Background: Lenvatinib and transarterial chemoembolization (TACE) are first-line treatments for unresectable hepatocellular carcinoma (HCC), but the objective response rate (ORR) is not satisfactory. We aimed to predict the response to lenvatinib combined with TACE before treatment for unresectable HCC using machine learning (ML) algorithms based on clinical data. Methods: Patients with unresectable HCC receiving the combination therapy of lenvatinib combined with TACE from two medical centers were retrospectively collected from January 2020 to December 2021. The response to the combination therapy was evaluated over the following 4–12 weeks. Five types of ML algorithms were applied to develop the predictive models, including classification and regression tree (CART), adaptive boosting (AdaBoost), extreme gradient boosting (XGBoost), random forest (RF), and support vector machine (SVM). The performance of the models was assessed by the receiver operating characteristic (ROC) curve and area under the receiver operating characteristic curve (AUC). The Shapley Additive exPlanation (SHAP) method was applied to explain the model. Results: A total of 125 unresectable HCC patients were included in the analysis after the inclusion and exclusion criteria, among which 42 (33.6%) patients showed progression disease (PD), 49 (39.2%) showed stable disease (SD), and 34 (27.2%) achieved partial response (PR). The nonresponse group (PD + SD) included 91 patients, while the response group (PR) included 34 patients. The top 40 most important features from all 64 clinical features were selected using the recursive feature elimination (RFE) algorithm to develop the predictive models. The predictive power was satisfactory, with AUCs of 0.74 to 0.91. The SVM model and RF model showed the highest accuracy (86.5%), and the RF model showed the largest AUC (0.91, 95% confidence interval (CI): 0.61–0.95). The SHAP summary plot and decision plot illustrated the impact of the top 40 features on the efficacy of the combination therapy, and the SHAP force plot successfully predicted the efficacy at the individualized level. Conclusions: A new predictive model based on clinical data was developed using ML algorithms, which showed favorable performance in predicting the response to lenvatinib combined with TACE for unresectable HCC. Combining ML with SHAP could provide an explicit explanation of the efficacy prediction.

Even after curative resection, the prognosis for patients with intrahepatic cholangiocarcinoma (iCCA) remains disappointing due to the extremely high incidence of postoperative recurrence. A total of 280 iCCA patients following curative hepatectomy from three independent institutions were recruited to establish the retrospective multicenter cohort study. The very early recurrence (VER) of iCCA was defined as the appearance of recurrence within 6 months. The 3D tumor region of interest (ROI) derived from contrast-enhanced CT (CECT) was used for radiomics analysis. The independent clinical predictors for VER were histological stage, AJCC stage, and CA199 levels. We implemented K-means clustering algorithm to investigate novel radiomics-based subtypes of iCCA. Six types of machine learning (ML) algorithms were performed for VER prediction, including logistic, random forest (RF), neural network, bayes, support vector machine (SVM), and eXtreme Gradient Boosting (XGBoost). Additionally, six clinical ML (CML) models and six radiomics-clinical ML (RCML) models were developed to predict VER. Predictive performance was internally validated by 10-fold cross-validation in the training cohort, and further evaluated in the external validation cohort. Approximately 30 % of patients with iCCA experienced VER with extremely discouraging outcome (Hazard ratio (HR) = 5.77, 95 % Confidence Interval (CI) = 3.73-8.93, P < 0.001). Two distinct iCCA subtypes based on radiomics features were identified, and subtype 2 harbored a higher proportion of VER (47.62 % Vs 25.53 %) and significant shorter survival time than subtype 1. The average AUC values of the CML and RCML models were 0.744 ± 0.018, and 0.900 ± 0.014 in the training cohort, and 0.769 ± 0.065 and 0.929 ± 0.027 in the external validation cohort, respectively. Two radiomics-based iCCA subtypes were identified, and six RCML models were developed to predict VER of iCCA, which can be used as valid tools to guide individualized management in clinical practice.

Background & Aims: Glucose-lowering drug is associated with various cancers, but the causality with gastrointestinal cancer risk is rarely reported. We aimed to explore the causality between them in this Mendelian randomization (MR) study. Methods: Two-sample MR, summary-data-based (SMR), mediation MR, and colocalization analyses was employed. Ten glucose-lowering drug targets (PPARG, DPP4, GLP1R, INSR, SLC5A2, ABCC8, KCNJ11, ETFDH, GPD2, PRKAB1) and seven types of gastrointestinal cancer (anal carcinoma, cardia cancer, gastric cancer, hepatocellular carcinoma (HCC), intrahepatic cholangiocarcinoma (ICC), pancreatic cancer, rectum cancer) were included. Patients with gastrointestinal cancers from six different large GWAS databases, including the UK Biobank and Finnish cohorts were incorporated, for discovery and external validation. Meta-analysis was employed to integrate the results from both discovery and validation cohorts, thereby ensuring the reliability of findings. Results: ABCC8/KCNJ11 were associated with pancreatic cancer risk in both two-sample MR (odds ratio (OR): 15.058, per standard deviation unit (SD) change of glucose-lowering durg target perturbation equivalent to 1 SD unit of HbA_1c lowering; 95% confidence interval (95% CI): 3.824–59.295; P-value = 0.0001) and SMR (OR: 1.142; 95% CI: 1.013–1.287; P-value = 0.030) analyses. The mediation effect of body mass index (OR: 0.938; 95% CI: 0.884–0.995; proportion of mediation effect: 3.001%; P-value = 0.033) on ABCC8/KCNJ11 and pancreatic cancer was uncovered. Strong connections of DPP4 with anal carcinoma (OR: 0.123; 95% CI: 0.020–0.745; P-value = 0.023) and ICC (OR: 7.733; 95% CI: 1.743–34.310; P-value = 0.007) were detected. PPARG was associated with anal carcinoma (OR: 12.909; 95% CI: 3.217–51.795; P-value = 0.0003), HCC (OR: 36.507; 95% CI: 8.929-149.259; P-value < 0.0001), and pancreatic cancer (OR: 0.110; 95% CI: 0.071–0.172; P-value < 0.0001). SLC5A2 was connected with pancreatic cancer (OR: 8.096; 95% CI: 3.476–18.857; P-value < 0.0001). Weak evidence indicated the connections of GLP1R, GPD2, and PRKAB1 with anal carcinoma, cardia cancer, ICC, and rectum cancer. In addition, the corresponding results were consistently validated in both the validation cohorts and the integrated outcomes. Conclusions: Some glucose-lowering drugs were associated with gastrointestinal cancer risk, which might provide new ideas for gastrointestinal cancer treatment.

Background: Although sarcopenia has been reported as a negative prognostic factor in patients with hepatocellular carcinoma (HCC), the lack of studies with a prospective design utilizing comprehensive sarcopenia assessment with composite endpoints is an important gap in understanding the impact of sarcopenia in patients with HCC. The aim of this study was to investigate the relationship between sarcopenia and postoperative 1-year mortality and health-related quality of life (HRQOL) based on sarcopenia assessment. Methods: The study cohort, who received resection surgery for HCC between May 2020 and August 2021, was assessed for sarcopenia based on grip strength, the chair stand test, skeletal muscle mass, and gait speed. The primary outcome measures were 1-year mortality and HRQOL determined using the QLQ-C30 questionnaire. In addition, we collected hospital costs, postoperative hospital stays, complications, 30-day and 90-day mortality, and 90- and 180-day readmission rates. Univariate and multivariate linear regression analyses were conducted to examine factors associated with global health status. Results: A total of 153 eligible patients were included in the cohort. One-year mortality was higher in patients with sarcopenia than in those without sarcopenia (p = 0.043). There was a correlation between sarcopenia and the surgical approach to global health status (p = 0.025) and diarrhea (p = 0.003). Conclusions: Preoperative sarcopenia reduces postoperative survival and health-related quality of life in patients with HCC.

Sarcopenia is associated with adverse prognosis of intrahepatic cholangiocarcinoma (iCCA) after surgery. 321 patients with iCCA undergoing surgery were retrospectively recruited and assigned to training and validation cohort. Skeletal muscle index (SMI) was assessed to define sarcopenia. Logistic regression and cox regression analysis were used to identify risk factors. A novel sarcopenia-based nomogram was constructed and validated by ROC curves, calibration curves, and DCA curves. 260 patients were included for analysis. The median age was 63.0 years and 161 patients (61.9%) were diagnosed with sarcopenia. Patients with sarcopenia exhibited a higher rate of postoperative complications, a worse OS and RFS than patients without sarcopenia. Sarcopenia, low albumin and intraoperative blood transfusion were independent risk factors of postoperative complications, while sarcopenia and low albumin were risk factors of high CCI≥26.2. Sarcopenia, high PS score, low-undifferentiated differentiation, perineural invasion, TNM stage III-IV were risk factors of OS, and a novel nomogram based on these five factors was built to predict the 12-, 24-, and 36-months OS, with the mean AUC > 0.6. Sarcopenia is negatively associated with both postoperative complications and survival prognosis of iCCA undergoing hepatectomy.

Despite a rising trend in intrahepatic cholangiocarcinoma (ICC) incidence in the elderly population worldwide, the benefit of surgery for those patients is still controversial. Data from 811 elderly patients diagnosed with non-metastatic ICC were obtained from the US surveillance, epidemiology, and end results (SEER) program database. Propensity score matched (PSM) was conducted for the better balance of baseline. The associations between tumor characteristics and surgery with overall survival (OS) and cancer specific survival (CSS) were estimated using hazard ratios (HR) and 95% confidence intervals (CI). The results showed that ICC patients above 60 years old taking surgery had better OS (hazard ratio [HR], 0.258; 95% CI, 0.205–0.324) and CSS (hazard ratio [HR], 0.239; 95% CI, 0.188–0.303) than patients without surgery. Similar trends in patients above 65 years old, above 70 years old, above 75 years old, and above 80 years old were observed, separately. This benefit was also showed in lymph node-negative (N0) and lymph node-positive (N1) subgroups and N0 patients are more likely to take an advantage from surgery than N1 patients. The different outcomes between surgery and non-surgery suggest that surgical treatment may be recommended for elderly ICC if the tumor is resectable to ensure optimal treatment.