Dataset Information

Machine Learning for Building Immune Genetic Model in Hepatocellular Carcinoma Patients.

ABSTRACT:

Background

Hepatocellular carcinoma (HCC) is the leading liver cancer with special immune microenvironment, which played vital roles in tumor relapse and poor drug responses. In this study, we aimed to explore the prognostic immune signatures in HCC and tried to construct an immune-risk model for patient evaluation.

Methods

RNA sequencing profiles of HCC patients were collected from the cancer genome Atlas (TCGA), international cancer genome consortium (ICGC), and gene expression omnibus (GEO) databases (GSE14520). Differentially expressed immune genes, derived from ImmPort database and MSigDB signaling pathway lists, between tumor and normal tissues were analyzed with Limma package in R environment. Univariate Cox regression was performed to find survival-related immune genes in TCGA dataset, and in further random forest algorithm analysis, significantly changed immune genes were used to generate a multivariate Cox model to calculate the corresponding immune-risk score. The model was examined in the other two datasets with recipient operation curve (ROC) and survival analysis. Risk effects of immune-risk score and clinical characteristics of patients were individually evaluated, and significant factors were then used to generate a nomogram.

Results

There were 52 downregulated and 259 upregulated immune genes between tumor and relatively normal tissues, and the final immune-risk model (based on SPP1, BRD8, NDRG1, KITLG, HSPA4, TRAF3, ITGAV and MAP4K2) can better differentiate patients into high and low immune-risk subpopulations, in which high score patients showed worse outcomes after resection (p < 0.05). The differentially enriched pathways between the two groups were mainly about cell proliferation and cytokine production, and calculated immune-risk score was also highly correlated with immune infiltration levels. The nomogram, constructed with immune-risk score and tumor stages, showed high accuracy and clinical benefits in prediction of 1-, 3- and 5-year overall survival, which is useful in clinical practice.

Conclusion

The immune-risk model, based on expression of SPP1, BRD8, NDRG1, KITLG, HSPA4, TRAF3, ITGAV, and MAP4K2, can better differentiate patients into high and low immune-risk groups. Combined nomogram, using immune-risk score and tumor stages, could make accurate prediction of 1-, 3- and 5-year survival in HCC patients.

SUBMITTER: Liu J

PROVIDER: S-EPMC7994091 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:ObjectiveTo investigate the value of differential diagnosis of hepatocellular carcinoma (HCC) and non-hepatocellular carcinoma (non-HCC) based on CT and MR multiphase radiomics combined with different machine learning models and compare the diagnostic efficacy between different radiomics models.BackgroundPrimary liver cancer is one of the most common clinical malignancies, hepatocellular carcinoma (HCC) is the most common subtype of primary liver cancer, accounting for approximately 90% of cases. A clear diagnosis of HCC is important for the individualized treatment of patients with HCC. However, more sophisticated diagnostic modalities need to be explored.MethodsThis retrospective study included 211 patients with liver lesions: 97 HCC and 124 non-hepatocellular carcinoma (non-HCC) who underwent CT and MRI. Imaging data were used to obtain imaging features of lesions and radiomics regions of interest (ROI). The extracted imaging features were combined to construct different radiomics models. The clinical data and imaging features were then combined with radiomics features to construct the combined models. Support Vector Machine (SVM), K-nearest Neighbor (KNN), RandomForest (RF), eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), Multilayer Perceptron (MLP) six machine learning models were used for training. Five-fold cross-validation was used to train the models, and ROC curves were used to analyze the diagnostic efficacy of each model and calculate the accuracy rate. Model training and efficacy test were performed as before.ResultsStatistical analysis showed that some clinical data (gender and concomitant cirrhosis) and imaging features (presence of envelope, marked enhancement in the arterial phase, rapid contouring in the portal phase, uniform density/signal and concomitant steatosis) were statistical differences (P < 0.001). The results of machine learning models showed that KNN had the best diagnostic efficacy. The results of the combined model showed that SVM had the best diagnostic efficacy, indicating that the combined model (accuracy 0.824) had better diagnostic efficacy than the radiomics-only model.ConclusionsOur results demonstrate that the radiomic features of CT and MRI combined with machine learning models enable differential diagnosis of HCC and non-HCC (malignant, benign). The diagnostic model with dual radiomic had better diagnostic efficacy. The combined model was superior to the radiomic model alone.