Abstract
Hepatocellular carcinoma (HCC) is one of the most prevalent life-threatening human cancers and the leading cause of cancer-related mortality, with increased global incidence within the last decade. Identification of effective diagnostic and prognostic biomarkers would enable reliable risk stratification and efficient screening of high-risk patients, thereby facilitating clinical decision-making. Herein, we performed a comprehensive, robust DNA methylation analysis based on genome-wide DNA methylation profiling. We constructed a diagnostic signature with five DNA methylation markers, which precisely distinguished HCC patients from normal controls. Cox regression and LASSO analysis were applied to construct a prognostic signature with four DNA methylation markers. A one-to-one correlation analysis was carried out between genes of the whole genome and our prognostic signature. Exploration of the biological function and the role of the underlying significantly correlated genes was conducted. A mixed dataset of 463 HCC patients and 253 normal controls, derived from six independent datasets, was used to valid the diagnostic signature. Results showed a specificity of 96.84% and sensitivity of 96.77%. Class scores for the diagnostic signature were significantly different between normal controls, individuals with liver diseases, and HCC patients. The present signature has the potential to serve as a biomarker to monitor health in normal controls. Additionally, HCC patients were successfully separated into low-risk and high-risk groups by the prognostic signature, with a better prognosis for patients in the low-risk group. Kaplan-Meier and ROC analysis confirmed that the prognostic signature performed well. We found eight of the top ten genes to positively correlate with risk scores of the prognostic signature, and to be involved in cell cycle regulation. This eight-gene panel also served as a prognostic signature. The robust evidence presented in this study therefore demonstrates the effectiveness of the prognostic signature. In summary, we constructed diagnostic and prognostic signatures, which have potential for use in diagnosis, surveillance, and prognostic prediction for HCC patients. Eight genes that were significantly and positively correlated with the prognostic signature were strongly associated with cell cycle processes. Therefore, the prognostic signature can be used as a guide by which to measure responsiveness to cell-cycle-targeting agents.
Highlights
Hepatocellular carcinoma (HCC) is a leading cause of cancerrelated deaths worldwide, especially in developing countries, with more than half a million deaths per year (Yang and Roberts, 2010)
Profiles of 371 HCC patients and 50 normal controls with clinical survival information were obtained from The Cancer Genome Atlas (TCGA) using UCSC Xena1, with the remaining profiles of 461 HCC patients and 253 normal controls from six independent datasets downloaded from the Gene Expression Omnibus (GEO) database (GSE54503, GSE56588, GSE60753, GSE75041, GSE77269, and GSE89852)
After 100 simulations, the diagnostic signature consisting of the top 5 methylation markers discriminated HCC patients and normal controls with the highest average balanced accuracy (BA) (Figure 1B and Supplementary Figure S2B, Supplementary Tables S1, S2). These five DNA methylation markers correspond to LDL receptor-related protein 5 (LRP5), T-box transcription factor 15 (TBX15), NCK associated protein 1 like (NCKAP1L), paired like homeodomain 1 (PITX1), and homeobox A10 (HOXA10), respectively
Summary
Hepatocellular carcinoma (HCC) is a leading cause of cancerrelated deaths worldwide, especially in developing countries, with more than half a million deaths per year (Yang and Roberts, 2010). Based on current clinical practice guidelines, surgical resection is the optimal treatment for patients with a single tumor lesion and well-preserved liver function. Even in this subgroup, the 5-year post-treatment recurrence rate approaches 70%, with no adjuvant therapy available (Villanueva et al, 2011; European Association for the Study of the Liver, 2018). Pre-treatment evaluation of patients can help identify individuals with a high risk for recurrence and metastasis as well as poor prognosis. There is an urgent need to identify accurate and effective biomarkers for diagnosis and prognosis
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
