Machine learning and experimental analyses identified miRNA expression models associated with metastatic osteosarcoma

Area of Science:

• Oncology
• Molecular Biology
• Bioinformatics

Background:

• Osteosarcoma (OS) is the most common primary bone cancer with high metastatic potential and poor prognosis.
• Early diagnosis of metastasis is crucial for effective treatment and improved patient outcomes.
• MicroRNA (miRNA) expression profiles are increasingly recognized for their role in cancer progression and metastasis.

Purpose of the Study:

• To identify early diagnostic biomarkers for osteosarcoma metastasis using miRNA expression profiles.
• To develop and validate machine learning-based diagnostic models for differentiating metastatic from non-metastatic osteosarcoma.
• To explore the potential of identified miRNA signatures in guiding therapeutic strategies.

Main Methods:

• RNA sequencing and analysis of public microarray data to identify differentially expressed miRNAs in osteosarcoma.
• Application of seven feature selection algorithms and five machine learning systems for model development.
• Network analysis and RT-qPCR validation in patient biopsies to confirm miRNA expression levels and diagnostic power.
• KEGG and Gene Ontology (GO) term enrichment analysis to identify associated biological pathways.

Main Results:

• miR-34c-3p and miR-154-3p were identified as the most promising diagnostic biomarkers for metastatic osteosarcoma.
• The combined signature of miR-34c-3p and miR-154-3p demonstrated 90% diagnostic power for osteosarcoma and 85% for metastatic osteosarcoma (AUC=0.90 and 0.85, respectively).
• Lower expression of miR-34c-3p and miR-154-3p was confirmed in OS samples compared to benign tissues, with a direct correlation to tumor grade.
• Epithelial-to-mesenchymal transition (EMT), adhesion junction, and focal adhesion pathways were significantly associated with the top miRNAs.

Conclusions:

• Novel miRNA expression signatures, specifically miR-34c-3p and miR-154-3p, hold significant potential for the early detection of metastatic osteosarcoma.
• These biomarkers may facilitate quantitative and rapid metastasis detection, aiding in the development of targeted molecular therapies.
• Further clinical validation is warranted to confirm the utility of miR-34c-3p and miR-154-3p as diagnostic biomarkers for osteosarcoma metastasis.