44 chapter 2 their models. Agretti et al. assessed a frequently quoted set of microRNAs consisting of miR-146b, miR-155, miR-187, miR-197, miR-221, miR-222 and miR-224 [186, 187]. Published in 2008, Nikiforova et al. had demonstrated that this 7-microRNA set in FNAC samples had 100% sensitivity and 94% specificity if one of the included microRNAs showed an at least two-fold overexpression [186]. Analytic validation of this model by Agretti et al. showed differential upregulation in PTC of all of these microRNAs except miR-197. In particular, miR-146b showed a >30-fold higher expression in PTC. A decision tree including miR-146b, miR-155 and miR-221 was 98% accurate in the test set, but validation in an independent set of indeterminate FNAC samples was unsuccessful, yielding mere 60% sensitivity and 58% specificity [187]. Vriens et al. used a microRNA array to detect 10 genes that were up- or downregulated by ≥5fold in thyroid malignancies. Four microRNAs (miR-100, miR-125b, miR-138 and miR-768-3p) were significantly downregulated and accurately differentiated between benign and malignant follicular and Hürthle cell neoplasms in the test set. In their validation set of 125 indeterminate FNAC samples, only miR-138 was moderately distinctive with 81% NPV. For Hürthle cell carcinoma, miR-138 and miR768-3p were both 98% accurate [198]. Finally, in a recent Italian study only miR-375 accurately differentiated between benign and malignant neoplasms. Subsequently, in TIR3 cytology excluding Hürthle cell lesions, a 12-fold or higher overexpression of miR-375 perfectly distinguished benign from malignant lesions with 100% accuracy. It was also significantly differently expressed between TIR3A and TIR3B categories and correlated with a different malignancy risk [199]. Availability and limitations of microRNA expression analysis MicroRNA expression analysis has advantages over other techniques. MicroRNAs are more stable than mRNA at maintaining their expression in formalin-fixed paraffin-embedded (FFPE) tissue samples as well as FNAC specimens, irrespective of the preservation method (e.g. archived FNAC slides or nucleic acid preservation solutions) [186, 199]. Recently microRNA expression was even successfully measured in serum [200]. Moreover, microRNA expression levels measured with generic methods (e.g. quantitative RT-PCR) correspond well to their biological effect, as microRNAs affect biological processes without the additional step of protein synthesis [186]. However, general limitations of FNAC also translate to concerns with microRNA analysis: scant cellularity or low levels of malignant cells in FNAC specimens could cause a false-negative microRNA test result [187]. Another limitation is the plurality of microRNAs associated with DTC in histopathological studies, causing vast heterogeneity between the limited number of studies in indeterminate cytology. Validation studies of the same microRNA set are lacking. Simultaneously, new microRNAs are still correlated to thyroid carcinoma. Ongoing research has yet to compose the optimal set of microRNAs. Recently, the first commercial test was marketed as the ThyraMIR™ (Interpace Diagnostics, Parsippany, NJ, USA). It evaluates the expression levels of miR-29b-1–5p, miR-31–5p, miR-138–1-3p, miR-139–5p, miR-146b-5p, miR-155, miR-204–5p, miR-222–3p, miR375, and miR-551b-3p. The ThyraMIR™ demonstrated 57% sensitivity and 92% specificity in 109
RkJQdWJsaXNoZXIy MTk4NDMw