MicroRNAs accurately identify cancer tissue origin

Area of Science:

• Molecular Biology
• Oncology
• Biomarkers

Background:

• MicroRNAs (miRNAs) are noncoding RNAs involved in oncogenesis.
• miRNAs exhibit significant tissue specificity, suggesting potential for cancer classification.
• Identifying the tissue of origin for cancers of unknown primary origin is a major clinical challenge.

Purpose of the Study:

• To investigate the utility of miRNA expression profiling for classifying the tissue of origin in cancers.
• To develop and validate a miRNA-based classifier for identifying unknown primary tumors.

Main Methods:

• miRNA expression levels were measured in 400 samples from 22 tumor types using microarray analysis.
• A transparent classifier was constructed using data from 253 samples and 48 specific miRNAs.
• Independent blinded test sets were used for validation, including quantitative RT-PCR.

Main Results:

• The miRNA classifier achieved high confidence (>90% accuracy) in classifying two-thirds of samples.
• Overall high-confidence accuracy reached 89% in an independent blinded test set of 83 samples.
• Classification accuracy reached 100% for most tissue classes, including 131 metastatic samples.

Conclusions:

• MicroRNAs are effective biomarkers for determining the tissue of origin in cancers.
• miRNA profiling offers a promising approach to address the clinical problem of unknown primary cancers.
• This study validates the diagnostic potential of miRNAs in oncology.