Sanja Goč*, Zorana Dobrijević, Tamara Janković, Filip Janjić, Ninoslav Mitić and Miroslava Janković
University of Belgrade, Institute for the Application of Nuclear Energy, INEP
sanjagoc [at] inep.co.rs
Abstract
Pancreatic cancer is one of the most lethal malignancies worldwide due to the lack of effective early diagnostic tools and its asymptomatic early stages. It is the 6th leading cause of cancer-related mortality, with a 5-year survival rate ranging from 44% in localized disease to 3% in metastatic cases. CA19-9, the most commonly used biomarker, has limited sensitivity and low positive predictive value, restricting its use in screening. Therefore, novel biomarkers are needed, and extracellular vesicle-associated miRNAs (EV-miRNAs) are increasingly recognized for their stability and potential in non-invasive cancer diagnosis. In this study, we performed an in silico analysis of plasma EV-miRNA expression profiles using the GEO dataset GSE221185 (BioProject: PRJNA913165), including 16 pancreatic cancer patients and 13 healthy controls. Seventeen differentially expressed miRNAs (5 downregulated and 12 upregulated) were identified, of which nine candidates were prioritized based on network topology metrics, including the number of single-line regulators (NSR) and transcription factor ratio (TFR) (p < 0.05). Further refinement using dbDEMC 2.0 meta-profiling identified five miRNAs (miR-218-5p, miR-221-3p, miR-423-3p, miR-505-3p, and miR-766-3p) exhibiting |log2FC| ≥ 2. Receiver operating characteristic (ROC) analysis demonstrated that only hsa-miR-766-3p achieved statistically significant diagnostic performance (AUC = 0.750, p = 0.028). Network analysis using miRNet 2.0 revealed that the nine miRNAs collectively regulate 3,499 target genes. Based on degree and betweenness centrality, three regulatory clusters were identified: high-centrality hub miRNAs (miR-218-5p, miR-221-3p, miR-423-3p), intermediate regulators (miR-590-3p, miR-766-3p, miR-512-3p), and peripheral miRNAs. Functional enrichment analysis indicated involvement of key oncogenic pathways, including PI3K–AKT, EGFR/MAPK, JAK–STAT, cell cycle regulation, apoptosis, angiogenesis, and TGF-β/EMT signaling. The four final candidate miRNAs (miR-218-5p, miR-221-3p, miR-423-3p, and miR-766-3p) collectively targeted 1,922 genes, including 14 pancreatic cancer-associated genes. TargetLink2.0 confirmed validated interactions for miR-218-5p (CDK6, E2F2, EGFR, IKBKB) and miR-221-3p (PIK3R1). External validation using independent serum-based GEO datasets (GSE85589 and GSE59856) demonstrated significant diagnostic performance for miR-221-3p (AUC = 0.734, p < 0.0001) and miR-218-5p (AUC = 0.676, p = 0.016). Overall, miR-218-5p and miR-221-3p represent promising non-invasive diagnostic biomarkers for pancreatic cancer, warranting further experimental validation.
Keywords: Extracellular vesicles, microRNA, Pancreatic cancer
Acknowledgement: This work was supported by the Ministry of Science, Technological Development, and Innovation, Republic of Serbia [Contract No: 451-03-33/2026-03/200019]. This research aligns with the Agenda 2030- United Nations Sustainable Development Goal 3, promoting good health and well-being.