A circulating three-miRNA panel (,,) for early-stage ovarian cancer detection: a machine-learning bioinformatics approach
Source: PubMed Central Open Access, NCBI / U.S. National Library of Medicine
Background Ovarian cancer (OVCA) remains one of the most lethal gynecological malignancies, primarily due to late-stage diagnosis and the lack of reliable early-detection biomarkers. Circulating microRNAs (miRNAs) have emerged as promising non-invasive biomarkers for cancer detection and prognosis. Objective This study aimed to computationally identify circulating miRNAs associated with early-stage OVCA using publicly available datasets and bioinformatics workflows. Methods Differential expression analysis was performed on miRNA-Seq datasets from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO). Functional enrichment analysis and pathway annotation were performed using miEAA and PANTHER. A random forest-based machine-learning model was developed and optimized for miRNA biomarker classification. Results Differential expression analysis revealed distinct miRNA signatures between OVCA and other cancer types (BRCA, CESC, UCEC, and COAD), as well as between OVCA and control samples. Stage-specific analysis identified key miRNAs, including,, and, consistently associated with early-stage OVCA. Functional enrichment analysis highlighted key pathways, including TP53 and VEGFA signaling, central to OVCA pathogenesis. The random forest classifier demonstrated robust performance with an accuracy of 91.67% and an area under the curve (AUC) of 0.991. Conclusion This study identifies a panel of circulating miRNAs with significant diagnostic potential for early-stage OVCA. In
Abstract
Background Ovarian cancer (OVCA) remains one of the most lethal gynecological malignancies, primarily due to late-stage diagnosis and the lack of reliable early-detection biomarkers. Circulating microRNAs (miRNAs) have emerged as promising non-invasive biomarkers for cancer detection and prognosis. Objective This study aimed to computationally identify circulating miRNAs associated with early-stage OVCA using publicly available datasets and bioinformatics workflows. Methods Differential expression analysis was performed on miRNA-Seq datasets from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO). Functional enrichment analysis and pathway annotation were performed using miEAA and PANTHER. A random forest-based machine-learning model was developed and optimized for miRNA biomarker classification. Results Differential expression analysis revealed distinct miRNA signatures between OVCA and other cancer types (BRCA, CESC, UCEC, and COAD), as well as between OVCA and control samples. Stage-specific analysis identified key miRNAs, including,, and, consistently associated with early-stage OVCA. Functional enrichment analysis highlighted key pathways, including TP53 and VEGFA signaling, central to OVCA pathogenesis. The random forest classifier demonstrated robust performance with an accuracy of 91.67% and an area under the curve (AUC) of 0.991. Conclusion This study identifies a panel of circulating miRNAs with significant diagnostic potential for early-stage OVCA. Integration of these miRNAs into clinical workflows could enhance early detection and improve patient outcomes. Further validation using independent cohorts is warranted.