Cross-study transcriptomics and machine learning approaches reveal a predictive six-gene panel for hepatocellular carcinoma

Hepatocellular carcinoma (HCC) is a leading cause of cancer-related death worldwide, and its prognosis remains poor because most cases are diagnosed at an advanced stage. Current diagnostic methods, including liver function tests (LFTs), alpha-fetoprotein (AFP) panels, and ultrasound, lack sufficient specificity for early and comprehensive detection of HCC, highlighting the need for reliable early diagnostic biomarkers. In this study, we analyzed publicly available RNA-seq datasets to identify HCC-associated gene biomarkers using machine learning (ML) techniques, such as recursive feature elimination (RFE) and random forest importance (RFI). To improve clinical interpretability, we also developed a concise risk score based on logistic regression (LR). Additionally, STRING and Gene Ontology (GO) analyses were performed to assess the biological relevance and functional enrichment of the identified genes. We identified a 6 gene signature comprising CDKN3, LIFR, MKI67, TOP2A, SLC5A1, and VIPR1 that showed consistent differential expression between HCC and non-tumor samples across multiple independent datasets and demonstrated strong potential as an early-stage HCC biomarker panel. The LR based risk score derived from these six genes effectively distinguished HCC from normal samples with an accuracy of 0.94. Furthermore, a random forest (RF) model using 52 significant genes (52 SGs) achieved an accuracy of 0.96 on an independent test set, further indicating pronounced transcriptomic differences between tumor and non-tumor states. Functional enrichment analyses revealed significant associations with key biological processes, including cell cycle regulation, apoptosis, and cytoskeletal organization, supporting the role of these genes in HCC development and progression. Overall, these findings emphasize a strong 6 gene signature and a complementary ML framework that may assist in early HCC detection and improve understanding of the disease’s molecular basis.