Establishment of a Genomic-Clinicopathologic Nomogram for Predicting Early Recurrence of Hepatocellular Carcinoma After R0 Resection

Abstract

BackgroundA high rate of postoperative recurrence, especially early recurrence (ER) occurring within 1 year, seriously impedes patients with hepatocellular carcinoma (HCC) from achieving long-term survival. This study aimed to establish a genomic-clinicopathologic nomogram for precisely predicting ER in HCC patients after R0 resection. MethodsTwo reliable datasets from The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO) databases were selected as the training and validation cohorts, respectively. The prognostic genes related to ER were screened out by univariate Cox regression analysis and differential expression analysis. The gene-based prognostic index was constructed using LASSO and Cox regression analyses, and its independent prognostic value was assessed by Kaplan-Meier and multivariate Cox analyses. Gene set enrichment analysis (GSEA) was performed to explore the biological pathways related to the prognostic index. Finally, the nomogram integrating all the independent prognostic factors was established and comprehensively evaluated by calibration plots, the C-index, receiver operating characteristic curves, and decision curve analysis. ResultsNine dysregulated and prognostic genes related to ER (ZNF131, TATDN2, TXN, DDX55, KPNA2, ZNF30, TIMELESS, SFRP1, and COLEC11) were identified (all P < 0.05). The prognostic index model based on the 9 genes was successfully constructed using the TCGA cohort and showed a certain capability to discriminate the ER group from the non-ER group (P < 0.05) and good independent prognostic value in terms of predicting poor early recurrence-free survival (P < 0.05). Eight biological pathways significantly related to ER were identified by GSEA, such as “cell cycle”, “homologous recombination” and “p53 signaling pathway.” The genomic-clinicopathologic nomogram integrating the 9-gene-based prognostic index and TNM stage displayed significantly higher predictive accuracy and clinical application value than that of TNM stage model both in the training and validation cohorts (all P < 0.05). ConclusionsThe novel genomic-clinicopathologic nomogram may be a convenient and powerful tool for accurately predicting ER in HCC patients after R0 resection.