Backgrounds: Liver hepatocellular carcinoma (LIHC) is one of the most malignant tumors, of which prognosis is unsatisfactory in most cases and metastatic of LIHC often results in poor prognosis. In this study, we aimed to construct a metastasis- related mRNAs prognostic model to increase the accuracy of prediction of LIHC prognosis.
Methods: 374 LIHC samples and 50 normal samples were downloaded from TCGA database, involving transcriptomic and clinical data. Metastatic-related genes were acquired from HCMBD website at the same time. 343 samples were randomly divided into train dataset and test dataset with a proportion of 1:1 by using caret package in R. Kaplan-Meier method and univariate Cox regression analysis and lasso regression analysis were performed to obtain metastasis-related mRNAs which played significant roles in prognosis. Then, using multivariate Cox regression analysis, a prognostic prediction model was established. Transcriptome and clinical data were combined to construct a prognostic model and a nomogram for OS evaluation. Functional enrichment in high- and low-risk groups were also analyzed by GSEA. An entire set was applied to verify the model.
Results: 1895 metastasis-related mRNAs were screened and 8mRNAs were associated with prognosis. The overall survival (OS)-related prognostic model which was constructed based on 4 MRGs (MMP1, SPP1, STC2, CDCA8) significantly stratified LIHC patients into high- and low-risk groups. The AUC values of the 4-gene prognostic signature at 1 year, 2 years, and 3 years were 0.807,0.729 and 0.673. A risk score based on the signature was a significantly independent prognostic factor (HR=1.295; 95%CI=1.167-1.436; P<0.001) for LIHC patients. A nomogram which incorporated the 4-gene signature and clinical features was also built for prognostic prediction. GSEA results that low- and high-risk group had an obviously difference in part of pathways. The value of this model was validated in test dataset and entire set.
Conclusion: Metastasis-related mRNAs prognostic model was verified that it had a predictable value on the prognosis of LIHC, which could be helpful for gene targeted therapy.