PURPOSE: Patients with liver disease often experience nutritional insufficiency due to an interplay of metabolic disturbances and dietary alterations leading to decreased muscle mass and the development of protein-calori...PURPOSE: Patients with liver disease often experience nutritional insufficiency due to an interplay of metabolic disturbances and dietary alterations leading to decreased muscle mass and the development of protein-calorie malnutrition (PCM). This study aimed to evaluate the prevalence of PCM in patients with steatotic liver disease (alcohol-associated liver disease (ALD) and metabolic dysfunction-associated steatotic liver disease (MASLD)) and their impacts on mortality and healthcare utilization. METHODS: We identified hospitalizations with ALD, MASLD, and PCM using International Classification of Diseases codes in the National Inpatient Sample from 2016 to 2020. Descriptive analyses compared hospitalizations with and without PCM. Multivariable linear models adjusting for confounders evaluated the association between PCM and inpatient mortality, length of stay (LOS), and total charges. RESULTS: PCM was found to be significantly more prevalent among hospitalizations with ALD or MASLD than those with neither (ALD 175.5 versus 51.8; MASLD 149.0 vs. 50.7; neither: 49.3 per 1000 hospitalizations; p < 0.001). Among hospitalizations with ALD nor MASLD, PCM was significantly associated with higher mortality (ALD adjusted odds ratio [aOR] 1.61; 95% CI 1.55-1.66; MASLD aOR 1.89; 95% CI 1.84-1.93), LOS (ALD 3.78 more days; 95% CI 3.67-3.88; MASLD 5.65 more days; 95% CI 5.53-5.77), and total charges (ALD $45,013; 95% CI $42,549-$47,477; MASLD $74,502; 95% CI $71,214-$77,789). CONCLUSION: We found a higher prevalence of PCM among individuals with ALD compared to those with MASLD nor neither condition. PCM was associated with increased mortality, LOS, and total charges in those with ALD and MASLD. Our findings underscore the importance of early identification and management of PCM in patients with steatotic liver disease to mitigate adverse outcomes and reduce healthcare utilization.
BACKGROUND: Cholangiocarcinoma (CCA) is a common and highly refractory malignancy; however, biomarkers and therapeutic targets associated with its progression remain incompletely understood. Crotonylation is an emerging...BACKGROUND: Cholangiocarcinoma (CCA) is a common and highly refractory malignancy; however, biomarkers and therapeutic targets associated with its progression remain incompletely understood. Crotonylation is an emerging lysine acylation modification that has been implicated in multiple diseases. AIMS: In this study, we aimed to identify crotonylation-related prognostic genes in CCA. METHODS: First, differentially expressed genes (DEGs) with consistent expression patterns were identified using the TCGA-CHOL and GSE26566 datasets. Key module genes associated with the expression patterns of known lysine crotonylation-related gene (LCRG) scores were subsequently identified through weighted gene co-expression network analysis (WGCNA), and candidate genes were obtained by intersection analysis. RESULTS: Three prognostic genes, complement factor H (CFH), phosphatidylcholine N-methyltransferase (PEMT), and tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein epsilon (YWHAE), were ultimately selected to construct and validate a prognostic risk model. In addition, immune infiltration analysis revealed significant differences in mast cell abundance between the high- and low-risk groups. Moreover, the expression levels of the immune checkpoints FMS-like tyrosine kinase 3 (FLT3) and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) also differed between the two risk groups. CONCLUSIONS: These findings may have important implications for CCA immunotherapy and provide a theoretical basis for improving diagnostic and prognostic assessment in CCA.
BACKGROUND: Major cardiac events can complicate endoscopic procedures. However, the true incidence and outcomes of these events are not well defined. AIMS: We aimed to define the true incidence of cardiac arrest during i...BACKGROUND: Major cardiac events can complicate endoscopic procedures. However, the true incidence and outcomes of these events are not well defined. AIMS: We aimed to define the true incidence of cardiac arrest during inpatient endoscopic procedures and the associated outcomes. METHODS: We performed a cross-sectional retrospective analysis of the National Inpatient Sample from 2016 to 2023. Endoscopic intraprocedural cardiac arrests were identified using ICD-10 diagnostic and procedural codes. Multivariable logistic regression was used to assess patient and procedural factors associated with cardiac arrest. Outcomes among patients experiencing endoscopic intraprocedural cardiac arrest were compared with those of all other cases of in-hospital cardiac arrest. RESULTS: Among 11,137,008 inpatient endoscopic procedures, 1,110 were complicated by cardiac arrest, corresponding to one cardiac arrest per 10,125 procedures. Among assessed patient factors, congestive heart failure had the strongest association with cardiac arrest (aOR = 2.78, 95% CI 2.04-3.78; p < 0.001). Compared with upper endoscopy, colonoscopy was associated with a lower risk of cardiac arrest (aOR = 0.50, 95% CI 0.36-0.69; p < 0.001), while endoscopic retrograde cholangiopancreatography had a similar risk (aOR = 0.75, 95% CI 0.49-1.15; p = 0.183). In-hospital mortality was lower among patients with endoscopy-associated cardiac arrest compared with other cases of in-hospital cardiac arrest (30.6% vs. 71.0%; aOR-0.17, 95% CI 0.13-0.23; p < 0.001). CONCLUSIONS: Intraprocedural cardiac arrest is an uncommon complication of inpatient endoscopy. Upper endoscopy carries a higher risk than colonoscopy, and patients with congestive heart failure are at increased risk. Outcomes following endoscopy-associated cardiac arrest are more favorable than for other in-hospital cardiac arrests.
INTRODUCTION: Rapid gastric emptying (RGE) is conventionally defined as ≥ 70% emptying of a standardized solid meal at 1 h using gastric emptying scintigraphy (GES). However, this threshold is not universally adopted, an...INTRODUCTION: Rapid gastric emptying (RGE) is conventionally defined as ≥ 70% emptying of a standardized solid meal at 1 h using gastric emptying scintigraphy (GES). However, this threshold is not universally adopted, and variability exists in how accelerated gastric emptying is defined in practice. We aimed to better characterize patients with conventionally defined RGE as well as those with accelerated emptying not meeting this threshold, and to evaluate what clinical differences exist across varying degrees of accelerated gastric emptying. METHODS: We identified a cohort of 258 adult patients (≥ 18 years old) with increased gastric emptying (≥ 30% emptied at 1 h) at a tertiary medical center. Patients with a history of esophageal, gastric or thoracic surgery were excluded. Patients were stratified into three cohorts based on 1-h gastric emptying percentages: 30-49%, 50-69%, and ≥ 70%. Manual chart review was performed to extract data on demographics, medications, laboratory values, GES indications, and management changes resulting from GES findings. RESULTS: The majority of patients (n = 205, 79.4%) were in the 30-49% emptying cohort. Only 10 patients (3.9%) met the conventional threshold of ≥ 70% emptying at 1 h. The most common indications for GES were nausea (39.1%), vomiting (33.7%), and abdominal pain (25.6%), with no statistically significant differences in indications across the 3 cohorts. Furthermore, there were no differences in age, sex, BMI, comorbidities, medications, or management changes between the cohorts. CONCLUSIONS: RGE that meets current consensus criteria is uncommon in clinical practice. Clinical characteristics and interventions were similar among cohorts with different degrees of accelerated emptying. The current cutoff of ≥ 70% emptying at 1 h may not represent a clinically distinct phenotype, emphasizing the need for better criteria to guide diagnosis and management.
BACKGROUND AND AIMS: Patients increasingly use the internet and artificial intelligence chatbots to obtain health information, yet the readability, quality, understandability, and actionability of AI-generated gastrointe...BACKGROUND AND AIMS: Patients increasingly use the internet and artificial intelligence chatbots to obtain health information, yet the readability, quality, understandability, and actionability of AI-generated gastrointestinal patient education remain unclear. This study compared gastrointestinal patient education from a professional society website with content generated by ChatGPT using validated health literacy instruments. METHODS: In this cross-sectional comparative study, 50 gastrointestinal patient education topics from the American Gastroenterological Association patient information website were paired with ChatGPT-generated responses using standardized prompts. Readability was assessed using the Flesch-Kincaid Grade Level, quality of treatment information was evaluated using the DISCERN instrument, and understandability and actionability were assessed using the Patient Education Materials Assessment Tool; scoring was performed by two blinded reviewers. Paired t tests were used to compare mean scores between sources, and intraclass correlation coefficients (ICCs) were used to assess interrater reliability between reviewers. RESULTS: Fifty paired topics were analyzed. The mean Flesch-Kincaid Grade Level was higher for ChatGPT than GI website materials (10.33 ± 1.5 vs 8.72 ± 1.7; mean difference, 1.61; P < .001). Differences in DISCERN scores (63.5 ± 5.7 vs 64.3 ± 5.4; mean difference, - 0.8), PEMAT understandability (87.9% ± 6.9% vs 86.5% ± 7.8%; mean difference, 1.4%; P = .33), and PEMAT actionability (78.6% ± 9.8% vs 77.9% ± 10.2%; mean difference, 0.6%; P = .73) were not statistically significant. Inter-rater reliability was excellent across all measures, with intraclass correlation coefficients of 0.97 (95% CI, 0.95-0.99) for PEMAT understandability, 0.96 (95% CI, 0.94-0.98) for PEMAT actionability, and 0.99 (95% CI, 0.98-0.99) for DISCERN. CONCLUSION: ChatGPT-generated gastrointestinal patient education demonstrated similar quality, understandability, and actionability compared with professional society materials but was written at a significantly higher reading level. Improving readability may enhance accessibility and support the safe integration of AI-generated patient education.
PURPOSE: Pancreatic adenosquamous carcinoma (PASC) is a rare and aggressive malignancy with poorer prognosis than pancreatic ductal adenocarcinoma (PDAC). This study aimed to identify distinctive endoscopic ultrasound (E...PURPOSE: Pancreatic adenosquamous carcinoma (PASC) is a rare and aggressive malignancy with poorer prognosis than pancreatic ductal adenocarcinoma (PDAC). This study aimed to identify distinctive endoscopic ultrasound (EUS) features for pre-operative differentiation between PASC and PDAC. METHODS: Forty-six PASC patients and 683 PDAC patients were retrospectively enrolled. Propensity score matching (1:2) balanced age, sex, and tumor size. EUS characteristics were compared, and independent predictors were identified by multivariate logistic regression. RESULTS: PASC tumors were significantly larger than PDAC tumors at baseline (p < 0.001). After matching, multivariate analysis identified five independent predictors of PASC: higher serum albumin (OR = 1.256), hyperechoic foci (OR = 7.733), non-infiltrative growth pattern (INF C: OR = 0.163), and absence of pancreatic duct dilation (positive: OR = 0.021; unknown: 0.008) (all p < 0.05). The model demonstrated good discrimination with an AUC of 0.931 (95% CI 0.888-0.974). PASC more frequently exhibited hyperechoic foci (69.6% vs. 34.8%) and expansive growth, while PDAC typically showed infiltrative growth and duct dilation. CONCLUSION: PASC demonstrates distinctive EUS features including hyperechoic foci, expansive growth pattern, and less frequent ductal obstruction compared with PDAC, which may facilitate pre-operative differentiation from PDAC.
BACKGROUND: Bleeding from gastric varices (GVs) is a severe complication of portal hypertension. Conventional endoscopic cyanoacrylate injection (Con-ECI) is an effective treatment, but it is associated with significant...BACKGROUND: Bleeding from gastric varices (GVs) is a severe complication of portal hypertension. Conventional endoscopic cyanoacrylate injection (Con-ECI) is an effective treatment, but it is associated with significant risks of rebleeding and systemic ectopic embolism. Clip-assisted ECI (Clip-ECI), where clips are deployed before glue injection, has been developed to mitigate these risks, but its efficacy remains debatable. METHODS: A comprehensive search of PubMed, Embase, Web of Science, and Scopus was conducted for studies up to August 2025. The primary outcomes were all-cause rebleeding and ectopic embolism. Secondary outcomes included variceal obliteration, mortality, and procedure-related complications. Risk ratios (RR) with 95% confidence intervals (CI) were pooled from randomized and non-randomized studies using a random-effects model. PROSPERO ID: CRD420251163015. RESULTS: Seven studies involving 783 patients were included. Compared to Con-ECI, Clip-ECI was associated with a significantly lower risk of variceal rebleeding (RR: 0.24; 95% CI [0.12, 0.48]; P < 0.001) and ectopic embolism (RR: 0.31; 95% CI [0.13, 0.76]; P = 0.01). The pooled rates of variceal rebleeding and ectopic embolism in the Clip-ECI group were 7.1% and 2.1%, respectively. Also, Clip-ECI achieved a significantly higher rate of complete variceal obliteration (RR: 1.39; 95% CI [1.23, 1.57]; P < 0.001). There were no significant differences in all-cause mortality (P = 0.28), technical success (P = 0.35), or other adverse events. CONCLUSION: In patients with GVs, Clip-ECI significantly reduces the risks of rebleeding and ectopic embolism while improving variceal obliteration, without increasing adverse events, compared with Con-ECI.
BACKGROUND: Despite varying colorectal cancer (CRC) outcomes across distinct Asian American subgroups, Asian individuals are often assessed in aggregate. We performed disaggregated analyses to assess for heterogeneity in...BACKGROUND: Despite varying colorectal cancer (CRC) outcomes across distinct Asian American subgroups, Asian individuals are often assessed in aggregate. We performed disaggregated analyses to assess for heterogeneity in CRC outcomes across the six largest Asian origin groups in the United States. METHODS: We identified adults with a primary diagnosis of CRC using surveillance, epidemiology, and end results (SEER) registry data from 2006 to 2020. Individuals from the six largest Asian origin groups and a referent non-Hispanic White (NHW) group were included. We identified predictors for regional or distant stage CRC at the time of diagnosis using multivariable logistic regression and predictors for CRC related death using multivariable Cox proportional hazards models. RESULTS: Our cohort included 25,379 Asian Americans, who identified as Chinese (6585 [25.9%]), Filipino (6399 [25.2%]), Japanese (4309 [17.0%]), Vietnamese (3047 [12.0%]), Korean (2964 [11.7%]), and Asian Indian/Pakistani (2075 [8.2%]), in addition to 234,938 NHW individuals. In our multivariable logistic regression, Vietnamese (aOR 1.23; 95% CI 1.14-1.32), Korean (aOR 1.20; 95% CI 1.11-1.30), and Asian Indian/Pakistani (aOR 1.10; 95% CI 1.01-1.21) individuals had increased odds of regional or distant stage at time of diagnosis compared to NHW individuals. In our multivariable Cox model, Chinese (aHR 0.92; 95% CI 0.88-0.97) and Asian Indian/Pakistani (aHR 0.88; 95% CI 0.80-0.97) individuals had decreased likelihood of CRC related death compared to NHW individuals. Lower neighborhood-level socioeconomic status by Yost Index was associated with worse CRC outcomes in both multivariable models. CONCLUSIONS: In this national cancer registry analysis, marked heterogeneity was observed in CRC stage and survival across disaggregated Asian American groups. Disaggregated analyses are key to accurately survey CRC trends, identify groups at highest risk, and develop culturally relevant approaches to CRC prevention and care.
OBJECTIVES: Mounting evidence suggests that ferroptosis is closely involved in the development of severe acute pancreatitis (SAP). This study aimed to investigate the association between abnormal lipid metabolism and fer...OBJECTIVES: Mounting evidence suggests that ferroptosis is closely involved in the development of severe acute pancreatitis (SAP). This study aimed to investigate the association between abnormal lipid metabolism and ferroptosis and elucidate the role of the Keap1/Nrf2/SLC7A11/GPX4 pathway in pancreatic acinar cell injury during SAP. METHODS: Rats were divided into four experimental groups: sham, hyperlipidemia (HL), SAP, and HL-SAP. Sprague-Dawley rats were adopted to establish HL-SAP model through administering high-fat emulsions via gastric infusion for 14 consecutive days and sodium taurocholic injection. Measure serum amylase, blood lipid, and inflammatory cytokine levels, perform histological analysis, determine the expression levels of proteins in the Keap1/Nrf2/SLC7A11/GPX4 signaling pathway, and evaluate ferroptosis-related changes and oxidative stress. RESULTS: High-fat emulsion feeding successfully induced hyperlipidemia with elevated blood lipids, while injection of 3.5% sodium taurocholate triggered SAP accompanied by increased serum amylase. The combined intervention effectively established the HL-SAP model. Compared with the SAP rats, the HL-SAP rats exhibited more severe pancreatic damage (72 h mortality: 80 vs. 50%, respectively, plus elevated amylase, inflammation, histopathology scores). Higher amylase levels, intensified inflammation, and increased histopathological scores were also observed in HL-SAP rats. Moreover, HL-SAP rats showed markedly enhanced oxidative stress and ferroptosis-related phenotypes, including increased MDA, ROS and Fe⁺ levels, as well as decreased GSH and SOD levels. In both SAP and HL-SAP groups, abnormal lipid metabolism was associated with downregulated Nrf2, SLC7A11, and GPX4 expression and upregulated Keap1 expression, and these alterations were more prominent in the HL-SAP group. CONCLUSION: Abnormal lipid metabolism aggravates oxidative stress, inflammation, and pancreatic acinar cell injury in SAP, and promotes ferroptosis. These effects may be attributed to reduced activity of the Keap1/Nrf2/SLC7A11/GPX4 pathway.
BACKGROUND: Polycystic liver disease and other manifestations of ductal plate malformation (DPM) may be incidental or indeed develop significant complications. The genetic basis of 'isolated' DPM differs from that associ...BACKGROUND: Polycystic liver disease and other manifestations of ductal plate malformation (DPM) may be incidental or indeed develop significant complications. The genetic basis of 'isolated' DPM differs from that associated with polycystic kidney disease and remains incompletely understood. AIMS: We sought to genetically characterize a sizeable single-center cohort of patients with 'isolated' DPM, defined as absence of polycystic kidney disease. METHODS: 55 consecutive patients with isolated DPM were analyzed retrospectively with next-generation sequencing used to identify genetic variants. RESULTS: 54/55 completed genetic characterization. 40 patients had polycystic liver disease (PCLD) as the primary phenotype, 9 had congenital hepatic fibrosis (CHF), 1 had Caroli disease and 5 had multiple biliary hamartomas. 10 patients had mixed phenotype. The majority were female (78%) and mean age at presentation was 47 years. Of patients with PCLD, previous estrogen exposure was associated with cyst complications. Of the 54 genetically analyzed patients, pathogenic or likely pathogenic variants were identified in 31. PCLD was mainly associated with SEC63, PRKCSH and GANAB heterozygous variants with a minor contribution from PKHD1 and ALG8. PRKCSH followed by GANAB was associated with the most severe PCLD presentation. CHF was represented solely by PKHD1 variants half of which were missense, as opposed to truncating in all but one of the other genotypes. Importantly, 5 novel variants were identified (4 SEC63, 1 PRKCSH). CONCLUSIONS: We have shown that current genetic testing methods are often revealing in isolated DPM. Future broadening of genetic analysis will enhance diagnosis and understanding within this heterogeneous disease spectrum.
BACKGROUND: Pancreatic cancer incidence rises in low- and middle-income countries (LMICs), yet data on early-onset disease (20-54 years) are scarce. METHODS: Using Global Burden of Disease Study 2023 data (1990-2023), we...BACKGROUND: Pancreatic cancer incidence rises in low- and middle-income countries (LMICs), yet data on early-onset disease (20-54 years) are scarce. METHODS: Using Global Burden of Disease Study 2023 data (1990-2023), we analyzed incidence, mortality, and disability-adjusted life years (DALYs) for early-onset pancreatic cancer in LMIC adults. Trends were assessed via estimated annual percentage change (EAPC), stratified by World Bank income groups. Decomposition analysis quantified epidemiological/demographic drivers. Socioeconomic inequalities were evaluated using socio-demographic index (SDI); projections to 2050 employed autoregressive integrated moving average models. RESULTS: From 1990 to 2023, LMIC early-onset pancreatic cancer cases and deaths increased by 125% and 128%, respectively. Age-standardized incidence (ASIR EAPC: 0.68%) and mortality rates rose. Lower-middle-income countries showed the steepest incidence surge (+ 282%; ASIR EAPC: 1.65%). Young women (20-34 years) in low/lower-middle-income countries bore higher burden, contrasting higher-income regions. Burden correlated positively with national income and exhibited an S-shaped SDI association. Epidemiological changes drove increases in lower-income settings. Socioeconomic inequalities widened; age-standardized rates are projected to rise through 2050. CONCLUSION: Early-onset pancreatic cancer burden is escalating across LMICs, with distinct patterns by income level. Young women in lower-income settings are specifically vulnerable. Targeted prevention and resource allocation are urgently needed.
PURPOSE: Intestinal ultrasound (IUS) has been shown to be an accurate, non-invasive, point-of-care tool for monitoring disease activity in inflammatory bowel disease (IBD). While its diagnostic performance is well-establ...PURPOSE: Intestinal ultrasound (IUS) has been shown to be an accurate, non-invasive, point-of-care tool for monitoring disease activity in inflammatory bowel disease (IBD). While its diagnostic performance is well-established, there is a notable gap in research related to patient perspectives regarding the role of IUS in IBD management, including its impact on the patient-provider relationship, which has not been previously studied. METHODS: We conducted a quantitative study with a 21-question Likert scale (0-10) survey administered to patients undergoing IUS during routine IBD clinic visits. Survey questions assessed four domains: (1) patient experience with IUS, (2) preference of IUS versus other diagnostic modalities, (3) impact on the patient-provider relationship, and (4) the enhancement of disease understanding. RESULTS: Fifty-seven patients completed the survey. Overall, patients reported high satisfaction with IUS (mean score: 9.15), preferred IUS over bloodwork, stool testing, CT/MRI, and colonoscopy (mean score: 7.64), and felt it positively impacted patient-provider relationship (mean score: 8.07), and disease understanding (mean score: 7.74). IUS was strongly preferred over stool testing and colonoscopy (mean scores: 10 for both). Patients on steroids reported greater enhancement of the patient-provider relationship (8.70 vs. 7.82, p = 0.03), and those with Crohn's disease reported greater improvement in disease understanding compared to ulcerative colitis (8.16 vs. 7.18, p = 0.04). CONCLUSIONS: Patients expressed high satisfaction with IUS across all survey domains. These findings support the broader use of IUS into IBD care, highlighting its value not only as a diagnostic tool but also in enhancing patient experience and engagement.
Herran R, Wilcox S, Hodish G
… +12 more, Hassan SA, Mertens H, Almario CV, Gu P, Kayal M, Cohen-Mekelburg S, Regal RE, Bishu S, Siegel CA, DeJonckheere M, Higgins PDR, Berinstein JA
PURPOSE: Acute severe ulcerative colitis (ASUC) is a life-threatening medical emergency necessitating immediate hospitalization and rapid initiation of therapy. For decades, intravenous (IV) corticosteroids have remained...PURPOSE: Acute severe ulcerative colitis (ASUC) is a life-threatening medical emergency necessitating immediate hospitalization and rapid initiation of therapy. For decades, intravenous (IV) corticosteroids have remained first-line, with rescue therapies reserved for patients who are corticosteroid-refractory. However, the therapeutic landscape has evolved, shifting toward earlier initiation of advanced therapy. This study analyzes factors influencing patient decision-making for treatment of ASUC. METHODS: In this prospective, mixed-methods study, patients with ASUC who were infliximab and Janus kinase inhibitor naïve were enrolled. Participants were presented with educational resources on IV corticosteroids, upadacitinib, and infliximab and then asked to choose between continuing IV corticosteroids alone or starting early initiation of advanced therapy in addition to IV corticosteroids. Factors influencing their preferences were captured via semi-structured interviews, and clinical outcomes were followed. RESULTS: Eleven patients were prospectively enrolled. Analysis of semi-structured interviews revealed five dominant themes influencing treatment decision-making: (1) reliance on physician expertise, (2) contextualized treatment selection, (3) concrete risk assessment, (4) empathetic individualization, and (5) strong preference for oral therapy. Notably, we observed significant discordance between patient treatment preferences and medication initiated following discussion between the participant and their primary clinical team, particularly among participants favoring upadacitinib. CONCLUSION: These five principal themes offer actionable opportunities to assist patients in making well-informed decisions regarding ASUC therapy. Clinicians should utilize these insights to develop evidence-based patient decision aids that specifically address patient priorities. While patients often prioritize oral therapy, their choices often diverge from the therapy initiated after discussing with their primary clinical team. Future efforts should focus on bridging this discordance by integrating patient priorities with available evidence into the clinical decision-making process.
BACKGROUND: The immunosuppressive tumor microenvironment in hepatocellular carcinoma (HCC) limits therapeutic efficacy. This study aimed to develop an immune-related signature for risk stratification and to identify key...BACKGROUND: The immunosuppressive tumor microenvironment in hepatocellular carcinoma (HCC) limits therapeutic efficacy. This study aimed to develop an immune-related signature for risk stratification and to identify key molecules that drive immune evasion. METHODS: Transcriptomic data from HCC were analyzed to identify immune-related genes. A prognostic risk signature was constructed using Cox and least absolute shrinkage and selection operator regression and validated in an independent external cohort. Immune cell infiltration, immune checkpoints, and interleukin (IL)-17RA expression were characterized. Serum IL-17 levels were measured in patients with HCC, patients with cirrhosis, and healthy controls. Mechanistic studies included analyses of cell proliferation, programmed death-ligand 1 (PD-L1) expression, nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) activation, and the effects of IL-17RA knockdown. RESULTS: A six-gene risk signature stratified patients into high- and low-risk groups. The high-risk group exhibited poorer prognosis and a more immunosuppressive microenvironment. External validation confirmed model robustness, with the risk score as an independent prognostic factor (multivariable Cox, P < 0.05). IL-17RA was overexpressed in HCC and correlated with poor prognosis, immune checkpoints, and M2 macrophage infiltration. Serum IL-17 levels were elevated in patients with HCC, had diagnostic value (area under the curve [AUC] = 0.738), and predicted vascular invasion (AUC = 0.891). Mechanistically, IL-17 activated NF-κB, increased p-p65, induced p65 nuclear translocation, upregulated PD-L1, and stimulated HCC cell proliferation. Pharmacological NF-κB inhibition or genetic silencing of IL-17RA abrogated IL-17-induced PD-L1 upregulation. CONCLUSION: The six-gene prognostic signature and external validation suggest clinical utility. The IL-17/ IL-17 receptor alpha axis may drive immune escape by upregulating PD-L1 via NF-κB activation; therefore, it represents a potential prognostic biomarker and therapeutic target.
PURPOSE: Risk stratification for advanced colorectal polyps typically relies on colonoscopy and/or pathology findings. However, there is growing interest in whether noninvasive features available prior to colonoscopy can...PURPOSE: Risk stratification for advanced colorectal polyps typically relies on colonoscopy and/or pathology findings. However, there is growing interest in whether noninvasive features available prior to colonoscopy can help identify patients at higher risk. Such approaches may enhance clinical decision-making by prioritizing surveillance for individuals most likely to harbor high-risk polyps, when colonoscopy resources are limited while potentially reducing unnecessary procedures in lower-risk patients. Importantly, the use of noninvasive, pre-procedural information may also help promote more equitable access to risk stratification, particularly in settings where colonoscopy resources are limited or unevenly distributed. We aimed to develop and externally validate machine learning models to predict high-risk colorectal polyps using only noninvasive, pre-colonoscopy demographic, clinical, and behavioral features in a diverse, predominantly African American, urban cohort. METHODS: We conducted a retrospective cohort study using demographic, lifestyle, and comorbidity data from patients who underwent colonoscopy at Howard University Hospital to develop and validate several machine learning models, including neural networks, random forest, support vector machines (SVM), Naïve Bayes, logistic regression, decision trees, k-nearest neighbors (KNN), and XGBoost, for predicting high-risk colorectal polyps. High-risk polyps (HRP) were defined as villous or tubullovillous adenomas, high-grade dysplasia, polyps 10 mm in size, and/or the presence of 3 polyps per procedure; all other cases were classified as low-risk polyps (LRP). The dataset included 4,681 patients from 2015 to 2022 used for internal validation and 1,562 patients from 2023 to 2024 used for external validation. Model performance was evaluated using the area under the receiver operating characteristic curve (ROC-AUC), precision-recall area under the curve (PR-AUC), accuracy, precision, recall, and F1 score. Model interpretability and feature contribution were assessed using SHapley Additive exPlanations (SHAP). RESULTS: Overall predictive performance was moderate using noninvasive pre-colonoscopy features. The neural network demonstrated the strongest overall discrimination, achieving the highest internal validation performance (ROC-AUC 0.78, PR-AUC 0.75, accuracy 0.72), but showed reduced performance in the external cohort (ROC-AUC 0.67, accuracy 0.66), suggesting potential overfitting or temporal feature drift. In contrast, simpler models including Naïve Bayes, SVM, and XGBoost exhibited lower internal performance (ROC-AUC 0.54-0.59) but more stable generalization to the external cohort (ROC-AUC 0.52-0.63; accuracy approximately 0.53-0.60). Model interpretability analysis using SHAP identified age, smoking status, sex, occupation, race, colonoscopy indication, and family history of colorectal cancer as the most influential predictors, highlighting contributions from both traditional clinical and sociodemographic factors. CONCLUSIONS: Prediction of HRP using routine pre-colonoscopy data is feasible but demonstrates limited generalizability across cohorts. These findings highlight the clinical potential and limitations of pre-procedural risk modeling, especially in diverse, underserved populations. Integration of additional data modalities may be required to achieve clinically robust and equitable prediction tools.