Abstract
The detection of somatic mutations in indeterminate or follicular proliferation fine-needle aspiration cytologies (FNACs) is able to clarify only a subgroup of those FNACs. Therefore, further markers to differentiate this problematic FNAC category by the identification of mutation negative thyroid cancers and benign nodules are urgently needed. Our objective was to evaluate previously published miRNA markers and discover novel ones from all publicly available miRNA expression profiling data sets. By literature review and data repository search we gathered 3 data sets describing human miRNA expression profiles of follicular thyroid cancer (FTC) and follicular adenoma (FA) samples. Literature review summarized 27 previously published miRNAs, which were validated in the 3 available data sets. By means of uniform statistical analysis 6 further miRNAs were identified and tested in an independent, previously published microarray data set. Meta-analysis confirmed 7 out of 27 previously published, and 4 out of 6 de novo identified miRNAs. The low confirmation rate of previously published miRNA markers was induced by low numbers of samples in the analyzed studies and high false discovery rates that were higher than 0.2. Finally, miR-637, miR-181c-3p, miR-206, and miR-7-5p were discovered as de novo potential FTC markers and validated in at least one independent, previously published data set. Two out of these new identified miRNAs (miR-7-5p and miR-206) were validated by qPCR in an independent sample set of 32 FTC and 46 FA samples. Especially miR-7-5p was able to differentiate benign and malignant thyroid tumors in several datasets.
Key words
miRNA - follicular thyroid tumors - microarray - integrated statistical analysis
