DataSheet_1_Prognostic Implication of a Novel Metabolism-Related Gene Signature in Hepatocellular Carcinoma.docx (137.37 kB)
Download file

DataSheet_1_Prognostic Implication of a Novel Metabolism-Related Gene Signature in Hepatocellular Carcinoma.docx

Download (137.37 kB)
dataset
posted on 04.06.2021, 04:58 by Chaoyan Yuan, Mengqin Yuan, Mingqian Chen, Jinhua Ouyang, Wei Tan, Fangfang Dai, Dongyong Yang, Shiyi Liu, Yajing Zheng, Chenliang Zhou, Yanxiang Cheng
Background

Hepatocellular carcinoma (HCC) is one of the main causes of cancer-associated deaths globally, accounts for 90% of primary liver cancers. However, further studies are needed to confirm the metabolism-related gene signature related to the prognosis of patients with HCC.

Methods

Using the “limma” R package and univariate Cox analysis, combined with LASSO regression analysis, a metabolism-related gene signature was established. The relationship between the gene signature and overall survival (OS) of HCC patients was analyzed. RT-qPCR was used to evaluate the expression of metabolism-related genes in clinical samples. GSEA and ssGSEA algorithms were used to evaluate differences in metabolism and immune status, respectively. Simultaneously, data downloaded from ICGC were used as an external verification set.

Results

From a total of 1,382 metabolism-related genes, a novel six-gene signature (G6PD, AKR1B15, HMMR, CSPG5, ELOVL3, FABP6) was constructed based on data from TCGA. Patients were divided into two risk groups based on risk scores calculated for these six genes. Survival analysis showed a significant correlation between high-risk patients and poor prognosis. ROC analysis demonstrated that the gene signature had good predictive capability, and the mRNA expression levels of the six genes were upregulated in HCC tissues than those in adjacent normal liver tissues. Independent prognosis analysis confirmed that the risk score and tumor grade were independent risk factors for HCC. Furthermore, a nomogram of the risk score combined with tumor stage was constructed. The calibration graph results demonstrated that the OS probability predicted by the nomogram had almost no deviation from the actual OS probability, especially for 3-year OS. Both the C-index and DCA curve indicated that the nomogram provides higher reliability than the tumor stage and risk scores. Moreover, the metabolic and immune infiltration statuses of the two risk groups were significantly different. In the high-risk group, the expression levels of immune checkpoints, TGF-β, and C-ECM genes, whose functions are related to immune escape and immunotherapy failure, were also upregulated.

Conclusions

In summary, we developed a novel metabolism-related gene signature to provide more powerful prognostic evaluation information with potential ability to predict the immunotherapy efficiency and guide early treatment for HCC.

History

References