Establishment Of A Prognosis Nomogram For Hepatoblastomas And Identification Of Potential Biomarkers And Immune Characteristics By Bioinformatics Methods

Objective:Microvascular invasion may be a prognosis factor for hepatoblastomas(HBs).This study aimed to construct a nomogram model to predict the survival outcome in HBs.Methods:We retrospectively analyzed 331 HBs who underwent liver resection at our institution from January 2014 to October 2021.First,we applied propensity score matching(PSM)to balance the clinical characteristics between the positive and negative groups for microvascular invasion in a 1:1 ratio,and compared the survival prognosis of children in both groups.Second,all patients were divided into training and validation groups in a 7:3 ratio.The Fine-Gray test was used to find variables with significant differences to identify risk factors for final inclusion in the multivariate model.After that,the competing risk model was constructed for multivariate analysis to identify high-risk factors and establish nomogram to predict the probability of recurrence-free survival(RFS)at 1,2 and 3 years.ROC and calibration curves were used to validate the predictive power of the model.Clinical utility assessment was then evaluated by plotting decision curves.We also validated two risk staging approaches to assess the prognosis of HBs,including the CHIC-HS as well as the SIOPEL staging model.Results:Competing risk model showed that serum AFP level,multifocality,distant metastasis,macrovascular involvement,histology type and microvascular invasion were independent risk factors affecting recurrence in HBs.The AUC values of nomogram for predicting RFS at 1,2,and 3 years were 0.803,0.895,0.910,and 0.923,0.953,0.963 in the training and validation groups,respectively.The results showed that the AUC of the nomogram was superior to that of CHIC-HS and SIOPEL staging.The calibration curves showed that the predicted probabilities were in good agreement with the actual probabilities.The decision curve analysis showed the nomogram had good clinical utility.Conclusion:First,we demonstrated that positive microscopic mcrovascular invasion is an independent risk factor for poor prognosis in HBs.Second,by combining imaging results and other clinical data,we established a nomogram to predict RFS probability for patients with HB,which could be a potential tool to guide personalized treatment.Objective:This study aimed to investigate the immune characteristics of hepatoblastoma(HB)and identify potential biomarkers for its diagnosis.Methods:First,three datasets(GSE132037,GSE51701,GSE131329)of HB were downloaded from Gene Expression Omnibus(GEO)database.Differentially expressed genes(DEGs)between HB and normal liver samples were identified using the "limma"package in R software.The intersection of DEGs obtained from datasets GSE132037 and GSE51701 were called the common DEGs.The potential biomarkers of HB were selected using the Random-forest algorithm and LASSO method in the GSE132037 dataset.The reliability of each biomarker was then validated in the GSE131329 dataset.Differentially expressed immune-related genes(DEIRGs)were screened by combining immune-related genes obtained from InnateDB and ImmPort databases with common DEGs.Functional and pathway enrichment analyses were performed for the DEIRGs.The protein-protein interaction(PPI)network of DEIRGs was constructed using Cytoscape software.The CIBERSORT algorithm was used to estimate the fraction of 22 immune cell types in HB tissue from the above microarray datasets.Finally,qRT-PCR was applied to validate the potential markers in HB cells and normal hepatocyte lines.Results:Nine hundred six common DEGs,and 143 DEIRGs,were identified.The biomarkers TRPC1(AUC=0.907),OGDHL(AUC=0.923),SLC1A1(AUC=0.890),and MAP7(AUC=0.971)showed high accuracy in the validation set.GO enrichment and KEGG pathway analyses of DEIRGs were focused on complement activation,NF-kappa B signaling pathway,cell cycle,and other pathways.Four hub genes(C4B,C9,C8A,MBL2)were screened.CIBERSORT analysis showed that CD8+T cells,M0 macrophages,and resting Mast cells had higher abundance in HB infiltration,while Dendritic activated cells,Neutrophils,and Monocytes had less abundance in HB infiltration.The expression trends of four markers and four hub immune-related genes in HuH6 as well as HepG2 cell line and normal hepatocyte line WRL68 were largely consistent with the database screening results as verified by qRT-PCR.Conclusion:Our study provided potential biomarkers for the precisely targeted therapy and revealed the immune characteristics of HB.