OBGECTIVE: Pancreatic Ductal Adenocarcinoma (PDAC) remains a lethal malignancy with limited therapeutic options. This study aimed to identify brown adipocyte-related genes (BARGs) influencing PDAC prognosis and explore their roles in the tumor microenvironment (TME) and immunotherapy response. METHODS: Transcriptomic and proteomic data from TCGA, GEO, ICGC, and CPTAC databases were analyzed to screen prognostic BARGs. Immune infiltration, immunotherapy prediction (via TIDE, IRnet, and TCIA), and drug sensitivity analyses were conducted. Single-cell RNA sequencing (CRA001160 dataset) and experimental validation (qPCR in pancreatic cancer cell lines) were performed to validate findings. RESULTS: CALU emerged as a core prognostic gene, significantly overexpressed in PDAC tissues and correlated with advanced tumor grade. High CALU expression was linked to stromal cell activation (e.g., cancer-associated fibroblasts, M1 macrophages) and suppressed T-cell infiltration, indicating immunosuppressive TME remodeling. CALU predicted resistance to CTLA4 inhibitors but showed no significant association with PD1 blockade. Drug sensitivity analysis revealed correlations between CALU and chemotherapeutic agents (e.g., TAK-715, LGK974). Single-cell analysis localized CALU to malignant and stromal cells, highlighting its role in PERIOSTIN-mediated fibroblast-malignant cell communication. Experimental validation confirmed elevated CALU expression in pancreatic cancer cell lines compared to normal cells. CONCLUSION: CALU is a critical regulator of PDAC progression, influencing stromal-TME interactions and immune evasion. It serves as a potential prognostic biomarker and therapeutic target, offering insights into combination strategies targeting stromal-immune crosstalk in PDAC.

e16370 Background: The optimal treatment strategy for patients with synchronous pancreatic ductal adenocarcinoma with liver metastases (sPDACLM) remains uncertain, and the role of surgery in selected patients is controversial. Comparative evidence regarding survival outcomes between upfront surgery (US) and surgery following systemic therapy (SS) in this population is limited. This study aimed to compare overall survival (OS) among patients with sPDACLM treated with different therapeutic strategies and identify prognostic factors associated with survival. Methods: This multicenter retrospective cohort study included patients with sPDACLM treated at seven tertiary medical centers in China between 2017 and 2025. Patients were categorized into three groups: upfront surgery group (US), effective systemic therapy followed by surgery (E-SS), defined as achievement of partial response [PR] per RECIST 1.1 with a > 85% reduction or normalization of serum CA19-9 after systemic therapy; and ineffective systemic therapy followed by surgery (I-SS), defined as fails to achieve PR or < 85% reduction in CA19-9 after systemic therapy. The primary endpoint was OS. Results: A total of 72 patients were included, comprising 21 in the US group, 24 in the E-SS group, and 27 in the I-SS group. Baseline characteristics differed substantially among groups, with 21 of 24 covariates demonstrating a standardized mean difference (SMD) > 0.1. Kaplan–Meier analysis showed significantly improved survival in the E-SS group compared with the US group (median OS: 31.9 vs. 8.0 months; 1-year OS: 90.47% vs. 33.33%; 3-year OS: 40.50% vs. 15.23%) and the I-SS group (median OS: 17.0 months; 1-year OS: 66.62%; 3-year OS: 0%) (P < 0.0001). Univariate Cox regression identified effective systemic therapy as a favorable prognostic factor for OS (hazard ratio [HR], 0.265; 95% CI, 0.123–0.571; P < 0.001). After adjustment for clinically relevant covariates, multivariable Cox analysis confirmed that effective systemic therapy (adjusted HR [aHR] = 0.225, 95% CI: 0.099–0.512; P < 0.001) and tumors located in the pancreatic body–tail region (aHR = 0.443, 95% CI: 0.222–0.882; P = 0.03) were independently associated with prolonged OS. Sensitivity analyses yielded consistent results, supporting the robustness of the findings. Conclusions: Among patients with sPDACLM, treatment strategies incorporating systemic therapy are associated with improved survival compared with upfront surgery. Moreover, tumors arising in the pancreatic body–tail region represent an independent favorable prognostic factor. These findings support prioritizing systemic therapy in the management of sPDACLM, with subsequent surgical resection following systemic therapy offering a clinically meaningful survival benefit.

ABSTRACT Pancreatic ductal adenocarcinoma (PDAC) remains one of the most lethal malignancies, with chemotherapy response evaluation currently relying on the Response Evaluation Criteria in Solid Tumors (RECIST). However, RECIST inadequately captures the unique biology of PDAC, particularly stromal fibrosis–induced radiographic pseudoprogression. We developed and externally validated a multivariable prognostic prediction model, the chemotherapy Progression Decision (cPD) score, integrating biological markers and imaging metrics to guide therapeutic decision‐making after the first RECIST assessment. This multicenter retrospective study enrolled 616 PDAC patients across three cohorts: the Training Cohort (n = 155), Validation Cohort 1 (n = 263), and Validation Cohort 2 (n = 198). Multivariable Cox regression identified four independent prognostic predictors: neutrophil‐to‐lymphocyte ratio, baseline carbohydrate antigen 19‐9, tumor maximum cross‐sectional rate change ratio, and emergence of new lesions. These were integrated into an integer‐based score (8–13 points) stratifying patients into good prognosis (GP), poor prognosis (PP), and critical prognosis (CP) groups. The cPD model demonstrated robust discrimination and calibration, with significant survival differences across groups in all cohorts. Notably, the cPD score identified a subset of patients (GP and PP groups) for whom continuing the original chemotherapy yielded significantly better survival than switching regimens, even when RECIST classified the disease as progressive.

Hepatocellular carcinoma (HCC) is a common malignancy and a leading cause of cancer-related mortality. Current guidelines and staging systems provide coarse categories, but often miss within-stage heterogeneity and the clinical context in electronic medical records (EMRs). We present HCC-STAR (Hepatocellular Carcinoma Staging, Treatment And pRognosis), a clinically aligned large language model that reads routine EMR narratives and jointly outputs risk score-based staging, ranked guideline-consistent treatments with evidence-based rationales, and individualized survival estimates. We curated about 30,000 HCC cases from SEER and expanded them into EMR-style narrative training data using a clinician-validated, prompt-based augmentation workflow. On this corpus, we developed a knowledge-aligned reasoning framework optimized with a step-verifiable composite reward, moving beyond text-level memorization of clinical guidelines. In a multi-center cohort of 6,668 patients from 12 hospitals in China, HCC-STAR achieved state-of-the-art performance in treatment recommendation and risk stratification compared with clinical guidelines and competitive models, including GPT-5 and Gemini-2.5 Pro. Hypothetical overall-survival analysis showed a median survival of 51 months under adherence to HCC-STAR recommendations, compared with 29 and 32 months under BCLC and CNLC. In clinician-centric evaluations, blinded hepatobiliary specialists rated HCC-STAR's reasoning and evidence-based justifications as trustworthy. The model surpassed resident and attending physicians in treatment accuracy and helped physicians make more accurate decisions faster when used as an assistant. These findings support HCC-STAR as a reliable and verifiable decision-support system for risk stratification and precision therapy in HCC.

Acetaminophen-induced liver injury (AILI) is a major cause of acute liver failure, yet single-target therapies like N-acetylcysteine remain inadequate due to its complex pathogenesis. To address this challenge, we propose a dual-action defense-regeneration strategy that concurrently blocks multi-death pathways and promotes hepatocyte regeneration. Specifically, the therapeutic metal gallium is doped into V2C MXene nanozymes (Ga-V2C) to surpass conventional nanozymes by integrating sustained antioxidant activity for death signaling blockade, multi-pathway regulation of cell death networks, and activation of pro-regenerative molecules. In vivo, Ga-V2C nanozymes exhibited superior protective efficacy over N-acetylcysteine against AILI. Mechanistic investigations revealed that the Ga-V2C nanozymes disrupt the synergistic amplification of liver injury by simultaneously inhibiting three key death pathways: oxidative stress (via ROS scavenging, reduce JNK phosphorylation, and activated Nrf2/HO-1), apoptosis (via restored Bcl-2/Bax balance), and ferroptosis (by suppressed iron-dependent lipid peroxidation and upregulated SLC7A11/FTH1/FTL1). Notably, Ga-V2C nanozymes fostered a pro-regenerative microenvironment by activating Wnt/βCAT pathways signaling and key cell cycle drivers (CCND1, MYC, PCNA), thereby enhancing hepatocyte regeneration. This work not only offers a promising therapeutic approach for AILI but also significantly expands the scope of nanozyme-based therapeutics for complex diseases requiring multi-target intervention.

Background Patients with initially unresectable intrahepatic cholangiocarcinoma (iCCA) have poor prognoses and the current first-line treatments remain unsatisfactory. Our study aimed to evaluate gemcitabine-based chemotherapy plus PD-1/PD-L1 inhibitors and tyrosine kinase inhibitors (TKIs) for iCCA. Methods In this multicenter retrospective cohort study, 392 patients were included in the full cohort analysis. For the primary analysis, 88 patients received gemcitabine-based chemotherapy (GEMCIS or GEMOX), and 177 patients received the same chemotherapy backbones combined with PD-1/PD-L1 inhibitors and TKIs. The primary outcome was to evaluate overall survival (OS). The propensity score matching (PSM) method was utilized to reduce potential confounders. Results Overall, 392 patients were included from 12 hospitals across China, with data from January 2016 to December 2024. In the full cohort analysis, the combination of GEMCIS/GEMOX with PD‑1/PD‑L1 inhibitors and TKIs significantly prolonged median OS compared to GEMCIS/GEMOX alone (26.7 vs. 15.3 months; hazard ratio [HR] 0.47, 95% CI: 0.35–0.63, P < 0.001). In the primary analysis, the combination strategy was associated with a median OS of 24.8 months and median progression-free survival (PFS) of 11.0 months, with HRs of 0.56 (95% CI: 0.38–0.83, P = 0.004) for OS and 0.46 (95% CI: 0.32–0.66, P < 0.001) for PFS after PSM. Importantly, our treatment strategy increased the rate of conversion surgery to 42% for locally advanced disease, ultimately resulting in significantly improved overall outcomes (HR for OS: 0.17 [95% CI: 0.05–0.61]). Conclusion Combining gemcitabine-based chemotherapy with PD-1/PD-L1 inhibitors and TKIs may provide a promising new first-line treatment option for patients with initially unresectable iCCA.

Background/Objectives: The limited targeting efficiency and systemic toxicity of conventional medicine present significant challenges in the treatment of skeletal disorders, such as bone tuberculosis. To address these limitations, we developed a bone-targeting nanomicelle delivery system functionalized with alendronate (ALN), designated ALN-PLGA-mPEG@RPT, to improve the targeted delivery and therapeutic efficacy of rifapentine (RPT) in bone tissue. Methods: The ALN-PLGA-mPEG blank micelles, prepared in accordance with our research group’s optimized protocol, were loaded with RPT and subjected to systematic formulation optimization. The resulting nanomicellar system was comprehensively characterized in terms of its physicochemical properties, including particle size and polydispersity index (PDI). Additionally, drug-loading capacity, encapsulation efficiency, and in vitro release curve were evaluated. Bone-targeting efficacy was assessed using in vivo imaging techniques, while biodistribution and safety profiles were determined through in vivo distribution studies and histopathological examination. Results: The optimized ALN-PLGA-mPEG@RPT nanomicelles exhibited a mean particle size of 101.90 ± 4.17 nm, and a PDI of 0.242 ± 0.021. The formulation achieved a drug loading of 16.74 ± 0.51% with an encapsulation efficiency of 50.27 ± 1.91%. In vitro release studies confirmed a sustained-release profile, with only 25% of RPT released within 12 h. In vivo imaging revealed significantly enhanced bone-targeting capability in the ALN-modified group, showing a 1.93-fold higher drug accumulation in bone tissue compared to blood. Histopathological analysis indicated no observable pathological alterations in major organs. Conclusions: The ALN-PLGA-mPEG@RPT nanomicelle system exhibits favorable bone-targeting efficiency, sustained-release properties, and biocompatibility, representing a promising strategy for the precise treatment of bone tuberculosis and other skeletal diseases.

Multi-phase contrast-enhanced CT (CECT) scans are often scattered across multiple institutions and contain incomplete phase in parts of institutions due to the strict data-protection regulations and the disparity of phase integrality. While federated learning (FL) enables training a privacy-preserving model collaboratively across institutions, it often suffers from significant performance degradation for liver lesion segmentation caused by heterogeneity in different institutions. To tackle the challenges, we present FedOG to guide a deep convolutional neural network collaboratively segment liver lesions for minimizing interference on 3,668 CECT multi-phase CECT scans from five different institutions. Specifically, FedOG adjusted the gradients from local models trained with incomplete phases of CECTs via orthogonal gradient decomposition to alleviate the interference. During each adjustment, the optimal gradient for updating the global model is determined by Bayesian optimization. Experiment results have shown that FedOG improves the Dice score by 1.67% on a large real-world clinical dataset, 1.13%, and 3.03% on two publicly available datasets. We anticipate our study will enable a heterogeneity-robust, search-efficient, and privacy-preserving federated training framework using multi-phase CECT. We also found out that FedOG is especially beneficial for underdeveloped regions where institutions often have missing or low-quality phases of multi-phase CECT scans.

e16403 Background: Patients with pancreatic ductal adenocarcinoma (PDAC) liver metastases have a poor prognosis, and the best treatment remains uncertain. While systemic chemotherapy is standard, the benefit of liver metastasis resection in oligometastatic patients lacks high-level evidence, with many studies affected by bias and mixed results. This study evaluates the effect of synchronous liver metastasis resection on overall survival, adjusted for confounders through multicentre, propensity score-matched analysis, and identifies independent prognostic factors. Methods: This retrospective multicenter study in seven Chinese centers enrolled patients with pancreatic cancer and liver metastases from January 2017 to July 2025. Patients were divided into complete liver lesion management and partial/untreated groups based on intervention extent. Propensity score matching used covariates like sex, tumor location, and metastasis burden, with oligometastasis defined as < 3 lesions without extrahepatic spread. The primary endpoint was overall survival. Kaplan- Meier and Cox regression analysis compared groups and identified prognostic factors. Results: After matching, 48 patients (24 in each group) formed a balanced cohort. No significant baseline differences were found. Median survival was 19 months in the complete management group and 25 months in the partial/untreated group, with no significant difference (HR = 1.07. 07, P = 0. 853). Univariate analysis showed preoperative neoadjuvant chemotherapy, higher BMI, and lower ALT levels were associated with better prognosis. Multivariate analysis confirmed preoperative neoadjuvant chemotherapy and higher BMI as independent favorable factors. Subgroup analysis indicated complete metastasis management did not improve survival among those responding to neoadjuvant chemotherapy. Conclusions: This study found that aggressive liver metastasis resection does not significantly prolong survival. Systemic therapy, especially neoadjuvant chemotherapy, and nutritional status (BMI) are stronger prognostic factors. The results suggest focus should be on systemic and health management rather than aggressive surgery, providing high- level evidence supporting' chemotherapy over surgery" in this setting.

The aim of our study was to determine the outcomes of liver transplant recipients receiving either lamivudine (LAM) monotherapy or LAM combined with low-dose intramuscular (IM) hepatitis B Immunoglobulin (HBIG) therapy. We performed a retrospective review of the medical records of patients that had had liver transplantation in a single center for HBV-related liver diseases from December 1999 to June 2004. A total of 165 patients received LAM monotherapy (51 patients) or combined prophylaxis (114 patients) post-liver transplantation (LT) with a mean follow-up of 20.13 months. Hepatitis B relapsed in 21 patients of the hepatitis B surface antigen (HBsAg) carriers who received LAM monotherapy, with a 1- and 2-yr actuarial risk of 27.4% and 39.7%. Recurrence occurred in 16 patients of 114 patients receiving the combined prophylaxis, with a 1- and 2-yr recurrence rate of 13.5% and 15.2% (P = 0.024). A total of 25 cases (67.6%) with YMDD mutants were detected in all the 37 patients, 14 cases (66.7%) in the monotherapy group and 11 cases (68.8%) in the combination group. In conclusion, LAM and low-dose intramuscular HBIG treatment demonstrates a better result than LAM monotherapy, as prophylaxis against post-LT reinfection of the graft, but the safety and efficacy as a substitution for high-dose intravenous HBIG with LAM needs to be investigated further.

Acute liver injury (ALI), driven by diverse insults such as drug toxicity and ischemia-reperfusion, poses a high mortality risk and lacks targeted therapies. While reactive oxygen species (ROS), neutrophil extracellular traps (NETs), and a coordinated cell death pathway PANoptosis have been implicated, their interplay as a unified pathogenic axis remains elusive. Here, by integrating multi-omics analyses of clinical databases and patient samples, we systematically identified and validated the ROS/NETs/PANoptosis axis as a central driver of hepatocyte damage across multiple ALI etiologies. To therapeutically target this axis, we engineered a liver-targeted gallium-quercetin nanocomposite (Ga@Que) via coordination-driven self-assembly. Ga@Que effectively overcomes the poor bioavailability of natural quercetin. In murine models of acetaminophen-induced and ischemia-reperfusion liver injury, Ga@Que exhibited significant liver accumulation, potently scavenged ROS, suppressed neutrophil infiltration and NETs formation, and attenuated PANoptosis. Consequently, Ga@Que treatment markedly mitigated liver damage and inflammation, outperforming its individual components. Our study not only delineates a novel pathogenic paradigm in ALI but also introduces Ga@Que as a promising precision nanotherapeutic, offering a synergistic and translatable strategy to disrupt this deleterious cascade.

Early detection of focal liver lesions is critical for patient outcomes, but current imaging techniques have limitations. This study developed MULLET, an AI system using contrast-enhanced CT, and assessed its performance in assisting radiologists with FLL detection and characterization. A retrospective multicenter, multi-reader, multi-case trial was conducted, in which 10 radiologists independently interpreted 375 patients' clinical images with and without MULLET assistance. This trial was registered on ClinicalTrials.gov (ID: NCT06068413) with the registration name "A Retrospective, Multicenter, Multiple-viewer-multiple-case (MRMC) Clinical Trial to Evaluate the Safety and Efficacy of CT Image-assisted Detection Software for Focal Liver Lesions" in April 2023. Diagnostic performance was compared using area under the receiver operating characteristic curve (AUC) analysis. Without MULLET, the average AUC was 0.8188 (sensitivity: 74.45%, specificity: 87.77%). With MULLET, the average AUC significantly improved to 0.9268 (p < 0.0001), sensitivity increased to 89.95% (p = 0.0003), and specificity rose to 93.57% (p = 0.0063). MULLET also showed improved performance across different FLL sizes and subtypes. Moreover, the average reading time for radiologists was reduced by 21.8% (p < 0.0001). In conclusion, MULLET demonstrated promising performance in significantly improving radiologists' sensitivity, accuracy, and efficiency in FLL detection and diagnosis.

ABSTRACT The global landscape of liver cirrhosis has undergone a significant transformation with the widespread implementation of viral vaccines, and metabolic dysfunction‐associated steatotic liver disease (MASLD) cirrhosis has emerged as a predominant etiology. Given the differences in the pathogenic mechanisms between viral and MASLD cirrhosis, conventional diagnostic and therapeutic strategies for viral cirrhosis have demonstrated limited applicability in MASLD cirrhosis. The pathogenesis of MASLD cirrhosis involves multifaceted interactions between metabolic dysregulation, inflammatory responses, immune dysregulation, and gut–liver axis disruption, presenting substantial challenges in clinical management. Although liver biopsy remains the gold standard for diagnosis, complications, including hemorrhage, infection, pain, and pneumothorax, lead to suboptimal patient compliance, thereby prompting the development of various noninvasive diagnostic tools. Current therapeutic approaches primarily target key pathological mechanisms through metabolic regulation of glucose and lipid homeostasis, inflammation control, and microbial balance restoration. This review summarized the epidemiological transition from viral to MASLD cirrhosis and aimed to provide a conceptual framework for optimizing diagnostic and therapeutic strategies in the MASLD era.

Pupil examination has been used as a basic measure in critically ill patients and has great importance for the prognosis and management of disease. An automated pupillometer is a computer-based infrared digital video system by which the accuracy and precision of the pupil examination are markedly improved. We conducted an observational study of pupil assessment with automated pupillometry in clinical liver transplantation settings, including pretransplant evaluations and posttransplant surveillance. Our results showed that unconscious patients (grade 4 hepatic encephalopathy) had a prolonged latency phase (left side: 283 +/- 80 milliseconds; right side: 295 +/- 96 milliseconds) and a reduced pupillary constrictive ratio (left direct response: 0.23 +/- 0.10; left indirect response: 0.21 +/- 0.07; right direct response: 0.20 +/- 0.08; right indirect response: 0.21 +/- 0.08) in comparison with normal and conscious patients. After liver transplantation, the recovery of pupillography in these patients was slower than that in conscious patients. However, the surviving recipients without major complications all had a gradual recovery of pupillary responses, which occurred on the first or second posttransplant day. We also reported 4 cases of futile LT in the absence of pretransplant pupillary responses and other pupillary abnormalities revealed by automated pupillometry in our study. In conclusion, patients with grade 4 hepatic encephalopathy had a sluggish pupil response and a delayed recovery pattern after LT. An automated pupillometer is potentially a supplementary device for pretransplant screening and posttransplant monitoring in patients undergoing LT, but further prospective studies are required.