Diagnosis of thyroid by fine needle aspiration is challenging for the "indeterminate" category and can be supported by molecular testing. We set out to identify miRNA markers that could be used in a diagnostic setting to improve the discrimination of mutation-negative indeterminate fine needle aspirations. miRNA high-throughput sequencing was performed for freshly frozen tissue samples of 19 RAS and PAX8/PPARG mutation-negative follicular thyroid carcinomas, and 23 RAS and PAX8/PPARG mutation-negative follicular adenomas. Differentially expressed miRNAs were validated by quantitative polymerase chain reaction in a set of 44 fine needle aspiration samples representing 24 follicular thyroid carcinomas and 20 follicular adenomas. Twenty-six miRNAs characterized by a significant differential expression between follicular thyroid carcinomas and follicular adenomas were identified. Nevertheless, since no single miRNA had satisfactory predictive power, classifiers comprising two differentially expressed miRNAs were designed with the aim to improve the classification. Six two-miRNA classifiers were established and quantitative polymerase chain reaction validated in fine needle aspiration samples. Four out of six classifiers were characterized by a high specificity (≥94 %). The best two-miRNA classifier (miR-484/miR-148b-3p) identified thyroid malignancy with a sensitivity of 89 % and a specificity of 87 %. The high-throughput sequencing allowed the identification of subtle differences in the miRNA expression profiles of follicular thyroid carcinomas and follicular adenomas. While none of the differentially expressed miRNAs could be used as a stand-alone malignancy marker, the validation results for two-miRNA classifiers in an independent set of fine needle aspirations are very promising. The ultimate evaluation of these classifiers for their capability of discriminating mutation-negative indeterminate fine needle aspirations will require the evaluation of a sufficiently large number of fine needle aspirations with histological confirmation.
Keywords: Classifier; Fine needle aspiration cytology; Follicular thyroid adenoma; Follicular thyroid cancer; High-throughput sequencing; miRNA.