A Three-MicroRNA Panel in Serum as Novel Biomarker for Papillary Thyroid Carcinoma Diagnosis
Introduction
Thyroid cancer, particularly papillary thyroid carcinoma (PTC), accounts for 85–90% of all thyroid malignancies. Despite its generally favorable prognosis, timely and accurate diagnosis remains critical for effective management. Current diagnostic methods, such as fine-needle aspiration (FNA) biopsy and imaging techniques, face limitations including invasiveness, cost, and dependency on operator expertise. Consequently, non-invasive biomarkers for PTC are urgently needed. MicroRNAs (miRNAs), small non-coding RNAs regulating gene expression, have emerged as promising candidates due to their stability in circulation and disease-specific expression profiles. This study aimed to identify serum-based miRNA signatures for non-invasive PTC diagnosis and validate their clinical utility.
Methods
Study Design and Participants
The study enrolled 100 PTC patients, 30 nodular goiter (NG) patients, and 96 healthy controls (HCs) from 2015 to 2017. All PTC cases were pathologically confirmed, and blood samples were collected pre-treatment. Serum was isolated using serum separation tubes, centrifuged, and stored at −80°C. The study comprised four phases: screening, training, testing, and external validation.
Screening Phase: Pooled serum samples from 20 PTC patients, 20 NG patients, and 10 HCs were analyzed using the Exiqon miRNA qPCR Panel (179 miRNAs). Candidate miRNAs were selected based on fold change (FC >1.5 or <0.67) and consistent dysregulation in PTC vs. HCs. Four literature-reported miRNAs (miR-95-5p, miR-190a-5p, miR-151a-5p, and miR-222-3p) were also included.
Training and Testing Phases: Individual serum samples from 34 PTC and 35 HCs (training) and 48 PTC and 45 HCs (testing) were analyzed via qRT-PCR. miRNAs with FC >1.5 and P <0.05 were validated.
External Validation: An independent cohort (18 PTC vs. 16 HCs) and NG comparison (30 PTC vs. 30 NG) further evaluated diagnostic accuracy.
Tissue and Exosome Analysis: miRNA expression was assessed in 23 PTC tumor tissues, adjacent normal tissues, and serum-derived exosomes (24 PTC vs. 24 HCs). TCGA data (59 PTC tissues) provided additional validation.
RNA Extraction and qRT-PCR
Total RNA from serum, tissues, and exosomes was extracted using the mirVana PARIS Kit, with cel-miR-39 spiked-in for normalization. Reverse transcription used Bulge-Loop™ primers, and qRT-PCR was performed on a LightCycler® 480 System. Data analysis followed the 2−ΔΔCt method, with U6 as the endogenous control for tissues.
Statistical Analysis
Mann-Whitney U tests compared miRNA expression. Diagnostic performance was evaluated via receiver operating characteristic (ROC) curves and area under the curve (AUC). Logistic regression constructed a multi-miRNA panel.
Results
Identification of Candidate miRNAs
The Exiqon panel identified 40 miRNAs (36 up-, 4 down-regulated) in PTC. Subsequent qRT-PCR validation in training and testing phases narrowed these to three consistently up-regulated miRNAs: miR-25-3p, miR-296-5p, and miR-92a-3p (all P 1.5). External validation confirmed these findings (P <0.001).
Diagnostic Performance
- Individual miRNAs: Combined data from all phases revealed AUCs of 0.623 (miR-25-3p), 0.621 (miR-296-5p), and 0.702 (miR-92a-3p).
- Three-miRNA Panel: Logistic regression generated the formula:
Logit(P) = 1.768 − 0.009 × miR-25-3p − 0.187 × miR-296-5p − 0.797 × miR-92a-3p.
The panel’s AUCs were 0.727 (training), 0.771 (testing), and 0.862 (external validation). Combined AUC reached 0.775, with 84.8% sensitivity and 62.2% specificity.
Differentiation from Benign Disease
The panel distinguished PTC from NG with an AUC of 0.969 (95% CI: 0.927–1.000), outperforming individual miRNAs.
Tissue and Exosome Analysis
- Tissues: miR-25-3p and miR-92a-3p were downregulated in PTC tumors vs. normal tissues (P =0.025 and P =0.043, respectively), contrasting serum findings. TCGA data corroborated miR-25 and miR-296 downregulation in tumors.
- Exosomes: miR-296-5p remained upregulated (P =0.019), while miR-25-3p and miR-92a-3p were downregulated (P <0.001 and P =0.006), suggesting distinct miRNA carriers in circulation.
Bioinformatics Analysis
DIANA-TarBase and miRPath implicated these miRNAs in cancer-related pathways (e.g., viral carcinogenesis, cell cycle) and biological processes (e.g., protein binding, metabolic regulation).
Discussion
This study identifies a serum three-miRNA panel (miR-25-3p, miR-296-5p, miR-92a-3p) with robust diagnostic accuracy for PTC. The panel’s superiority over individual miRNAs highlights the synergistic value of multi-analyte approaches.
Mechanistic Insights
- miR-25-3p: Promotes thyroid cancer progression via SOCS4 inhibition and MEK4 modulation, though its downregulation in tissues suggests complex regulatory roles.
- miR-296-5p: Exhibits dual oncogenic/tumor-suppressive roles across cancers; its consistent serum upregulation in PTC warrants further investigation.
- miR-92a-3p: Linked to VHL suppression in PTC tissues, its serum elevation may reflect tumor-secreted miRNA release.
Tissue vs. Serum Discordance
The inverse expression in serum and tissues underscores the complexity of miRNA biology. Serum miRNAs may originate from tumor microenvironment interactions, immune cells, or non-tumor sources, rather than direct tumor shedding.
Clinical Implications
The panel’s high AUC (0.969) in differentiating PTC from NG addresses a critical clinical need, as benign nodules often mimic malignancy on imaging. Non-invasive miRNA testing could reduce unnecessary FNAs and guide treatment decisions.
Limitations and Future Directions
Sample size constraints, especially in tissue/exosome analyses, warrant larger cohorts. Mechanistic studies are needed to clarify miRNA roles in PTC pathogenesis. Longitudinal studies assessing post-treatment miRNA dynamics could enhance prognostic utility.
Conclusion
The three-miRNA serum panel demonstrates significant potential as a non-invasive biomarker for PTC diagnosis, offering high sensitivity and specificity. Integration into clinical workflows could complement existing methods, improving diagnostic accuracy and patient outcomes. Further validation and mechanistic exploration are essential for translational application.
doi.org/10.1097/CM9.0000000000001107
Was this helpful?
0 / 0