ELISA and assessed virus-neutralizing activity using live virus and pseudovirus assays. By developing an iterative serial screening method, we identified asymptomatic breakthrough (post-vaccination) infections among unconfirmed cases. Our v7-v8 paired cohort resolved into three distinct groups: confirmed cases (21%), asymptomatic breakthrough infections (17%), and uninfected subjects (62%). In normalized v8 sera, confirmed cases exhibited an anti-S+++ (high) /anti-N+++ (high) phenotype, while uninfected subjects showed an anti-S+ (low)/anti-N+(baseline) phenotype. Statistical analysis validated a distinct asymptomatic group characterized by an antibody profile anti-S++ (intermediate) /anti-N+ (baseline). This approach may enable more accurate estimates of vaccine efficacy and infection prevalence. In a spike-vaccinated population, anti-N antibody is more a potential specific marker for COVID-19 symptomatic disease than an ideal marker for SARS-CoV-2 infection. To our knowledge, this is the first preliminary report of identification of asymptomatic breakthrough infection from a well-vaccinated population using self-matched longitudinal pairs of serum samples.

Children with cancer develop many short- and long-term side-effects of treatment1, but the amount of DNA damage associated with chemotherapy exposure is unclear2. Here we used mutational signatures to measure this damage using whole-genome-sequenced tumours from a multi-institutional cohort for which therapy dose and total exposure were uniformly collected3-5. Chemotherapy and radiotherapy were the only exogenous mutagens in relapsed childhood tumours and were often the dominant source of DNA alteration. Compared with treatment-naive tumours, post-therapy cancers carried nearly three times the number of private signatures, and two times the total burden of somatic mutations. Further, the mutagenic effects of different chemotherapies varied. Platinum-based therapies, for which we more than doubled the number of associated signatures, led to the highest number of variants in most patients. Using therapy exposure dates to track when therapy-associated mutations become detectable, we defined a minimum threshold for platinum-associated mutations to emerge. Remarkably, more than one-third of tumours treated with platinum drugs displayed detectable platinum signatures within one year. This work provides genomic evidence for the critical mutagenic effects of chemotherapy in childhood cancer, as a specific driver of tumour evolution. These data highlight opportunities for treatment de-escalation and the future possibility of tracking resistant clones before expansion.

AR pathway-independent prostate cancer (ARIPC), particularly neuroendocrine prostate cancer (NEPC), represents one of the most lethal states of metastatic castration-resistant prostate cancer. However, how fatty acid synthesis (FAS) is organized in ARIPC and whether distinct lipogenic states shape neuroendocrine lineage transdifferentiation remain unclear. By integrating single-cell and bulk transcriptomic analyses of mCRPC cohorts, we identify NEPC as a fatty-acid-synthesis-low state associated with poor survival. Within this context, fatty acid synthase (FASN) emerges as a key indicator and functional contributor to lipogenic activity. FASN depletion suppresses lipogenesis while increasing NEPC-associated programs, migration, and metastatic colonization. We further identify FGFRL1 as the FGF family member most consistently associated with fatty acid synthesis activity in ARIPC. FGFRL1 depletion reduces FASN expression and relative free fatty-acid content, while targeted GC-MS supports broader fatty-acid remodeling and fluorescent uptake assays show increased exogenous fatty-acid uptake. FASN restoration partially restores relative free fatty-acid content and attenuates NEPC-associated and migratory phenotypes. Directional perturbation, rescue, AKT phosphorylation, and co-immunoprecipitation analyses further support the functional FGFRL1-FASN relationship. ONECUT2 is prioritized as a candidate downstream transcriptional regulator whose expression correlates with the neuroendocrine program. Together, these findings support an FGFRL1-FASN metabolic axis that regulates neuroendocrine lineage transdifferentiation and metastatic progression in ARIPC.

Effective hemorrhage control remains an important unmet need in trauma care and emergency medicine. Therefore, it is essential to achieve rapid and effective hemostasis. Bioadhesives offer an attractive approach to wound sealing, achieving simultaneous leakage prevention, mechanical support, and hemostasis. However, available bioadhesives are limited in clinical applications by weak mechanical strength, insufficient biocompatibility, and strict storage condition. In this study, a thrombin-independent fibrinogen-based hydrogel was developed, featuring rapid gelation, robust mechanical strength, strong tissue adhesion, and excellent shelf stability. This hydrogel is prepared from knob-peptide-grafted gelatin methacryloyl (GMK) and fibrinogen methacryloyl (FM). GMK/FM hydrogel can quickly form the knob-hole crosslinking and then be reinforced through photo-crosslinking. Compared with clinical standard-of-care fibrin glue, GMK/FM hydrogel demonstrates comparably rapid gelation (< 2 s), alongside enhanced mechanical strength, tissue adhesion and storage stability (4 weeks at room temperature). In rat severe hemorrhage models of liver and abdominal aorta, GMK/FM hydrogel decreased hemostatic time and blood loss remarkably. It demonstrates exceptional hemostatic performance, outstanding biocompatibility, and storage stability, holding significant promise for clinical applications and emergency uses.

This study addresses the challenges of Error Coupling and model homogenization that commonly arise in dual-model collaborative learning for semi-supervised medical image segmentation by proposing a Divergent-Convergent Framework (DCF). The core innovation of this framework lies in abandoning the traditional blind pursuit of strong prediction consistency and instead dynamically quantifying the confidence and disagreement between the two models through a Guidance Mask (GM). In regions of high confidence and low disagreement, a Convergence Stabilization Mechanism is applied to reinforce the learning of robust pseudo-labels; in regions of low confidence or high disagreement, a Divergent Exploration Mechanism is activated, guiding the models to perform differentiated exploration along two dimensions: internal semantic confusion and external feature orthogonality. This effectively maintains model diversity while suppressing the propagation of shared errors. Systematic experiments on six publicly available datasets - four 2D (ACDC, PROMISE12, Hippocampus, ATLAS) and two 3D (BraTS2019, Pancreas-CT) - demonstrate that DCF significantly outperforms state-of-the-art semi-supervised methods across multiple labeled ratios, including 5%, 10%, and 20%. Qualitative analyses and cross-domain evaluations further validate the method's advantages in segmenting regions with ambiguous boundaries and its robustness to domain shifts. Moreover, DCF can be directly applied to fine-tune segmentation foundation models such as MedSAM and MedSAM2, consistently improving their performance under the same labeled budget, thereby demonstrating its practical value as a label-efficient strategy for clinical deployment.

Intravital microscopy (IVM) using dorsal skinfold chambers (DSCs) enables real-time, high-resolution imaging of the tumor microenvironment. Conventional, metal-based DSC systems often cause animal distress and require technically demanding surgical implantation. To address these challenges and adhere to the 3Rs principles of animal welfare, this study presents a novel, stereolithography (SLA)-based, 3D-printed DSC featuring a lightweight (0.69 g) biocompatible resin structure and a suture-free "one-click" fixation system. This affordable device simplifies surgical installation and minimizes postoperative inflammation, allowing continuous multimodal imaging-IVM, micro-CT, ultrasound, and in vivo fluorescence-for up to four weeks. The system is validated in mice bearing subcutaneous tumors generated from PDAC93-GFP pancreatic tumor organoids implanted in the DSC. Longitudinal, high-resolution imaging successfully tracks tumor expansion, active stromal remodeling characterized by progressive collagen fiber compaction and alignment, and the emergence of a highly tortuous peritumoral vascular network. Furthermore, in vivo tracking in Catchup mice reveals a distinct transition of tumor-associated neutrophils from random acute inflammatory migration to directed tumor-driven chemotaxis. Overall, this ergonomic and robust DSC design provides a highly reliable platform for sustained, high-quality assessment of tumor-stroma-immune interactions and therapeutic responses, while notably improving animal welfare.

Magnetic microswarms exhibit great potential for targeted delivery because of their excellent controllability and environmental adaptability. Enabling label-free cargo delivery and controllable release while endowing microswarms with resistance to environmental disturbances is key to expanding their application scope. In this study, we present a label-free microcargo delivery strategy based on a magnetic microswarm actuated by a magnetic tweezers system. In this approach, high-frequency magnetic fields enable autonomous cargo capture, while low-frequency fields trigger controlled release, forming a simple yet effective frequency-switching mechanism. Owing to the high-intensity magnetic field produced by the magnetic tweezers system, the microswarm exhibits significantly improved anti-interference capability, ensuring stable transport even under dynamic flow conditions. Leveraging visual feedback, the microswarm autonomously captures and stably transports various microscale cargos, including polystyrene microspheres and cell spheroids up to 400 µm in diameter. In complex structures and flowing fluid environments, microswarms that carry cargo successfully resist fluid impacts and achieve autonomous navigation. Furthermore, the controllable release of nonmagnetic cargos is realized through a frequency-switching mechanism. When a microswarm carries multiple cargos, small cargos are usually discharged first, indicating the potential of this release mechanism to sequentially release multiple cargos. This strategy circumvents the risks associated with permanent magnetic labeling and enhances the microswarm's anti-interference capability, providing essential technical support for the practical translation and clinical application of magnetic microrobots.

Pancreatic ductal adenocarcinoma (PDA) is among the deadliest malignancies, driven by metastatic progression and profound cellular heterogeneity. We previously identified glutathione S-transferase theta 1 (GSTT1) as a regulator of a slow-cycling, highly metastatic tumor cell population, suggesting that GSTT1High cells may possess stem-like properties. Here, we define the functional and molecular features of this subpopulation in metastatic PDA. Using a mCherry-tagged Gstt1 reporter system in metastatic murine PDAC cells, we enriched for Gstt1High cells and observed increased tumor sphere formation, accompanied by upregulation of stemness-associated genes including PROM1 (CD133) and activation of Wnt and FGF signaling pathways. In human PDA models, CD133HighGSTT1High cells exhibited enhanced tumor sphere initiation and expansion compared to other populations, defining a maximal stem-like state. Notably, sensitivity to FGFR inhibitors was observed only under tumor sphere conditions, highlighting a context-dependent therapeutic vulnerability. Mechanistically, FGFR3 expression correlated with GSTT1 and CD133 levels, and FGF signaling was required to sustain this state. GSTT1 knockdown reduced CD133 protein levels, impaired tumor sphere formation, and altered sensitivity to FGFR inhibition. These findings were largely recapitulated in patient-derived PDA organoids, where GSTT1 and PROM1 co-expression predicted increased tumor sphere formation and enhanced response to the multi-kinase inhibitor Nintedanib. Together, these results identify a GSTT1HighCD133High stem-like subpopulation in metastatic PDA and identify an FGFR-dependent signaling axis that sustains this state, representing a potential therapeutic vulnerability.

Malignant gliomas are lethal brain tumors characterized by profound local immunosuppression and a radically remodeled myeloid landscape. Although these tumors mobilize resident microglia and infiltrating monocyte-derived macrophages, the mechanisms governing their phenotypic convergence and diversification remain elusive. Here, we integrated single-cell profiling and spatial transcriptomics of glioma-associated microglia in the GL261 model. We identified distinct microglial states that aligned with tumor architecture, most notably Cst7-expressing disease-associated microglia (DAMs) that aggregated at the tumor invasive margin and exhibited a conserved transcriptional signature shared across various central nervous system pathologies. Interferon-γ and toll-like receptor signaling sequentially tuned stage-specific DAM features, including transient MHC-II expression and sustained PD-L1 upregulation, thereby recalibrating the local immune equilibrium by reshaping bidirectional DAM-T cell interactions during glioma progression. Our findings highlight microglial state transitions as a stage-specific layer of immune regulation in glioma that shapes T cell fate and support targeting microglial plasticity to rebalance anti-tumor immunity.

Gastro-entero-pancreatic neuroendocrine tumors (GEP-NETs) are a rare subset of cancers with increasing incidence. Due to their slow growth and lack of targetable mutations, the identification of effective treatments remains limited. One reason behind this stagnation is the lack of applicable, accurate study models. One solution is patient tumor organoids (PTOs) that maintain tumor characteristics and can scale for high throughput assays. In this study, PTOs were generated from 35 tumors of pancreatic, small intestinal, and gastric origin, obtained from 17 patients. Important subtypes including hormone functional and MEN1/VHL mutant GEP-NETs are represented, with each demonstrating growth in culture while maintaining GEP-NET immunohistochemistry and genomic characteristics. Half of G2/G3 tumors (10 of 20) could be cultured past passage 6, whereas G1 tumors (n = 15) were capable of growth until passage 4. Therapeutic targeting of the PTOs displayed both tissue-origin and grade-based response to standard of care and investigational therapies while maintaining patient tumor sensitivity and resistance. Last, a successful PTO xenograft model was developed from one PTO line. This study describes GEP-NET organoid development that demonstrates feasibility for expansion, enabling their use for translational investigations.

This work proposes an algorithm for seeking generalized feedback Nash equilibria (GFNEs) in noncooperative dynamic games. The focus is on cyber-physical systems with dynamics that are linear, stochastic, potentially unstable, and partially observed. We employ system-level synthesis (SLS) to reformulate the problem as the search for an equilibrium profile of closed-loop responses to noise, which can then be used to reconstruct a stabilizing output-feedback policy. Under this setup, we leverage monotone operator theory to design a GFNE-seeking algorithm capable of enforcing closed-loop stability, operational constraints, and communication constraints onto the control policies. This algorithm is amenable to numerical implementation, and we provide conditions for its convergence. We demonstrate our approach in a simulated experiment on the noncooperative stabilization of a decentralized power grid.

KRASG12D mutation drives oncogenic progression and creates an immunosuppressive microenvironment in cancers like pancreatic ductal adenocarcinoma and colorectal cancer. We investigate the immunomodulatory mechanisms of the KRASG12D inhibition and its synergy with natural killer (NK) cell therapies. We demonstrate that KRASG12D inhibition with MRTX1133 remodels the immune landscape by reducing myeloid-derived suppressor cell (MDSC) accumulation and facilitating infiltration and activation of NK and CD8+ T cells. Crucially, MRTX1133 reverses systemic immunosuppression, restoring the fitness of adoptively transferred NK cells. Mechanistically, KRASG12D inhibition impairs IFNGR1 palmitoylation and subsequent lysosomal degradation. MRTX1133 stabilizes IFNGR1 by reducing palmitoyltransferase expression and the palmitate pool. This stabilization increases IFN-γ/IFNGR signaling and up-regulates NK cell-activating ligands ICAM1 and ULBP1, thereby sensitizing cancer cells to NK cells. Consequently, combining MRTX1133 with IL-15 or adoptive NK cell therapy yields synergistic antitumor responses and prolonged survival. Our findings provide mechanistic rationale for combining KRASG12D inhibitors with NK cell-based immunotherapies to improve outcomes for patients with KRASG12D-mutant cancers.

Coal mining substantially alters watershed hydrology processes, yet quantitatively disentangling the dominant mechanisms within hydrological models remains challenging. Here, an enhanced Spatiotemporal Variable Source Mixed Runoff model, termed E-SVSMR, was developed to explicitly represent two key mining-induced perturbations: enhanced slope infiltration caused by mining fractures, parameterized through a soil-fissure dual-permeability system with an equivalent saturated hydraulic conductivity, and riverbed leakage, characterized by a nonlinear discharge-dependent function. The physical framework was initially validated at the hillslope scale via high-fidelity 3D variably-saturated flow modeling (COMSOL) in the Wujiayao catchment (78.7 km²). The physical basis of the framework was first evaluated at the hillslope scale in the Wujiayao catchment using high-resolution three-dimensional variably saturated flow simulations. Results demonstrate that goaf-induced fractures form preferential flow pathways and increase effective infiltration by approximately one order of magnitude. When upscaled to the Luzhuang catchment, the E-SVSMR markedly improved runoff simulation, increasing the Nash-Sutcliffe efficiency from 0.41 in the traditional SVSMR to 0.89. Continuous simulations from 1996 to 2009 indicated a cumulative mining-induced runoff loss of 1.73 × 10⁸ m³. Mechanistic decomposition of a representative storm event further suggested that enhanced slope infiltration accounted for approximately 70% of the runoff reduction, whereas riverbed leakage contributed the remaining 30%. Application to the independent Lingshi catchment further demonstrated the transferability of the model. By correcting the systematic post-1994 runoff overestimation in the baseline model, the E-SVSMR increased the Nash-Sutcliffe efficiency from -2.27 to 0.88, and successfully captured the timing and magnitude of mining-driven hydrological disturbances. Overall, the E-SVSMR provides a physically interpretable and transferable modeling framework for diagnosing mining impacts on runoff generation, with implications for flood forecasting and water-resource management in mining-affected catchments.

Although nanomaterials (NMs) have shown great promise across various fields due to their advantageous characteristics, concerns about their environmental consequences before large-scale production and application are increasing. The life cycle assessment (LCA) is a powerful tool for investigating the potential environmental impacts of emerging NMs. However, LCA practices are facing the challenge of lacking characterization factors (CFs) to estimate the environmental implications of NM release. CFs comprise of information on fate factors, exposure factors, and effect factors (EFs) using the USEtox framework. Therefore, this study aimed to address the insufficient toxicity testing data by developing eight machine learning (ML) regression algorithms to predict freshwater ecotoxicity EFs of NMs from their physicochemical properties. The results showed that the AdaBoost model exhibited the best predictive performance, with an R2 of 0.786, root-mean-square error (RMSE) of 0.711, and mean absolute error (MAE) of 0.625, and generalized well to an external validation set, achieving an RMSEₑₓₜ of 0.654 and MAEₑₓₜ of 0.550. The Shapley Additive Explanations (SHAP) analysis identified chemical composition, surface area, species, electronegativity, size, and diameter as the important features. Further, the estimated freshwater ecotoxicity EFs and CFs for graphene, derived from the optimal AdaBoost model and Monte Carlo simulation, were 23.15-200.24 potentially affected fraction (PAF)·m3·kg-1 and 231.49-2002.40 PAF·day·m3·kg-1, respectively. This study could not only help identify the potential freshwater ecotoxicity of NMs during the early development stage but also inform appropriate decisions for risk management in the absence of reliable data.

Tissue engineering (TE) remains a cornerstone of regenerative medicine, aiming to bypass the limitation of organ transplantation through the fabrication of functional tissue substitutes. Traditionally, TE has followed two primary paradigms: the top-down approach, utilising single cells seeded on a scaffold, and the bottom-up approach, employing cell spheroids as building blocks. While top-down offers architectural and structural control, bottom-up promotes self-assembly, native-like extracellular matrix deposition, and intercellular signalling. However, modern techniques increasingly blur this dichotomy, creating a spectrum of cell-based fabrication approaches. This review evaluates the diverse approaches across four major tissue classes: epithelial (pancreas as an example), connective (cartilage), muscle (heart), and nervous (brain) tissues. For each tissue, we examine notable studies to evaluate how different assembly methods recapitulate native tissue properties. By reviewing case studies across diverse tissue types, we highlight the relative strengths and limitations of various fabrication strategies. Although this review is limited by a selective cross-section of literature within a rapidly advancing technological landscape, it provides critical insights into optimising next-generation tissue constructs. In conclusion, we posit that there is no universal fabrication strategy; rather, the future of the field depends on tailoring approaches along this single-cell-to-spheroid spectrum based on the specific architectural and functional demands of the target tissue.

Accurate classification of diabetes into type 1, type 2, or monogenic forms remains a major challenge. This difficulty arises from overlapping clinical features and substantial heterogeneity within these conditions, particularly in adults, due to nonconforming presentations such as obesity coexisting with autoimmunity. Consequently, misclassification is common and can contribute to suboptimal therapeutic decisions, delayed achievement of glycemic targets, and increased risk of long-term complications. Growing evidence supports a spectrum-based view of diabetes pathophysiology rather than rigid categorical definitions, highlighting the need for improved biomarkers and a deeper mechanistic understanding of disease processes. To address this knowledge gap, over the past two decades we have conducted studies of human pancreata obtained through the Network for Pancreatic Organ donors with Diabetes (nPOD) and related programs, enabling in situ investigation of disease pathogenesis. Insights derived from such tissues and their biological relevance are based on limited clinical data at the time of terminal hospitalization. Challenges with postmortem diabetes classification have led to the creation of classification pathways that include not only clinical variables but also histopathology and genetics. For example, including the presence of insulin-negative islets and insulitic lesions on histopathology and clinical features such as C-peptide loss and increased type 1 diabetes genetic risk score, over clinical history alone, supports a more holistic diagnosis of type 1 diabetes, which may include atypical forms of diabetes. Continued integration of organ donor-based research with comprehensive clinical data is essential to refine classification systems, improve diagnostic accuracy, and advance personalized approaches to diabetes care.

Despite advancements in therapeutic cancer vaccines, clinical translation has been hindered by limited efficacy, with Sipuleucel-T remaining the only FDA-approved therapeutic cancer vaccine to date. However, recent advances in personalized mRNA vaccines, such as Moderna's mRNA-4157 and BioNTech's autogene cevumeran, have demonstrated significant reductions in recurrence risk and improved survival across several cancer types, renewing optimism in the field. Personalized cancer vaccines leverage patient-specific tumor antigens to initiate potent and targeted immune responses. This review outlines various classes of personalized vaccines, including DNA-, mRNA-, peptide-, dendritic cell-, and whole-cell-based platforms, and examines the immunological challenges they face, such as tumor heterogeneity, immunosuppressive microenvironments, and inadequate immune memory. To address these limitations, both conventional and nanotechnology-enhanced delivery systems have been developed. Notably, nanovaccines constructed from lipid-polymer hybrids, biomimetic membranes, and stimulus-responsive materials enable codelivery of neoantigens and immunostimulatory agonists, promoting enhanced lymph node targeting, dendritic cell activation, and antigen cross-presentation. Furthermore, biomimetic formulations incorporating autologous tumor membranes preserve native antigenic diversity and allow dynamic adaptation to evolving tumors. When integrated with artificial intelligence for antigen selection and multiomics for patient stratification, these platforms accelerate vaccine design and improve precision. Combination regimens with immune checkpoint inhibitors or other agents further potentiate efficacy and promote durable antitumor immunity. Increasing clinical evidence, especially in melanoma and pancreatic cancer, underscores the potential of these strategies to induce long-term protection and reduce recurrence. Overall, next-generation personalized cancer vaccines are advancing the transition from reactive treatment to proactive, precision-controlled cancer immunotherapy.