Development and validation of a mitochondrial energy metabolism-related risk model in hepatocellular carcinoma
Xin Gao,Mingyue Xu,Heng Wang,Zhaozhi Xia,Hongrui Sun,Meng Liu,Shuchao Zhao,Faji Yang,Zheyu Niu,Hengjun Gao,Huaqiang Zhu,Jun Lu,Xu Zhou
DOI: https://doi.org/10.1016/j.gene.2022.147133
IF: 3.913
2022-12-23
Gene
Abstract:Background Hepatocellular carcinoma (HCC) is one of the most prevalent cancers and ranks third inmortality. Mitochondria are the energy manufacturers of cells. Disruption of mitochondrial energy metabolism pathways is strongly correlated with the onset and progression of HCC. Aberrant genes in mitochondrial energy metabolism pathways may represent a unique diagnostic and therapeutic targets that act as indicators for HCC. Methods Gene expression data from 374 HCC patients and 50 controls were acquired from TCGA database. A total of 188 mitochondrial energy metabolism-related genes (MMRGs) were obtained from KEGG PATHWAY database. A total of 368 patients with survival data were randomly split into training and validation groups in a 7: 3 ratio. Prognosis-related MMRGs were selected by univariate Cox and LASSO analyses. Kaplan–Meier and ROC curves were employed to analyze the model precision, whereas the validation set was used for model verification. Furthermore, clinical examinations, immune infiltration analysis, GSVA, and immunotherapy analysis were conducted in the high- and low-risk groups. Finally, the risk model was combined with the clinical variables of HCC patients to perform univariate and multivariate Cox regression analyses to obtain independent risk indicators and draw a nomogram. Therefore, we evaluated the accuracy of the predictions using calibration curves. Results A total of 6032 differentially expressed genes (DEGs) were detected in the HCC and control samples. After overlapping DEGs with 188 MMRGs, 42 mitochondrial energy metabolism-related DEGs (DEMMRGs) were identified. A 17 specific genes-based risk score model of HCC was created, which revealed effectiveness in each TCGA training and validation dataset. Moreover, patients categorized by risk scores exhibited distinct immune infiltration status, immunotherapy responsiveness, and functional properties. Finally, univariate and multivariate Cox regression analyses revealed that risk score and stage T were independent predictive variables. Based on the T stage and risk score, a nomogram for estimating the survival of HCC patients was created. The calibration curves demonstrated that the prediction model had a high level of accuracy. Conclusions Our study constructed a mitochondrial energy metabolism-related risk model, that may be utilized to anticipate HCC prognosis and represent the immunological microenvironment of HCC.
genetics & heredity