Background: Thyroid nodules are common in adults, but only a small fraction of them are malignant. A large number of thyroid malignancies, such as noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) are highly indolent tumors. About one-third of thyroid nodules that underwent FNA fall into the indeterminate cytology (IC), their management being challenging. Different biomarkers were proposed to improve the accuracy of IC diagnosis.

The objective: Our aim is to integrate two advanced molecular tests, ThyroSeq v 2.1 and miRNA signature into an intelligent, personalized system in order to increase the diagnostic accuracy of FNA cytology in indeterminate thyroid nodules. Additionally, we intend to find a molecular signature to help identify NIFTP preoperatively.

Design: A prospective observational clinical trial of around 200 subjects with thyroid nodules will be conducted in a tertiary centre for 2 years in the Romanian population.

Methods: Residual material from thyroid FNA samples will be used for prospective molecular analysis that includes testing for the ThyroSeQ panel of mutations using NGS technology and miRNA using microarray technology. The target genes will be therefore validation by qRT-PCR. The molecular profile of indeterminate cytology will be identified in concordance with the test results of the benign and malignant nodules. The experimental data together with imagistic and cytopathologic data will be then integrated into a Computer-aided diagnostic system. Machine learning models will be used to ensure optimal outcomes. It will be developed an easy-to-use interface for the introduction and interpretation of all the possible information and show a very precise answer regarding diagnosis, risk stratification, and patient management. This system will have an impact on reducing the rate of unnecessary thyroid surgery, re-interventions, complications and additional costs of medical care.