Molecular classification of indeterminate thyroid nodules by microRNA profiling

Abstract

Thyroid nodules are common. By palpation can be found between 1 and 7% of the population. Brazilian and international guidelines recommend that nodes greater than 1 cm in patients with normal thyroid function, should be drained through Fine Needle Aspiration (FNA). About 65 to 80% of punctured nodes are considered benign and 5 to 15% malignant. However, between 15 and 30% are classified as indeterminate. These, by clinical for recommendations and risk of malignancy, are directed to surgical thyroidectomy and the postsurgical final histological analysis reclassifies 70 to 80 % of the cases as benign, showing the high rate of unnecessary surgeries. Several molecular techniques have been developed trying to help to solve this problem. Although clinically useful, none of the commercially available techniques present together highly sensitive and specificity, limiting its use and application. Using miRNAs expression data avaiable at ArrayExpress (EMBL-EBI) database, we develop a preliminar molecular classifier, in partnership with the swiss start up company SimplicityBio, which by a cross validation reach 90% of sensibility and 87% of specificity based on the analysis of only 4 miRNAs, in silico. The objective of this present study is to develop and validate a molecular classifier that, by analyzing molecular signatures generated by microRNA expression profile (microRNA profiling), is able to classify thyroid nodules classified by FNA as indeterminate into potentially benign or malignant, with high sensitivity and specificity. In this direction, we intend to analyze by RT-qPCR the expression of microRNAs selected in 160 benign and malignant thyroid samples classified as indeterminate by FNA. Half of the samples will be used for the identification and generation of specific molecular signatures of each class and the development of a molecular classifier. The other half of the samples will be used in a blind test which aims to validate the developed algorithm. The company, currently being set up, will be count with the partnership of the researchers and infrastructure of the Barretos Cancer Hospital. We intend to operate in a niche of the molecular diagnostics market hitherto unexplored in Brazil, establishing the Gene Expression Profiling technique as a technology platform, focused on personalized medicine for oncology area. Under the pillars of knowledge, open innovation and solution to the client, the company intends to differentiate from the competitors through a high-performance R&D. Since about 48,000 surgeries are performed unnecessarily each year in Brazil and generate an expense of R$ 135 million, the potential clinical impact in reducing this number of unnecessary surgeries and the economy that the health system can save and the more affordable price than competitors, once it is a national technology, are factors that contribute to the success of this product. (AU)