Sepsis-Associated Liver Injury (SALI) is an independent risk factor for death from sepsis. The aim of this study was to develop an interpretable machine learning model for early prediction of 28-day mortality in patients with SALI. Data from the Medical Information Mart for Intensive Care (MIMIC-IV, v2.2, MIMIC-III, v1.4) were used in this study. The study cohort from MIMIC-IV was randomized to the training set (0.7) and the internal validation set (0.3), with MIMIC-III (2001 to 2008) as external validation. The features with more than 20% missing values were deleted and the remaining features were multiple interpolated. Lasso-CV that lasso linear model with iterative fitting along a regularization path in which the best model is selected by cross-validation was used to select important features for model development. Eight machine learning models including Random Forest (RF), Logistic Regression, Decision Tree, Extreme Gradient Boost (XGBoost), K Nearest Neighbor, Support Vector Machine, Generalized Linear Models in which the best model is selected by cross-validation (CV_glmnet), and Linear Discriminant Analysis (LDA) were developed. Shapley additive interpretation (SHAP) was used to improve the interpretability of the optimal model. At last, a total of 1043 patients were included, of whom 710 were from MIMIC-IV and 333 from MIMIC-III. Twenty-four clinically relevant parameters were selected for model construction. For the prediction of 28-day mortality of SALI in the internal validation set, the area under the curve (AUC (95% CI)) of RF was 0.79 (95% CI: 0.73–0.86), and which performed the best. Compared with the traditional disease severity scores including Oxford Acute Severity of Illness Score (OASIS), Sequential Organ Failure Assessment (SOFA), Simplified Acute Physiology Score II (SAPS II), Logistic Organ Dysfunction Score (LODS), Systemic Inflammatory Response Syndrome (SIRS), and Acute Physiology Score III (APS III), RF also had the best performance. SHAP analysis found that Urine output, Charlson Comorbidity Index (CCI), minimal Glasgow Coma Scale (GCS_min), blood urea nitrogen (BUN) and admission_age were the five most important features affecting RF model. Therefore, RF has good predictive ability for 28-day mortality prediction in SALI. Urine output, CCI, GCS_min, BUN and age at admission(admission_age) within 24 h after intensive care unit(ICU) admission contribute significantly to model prediction.

Pentagalloyl glucose (PGG) is a natural hydrolyzable gallotannin abundant in various plants and herbs. It has a broad range of biological activities, specifically anticancer activities, and numerous molecular targets. Despite multiple studies available on the pharmacological action of PGG, the molecular mechanisms underlying the anticancer effects of PGG are unclear. Here, we have critically reviewed the natural sources of PGG, its anticancer properties, and underlying mechanisms of action. We found that multiple natural sources of PGG are available, and the existing production technology is sufficient to produce large quantities of the required product. Three plants (or their parts) with maximum PGG content were Rhus chinensis Mill, Bouea macrophylla seed, and Mangifera indica kernel. PGG acts on multiple molecular targets and signaling pathways associated with the hallmarks of cancer to inhibit growth, angiogenesis, and metastasis of several cancers. Moreover, PGG can enhance the efficacy of chemotherapy and radiotherapy by modulating various cancer-associated pathways. Therefore, PGG can be used for treating different human cancers; nevertheless, the data on the pharmacokinetics and safety profile of PGG are limited, and further studies are essential to define the clinical use of PGG in cancer therapies.

Gastric cancer is one of the most serious malignant tumor and threatens the health of people worldwide. Its heterogeneity leaves many clinical problems unsolved. To treat it effectively, we need to explore its heterogeneity. Single-cell transcriptome sequencing, or single-cell RNA sequencing (scRNA-seq), reveals the complex biological composition and molecular characteristics of gastric cancer at the level of individual cells, which provides a new perspective for understanding the heterogeneity of gastric cancer. In this review, we first introduce the current procedure of scRNA-seq, and discuss the advantages and limitations of scRNA-seq. We then elaborate on the research carried out with scRNA-seq in gastric cancer in recent years, and describe how it reveals cell heterogeneity, the tumor microenvironment, oncogenesis and metastasis, as well as drug response in to gastric cancer, to facilitate early diagnosis, individualized therapy, and prognosis evaluation.

Colorectal cancer (CRC) is a common and deadly cancer worldwide with a high lethality rate. Disulfidptosis has been found to be an emerging mode of death in cancer, and the purpose of this study was to explore the relationship between disulfidptosis-related lncRNAs (DRLs) and CRC and to develop a prognostic model for CRC and DRLs. The gene expression data and clinicopathologic information of colorectal cancer patients were obtained from The Cancer Genome Atlas (TCGA) and screened for DRLs based on correlation analysis. The least absolute shrinkage and selection operator (LASSO) and Cox regression were used to construct the prognostic model, and its validation was carried out by PCA and receiver operating characteristic (ROC) curves. We constructed nomograms combined with the model. Finally, the possible mechanisms by which lncRNAs affect CRC were explored by functional enrichment analysis, immune infiltration and immune escape analysis. In summary, we developed a prognostic marker consisting of lncRNAs associated with disulfidptosis to help clinicians predict the survival of different CRC patients and use different targeted therapies and immunotherapies depending on the condition.

Inflammation can influence the development of CRC as well as immunotherapy and plays a key role in CRC. Therefore, this study aimed to investigate the potential of inflammation-related genes in CRC risk prediction. Inflammation gene models were constructed and validated by combining transcriptomic and single-cell data from TCGA and GEO databases, and the expression of inflammation-related genes was verified by RT-qPCR. We identified two molecular subtypes and three genetic subtypes, two risk subgroups according to median risk values, constructed a prognostic model including thirteen genes (TIMP1, GDF15, UCN, KRT4, POU4F1, NXPH1, SIX2, NPC1L1, KLK12, IGFL1, FOXD1, ASPG, and CYP4F8), and validated the performance of each aspect of the model in an external database. Patients in the high-risk group had worse survival with reduced immune cell infiltration and a greater tumor mutational load. The risk score correlated strongly with the immune checkpoints PD1, PDL1, PDL2, and CTLA4, and it is possible that high-risk patients are more sensitive to treatment involving immune checkpoints. In the single-cell data, GDF15 was most significantly expressed in cancer cell populations. Therefore, we further validated their expression in cells and tissues using qPCR. In summary, we developed a prognostic marker associated with inflammatory genes to provide new directions for subsequent studies and to help clinicians assess the prognosis of CRC patients as well as to develop personalized treatment strategies.