Background
Liquid biopsies, for example, exosomal circular RNA (circRNA) can be used to assess potential predictive markers for hepatocellular carcinoma (HCC) in patients after curative resection. This study aimed to search for effective prognostic biomarkers for HCC in patients after surgical resection based on exosomal circRNA expression profiles. We developed two nomograms incorporating circRNAs to predict the postoperative recurrence-free survival (RFS) and overall survival (OS) of HCC patients.
Method
Plasma exosomes isolated from HCC patients and healthy individuals were used for circRNA microarray analysis to explore differentially expressed circRNAs. Pearson correlation analysis was used to evaluate the correlation between circRNAs and clinicopathological features. Cox regression analysis was used to explore the correlation between circRNA and postoperative survival time as well as recurrence time. A nomogram based on circRNA and clinicopathological characteristics was established and further evaluated to predict prognosis and recurrence.
Result
Among 60 significantly upregulated circRNAs and 25 downregulated circRNAs, hsa_circ_0029325 was selected to verify its power for predicting HCC outcomes. The high expression level of exosomal hsa_circ_0029325 was significantly correlated with OS (P = 0.001, HR = 2.04, 95% CI 1.41–3.32) and RFS (P = 0.009, HR = 1.62, 95% CI 1.14–2.30). Among 273 HCC patients, multivariate regression analysis showed that hsa_circ_0029325 (HR = 1.96, 95% CI 1.21–3.18), tumor size (HR = 2.11, 95% CI 1.33–3.32), clinical staging (HR = 2.31, 95% CI 1.54–3.48), and tumor thrombus (HR = 1.74, 95% CI 1.12–2.7) were independent risk factors for poor prognosis in HCC patients after radical resection. These independent predictors of prognosis were incorporated into the two nomograms. The AUCs under the 1-year, 3-year, and 5-year survival and recurrence curves of the OS and RFS nomograms were 0.755, 0.749, and 0.742 and 0.702, 0.685, and 0.642, respectively. The C-index, calibration curves, and clinical decision curves showed that the two prediction models had good predictive performance. These results were verified in the validation cohort with 90 HCC patients.
Conclusion
Our study established two reliable nomograms for predicting recurrence and prognosis in HCC patients. We also show that it is feasible to screen potential predictive markers for HCC after curative resection through exosomal circRNA expression profile analysis.