Introduction

Advanced fibrosis is the primary prognostic determinant in patients with metabolic dysfunction-associated steatotic liver disease (MASLD). Current estimates suggest that more than 25% of adults worldwide are affected by MASLD, and it is expected to become the leading cause of liver transplants in the future [1]. Current non-invasive tools like Fibrosis-4 (FIB-4) lack optimal balance between sensitivity and specificity. We aimed to develop and compare two ML models: Extreme Gradient Boosting (XGBoost) and logistic regression to predict advanced fibrosis using routinely available clinical data, with liver stiffness measurement via transient elastography as the reference standard.

Aims & Methods

We used data from NHANES 2017-2018. Adults with valid liver stiffness measurement from transient elastography were included. Advanced fibrosis was defined as liver stiffness >=8.2 kPa. Input features included AST, ALT, platelet count, BMI, triglycerides, glucose, and diabetes status. A total of 2,625 participants were included. Data were split 80:20 into training and testing sets respectively using stratified sampling to preserve fibrosis prevalence (~9%). To address class imbalance, the XGBoost model was trained using a weighted loss function. Logistic regression was used as a baseline comparator. Classification thresholds were optimized using Youden's Index to determine the best trade-off between over- and under-classification of fibrosis. Model performance was evaluated using AUROC, accuracy, sensitivity, and specificity.

Results

Both models demonstrated good discriminatory performance. The XGB model achieved an AUROC of 0.73, with sensitivity of 78.9%, specificity of 61.2%, and accuracy of 77.3%. While logistic regression yielded a slightly higher AUROC of 0.77, greater specificity (81.6%), and lower sensitivity (60.7%) and accuracy (62.7%).

Conclusion

Both ML models using simple clinical and biochemical inputs demonstrated reasonable performance in detecting advanced fibrosis among patients with MASLD. The XGBoost model provided a more balanced profile with improved sensitivity, while logistic regression offered better specificity. To address class imbalance (~9% prevalence), XGBoost used a weighted loss function, and both models were calibrated using Youden's Index to optimize classification thresholds. Limitations include reliance on a non-invasive fibrosis surrogate (transient elastography), and lack of external validation. Future work should validate these models in prospective cohorts and explore integration into digital clinical tools.

References

[1] McTeer, M., Applegate, D., Mesenbrink, P., Ratziu, V., Schattenberg, J. M., Bugianesi, E., Geier, A., Romero Gomez, M., Dufour, J. F., Ekstedt, M., Francque, S., Yki-Jarvinen, H., Allison, M., Valenti, L., Miele, L., Pavlides, M., Cobbold, J., Papatheodoridis, G., Holleboom, A. G., Tiniakos, D., … LITMUS Consortium investigators (2024). Machine learning approaches to enhance diagnosis and staging of patients with MASLD using routinely available clinical information. PloS one, 19(2), e0299487. https://doi.org/10.1371/journal.pone.0299487
[2] Satapathy, S. K., Bernstein, D. E., & Roth, N. C. (2022). Liver transplantation in patients with non-alcoholic steatohepatitis and alcohol-related liver disease: the dust is yet to settle. Translational gastroenterology and hepatology, 7, 23. https://doi.org/10.21037/tgh-2020-15

Introduction

Acetaminophen (APAP) overdose is a primary cause of drug-induced liver injury (DILI). Although the inflammatory microenvironment plays a critical role in APAP-induced liver injury (AILI), its cellular interaction network and molecular regulatory mechanisms remain incompletely elucidated. Krüppel-like factor 4 (KLF4), a zinc finger transcription factor, is known to regulate monocyte differentiation, but its functional role in the AILI inflammatory microenvironment remains unclear. This study aims to investigate the regulatory effects of myeloid KLF4 deletion on the AILI microenvironment.

Aims & Methods

By integrating multidimensional omics analysis and a myeloid-specific Klf4 knockout mouse model, we dynamically dissected microenvironmental remodeling during AILI progression. Liver samples were collected for RNA sequencing, combined with single-cell transcriptomics, flow cytometry, and molecular interaction validation to clarify the KLF4-dependent cellular subpopulation interaction network. In vivo liver imaging, histology, biochemical assays, and molecular experiments were performed to validate KLF4-dependent intercellular interactions. Primary hepatocytes and hepatocyte cell lines were used for in vitro verification.

Results

Myeloid-specific KLF4 deletion led to an increase in CXCL2⁺CXCL3⁺CD24⁺ monocytes, which activated the NF-κB signaling pathway to secrete CXCL2/CXCL3 chemokines. These chemokines bound to the CXCR2 receptor on neutrophils, driving neutrophil infiltration and neutrophil extracellular traps (NETs) formation, while amplifying the TNF-α signaling pathway. Concurrently, myeloid KLF4 knockout induced enrichment of CXCL1⁺IL17RA⁺ endothelial cells. Neutrophil-derived NETs promoted endothelial phenotypic switching toward pro-inflammatory polarization, while endothelial-derived CXCL1 further recruited neutrophils, establishing a "monocyte-neutrophil-endothelial" cellular subpopulation interaction that perpetuated microenvironmental inflammation.

Conclusion

This study reveals the molecular mechanism by which KLF4 drives inflammatory cascades in the AILI microenvironment through regulation of the myeloid-endothelial interaction network. We propose therapeutic strategies targeting NETs formation or the CXCL2/CXCL3-CXCR2 axis, providing novel targets for precise intervention in AILI.

Introduction

Primary biliary cholangitis (PBC) is a rare, chronic autoimmune liver disease characterized by progressive destruction of small intrahepatic bile ducts, leading to cholestasis and end-stage liver disease. Liver transplantation (LT) remains the definitive treatment, with excellent long-term outcomes. However, a subset of patients experiences disease recurrence (rPBC), which may impact prognosis. Due to the rarity of the condition, risk factors for rPBC remain poorly defined.

Aims & Methods

This study aims to determine the prevalence of rPBC after LT and identify potential risk factors associated with recurrence. A systematic search was conducted in MEDLINE, Embase, and Cochrane CENTRAL using a predefined strategy. We included full-text studies reporting the number of PBC patients who developed rPBC post-LT. Abstracts and studies lacking clear data on prevalence or risk factors were excluded. Prevalence estimates were pooled using a two-step approach with logit transformation and generalized linear mixed models. Hazard ratios (HRs) and 95% confidence intervals (CIs) were synthesized using a random-effects model with restricted maximum likelihood estimation. All analyses were performed using R Studio.

Results

Fifty-five studies met inclusion criteria, most conducted in Europe, followed by North America, Asia, and Australia; two studies were multi-continental. A total of 5,715 patients with PBC who underwent LT were included, with 9.98% being male. Reported prevalence of rPBC ranged from 0% to 50%, with a pooled prevalence of 15% (95% CI: 12%-19%). Significant risk factors for rPBC included younger age at LT (HR = 0.97, 95% CI: 0.96-0.98) and use of tacrolimus in maintenance immunosuppression (HR = 1.79, 95% CI: 1.09-2.95). Protective factors included cyclosporine-based maintenance immunosuppression (HR = 0.50, 95% CI: 0.28-0.87) and post-LT prophylactic use of ursodeoxycholic acid (HR = 0.42, 95% CI: 0.30-0.58). Donor age, donor-recipient sex mismatch, cold ischemia time, and use of tacrolimus, cyclosporine, or azathioprine in initial immunosuppression, as well as maintenance azathioprine or occurrence of acute cellular rejection, were not significantly associated with rPBC.

Conclusion

This is the largest meta-analysis to date evaluating rPBC, including 5,715 liver transplant recipients with PBC. The overall recurrence rate was 15%. Younger age at transplantation and tacrolimus-based immunosuppression were associated with higher recurrence risk, while cyclosporine and post-LT ursodeoxycholic acid appeared protective. These findings may inform risk stratification and guide post-transplant management in PBC. Prospective studies are needed to validate these associations and refine strategies to prevent rPBC.

Disclosure

RVM has received funding from PSC Support. GGLC has received a research grant from IPSEN.

Introduction

Acetaminophen-induced liver failure (AILF) has a high fatality rate and limited therapeutic approaches at present. Type 1 conventional dendritic cells (cDC1) are one of the major subgroups of cDCs in the liver, which is not only an important type in initiating adaptive immune response, but also can promote the generation, stability and function of Regulatory T (T_reg) cells through various mechanisms. cDC1 maintains liver immune homeostasis by regulating the balance between immune activation and tolerance. Nevertheless, the precise involvement and mechanistic contributions of hepatic cDCs in AILF remain to be elucidated.

Aims & Methods

Multiplex Immunohistochemistry (mIHC), flow cytometry, spatial transcriptomics (ST), single-cell RNA sequencing (scRNA-seq) and intravital 3D imaging were used to characterize the distribution and frequency of the cDC1. Bulk RNA-seq, metabolomic analyses, Cleavage Under Targets and Tagmentation (CUT&Tag) and Chromatin immunoprecipitation (ChIP)-qPCR assay were employed to explore the molecular mechanisms. CD11c^creXCR1^f/f, CD11c^creIDO1^f/f, CD11c^creP300^f/f and Foxp3^DTR mice, various mouse primary cells, human liver and serum samples were applied to verify the discoveries of bioinformatics analysis. Intravenous delivery of tolerogenic dendritic cells (tolDCs) was used to elucidate their regulatory effects on T_reg cells and roles in the AILF process.

Results

Our analyses revealed that the amount of cDC1 was significantly reduced across both clinical AILF cohorts and mouse AILF models, with the magnitude of depletion exhibiting dose- and duration-dependent associations with APAP exposure. mIHC, ST, scRNA-seq and intravital 3D imaging delineated that cDC1 depletion impaired T_reg cells priming capacity, ultimately disrupting hepatic immune homeostasis and exacerbating AILF. Intravenous tolDCs injection demonstrated therapeutic efficacy by orchestrating hepatic T_reg cells infiltration, attenuating inflammatory responses, potentiating liver regeneration, and ultimately reversing AILF-associated pathological manifestations. The therapeutic effect of intravenous tolDCs administration was abolished after diphtheria toxin-mediated T_reg cells depletion in Foxp3^DTR mice. CD11c^creXCR1^f/f mice andwild-type mice clearing cDC1 with XCL1-antibody had exacerbated liver pathology. Mechanistically, EP300-mediated histone 3 lysine 27 acetylation (H3K27ac) in IDO1 promoter region enforced stable tolerogenic states of cDC1, these epigenetically programmed cDC1 subsequently initiated tryptophan catabolism, generating kynurenine metabolites that triggered both differentiation and functional maturation of T_reg cells. Both the DC-specific knockout and inhibition of EP300 could reduce the H3k27ac levels of IDO1. And the CD11c^creIDO1^f/f mice with APAP exposure had less T_reg cells recruitment and worse phenotype.

Conclusion

This study elucidated that EP300 orchestrated tolDCs differentiation via H3K27ac/IDO1 signaling axis, this epigenetic-metabolic cascade licensed tolDCs to recruit and activate T_reg cells, thereby preserving hepatic immune homeostasis while promoting hepatocyte proliferation and tissue repair. These findings highlight novel therapeutic targets for reversing AILF pathogenesis.