Integrated Cross-Dataset Analysis of Transcriptional Signatures and Hub Genes Across Cirrhosis, Hepatocellular Carcinoma, and Cholangiocarcinoma

ABSTRACT Background Cirrhosis is a major risk factor for primary liver malignancies, but shared and distinct transcriptional features of cirrhosis-associated hepatocellular carcinoma (HCC) and cholangiocarcinoma (CCA) remain incompletely characterized. Methods We performed an integrated multi-dataset analysis of 12 public Gene Expression Omnibus (GEO) microarray datasets comprising normal, cirrhotic, HCC, intrahepatic CCA (iCCA), and extrahepatic CCA (eCCA) tissues. Differentially expressed genes (DEGs) were identified using adjusted p < 0.05 and |log₂FC| > 1, followed by functional enrichment, protein–protein interaction (PPI) analysis, hub-gene identification, and GEPIA2-based expression and survival analyses. Cross-dataset recurrence analyses were additionally performed to assess the consistency of DEG findings across the available datasets. GEO accession numbers were GSE14323, GSE15765, GSE17548, GSE40367, GSE45050, GSE45267, GSE60502, GSE63898, GSE84402, GSE89747, GSE101685, and GSE132305. Results HCC-related comparisons were enriched for cell-cycle and proliferation-associated genes, including TOP2A, RRM2, BIRC5, CCNA2, BUB1, BUB1B, and ASPM . CCA-related comparisons highlighted genes including LAMC2, AGRN, MME, EGR1, and FOSB . Cirrhotic tissues showed alterations involving metabolic and immune-related genes, including MT1M, ALB, APOA2, and SERPINC1 . Cross-dataset recurrence analysis showed variable but detectable agreement among datasets, reflecting the heterogeneity and unequal availability of public datasets across comparisons. Conclusions This integrated analysis identifies shared and disease-associated transcriptional patterns across cirrhosis, HCC, and CCA. The identified genes should be considered candidate molecular signatures. The absence of a direct Cirrhosis–CCA comparison, heterogeneity among datasets and CCA subtypes, and reliance on univariate survival analysis limit the clinical interpretation of the findings.