The molecular determinants of thyroid follicular nodules are incompletely understood and assessment of malignancy is a diagnostic challenge. Since microRNA (miRNA) analyses could provide new leads to malignant progression, we characterised the global miRNA expression in follicular adenoma (FA) and follicular carcinoma (FC). Comparison of carcinoma and adenoma with normal thyroid revealed 150 and 107 differentially expressed miRNAs respectively. Most miRNAs were down-regulated and especially miR-199b-5p and miR-144 which were essentially lost in the carcinomas. Integration of the changed miRNAs with differentially expressed mRNAs demonstrated an enrichment of seed sites among up-regulated transcripts encoding proteins implicated in thyroid tumourigenesis. This was substantiated by the demonstration that pre-miR-199b reduced proliferation when added to cultured follicular thyroid carcinoma cells. The down-regulated miRNAs in FC exhibited a substantial similarity with down-regulated miRNAs in anaplastic carcinoma (AC) and by gene set enrichment analysis, we observed a significant identity between target mRNAs in FC and transcripts up-regulated in AC. To examine the diagnostic potential of miRNA expression pattern in distinguishing malignant from benign nodules we employed a supervised learning algorithm and leave-one-out-cross-validation. By this procedure, FA and FC were identified with a negative predicted value of 83% (data generated by microarray platform) and of 92% (data generated by qRT-PCR platform). We conclude that follicular neoplasia is associated with major changes in miRNA expression that may promote malignant transformation by increasing the expression of transcripts encoding tumourigenic factors. Moreover, miRNA profiling may facilitate the diagnosis of carcinoma vs adenoma.