Thyroid lesions categories iii and IV: malignancy detection by computer analysis of nuclear images

Abstract

Although thyroid cancer is the most common endocrine neoplasia, its incidence is reduced. The differentiated thyroid carcinoma (DTC) corresponds to 90% of all thyroid carcinomas and is mainly represented by papillary (PC) and follicular carcinomas (FC) subtypes. According to the Bethesda System, the fine-needle aspiration (FNA) of thyroid nodules may be specified in six classes. As the risk of malignancy in classes III and IV is still unclear, there are divergences about which conduct to be used. Therefore, the computer analysis of certain nuclear parameters of follicular cells may show a better identification of malignancy in these two classes. Objectives: To evaluate nuclear parameters, using computerized image analysis, in cytological material obtained from thyroid nodules cytologically diagnosed in categories III and IV, and relate them with the final histological diagnose. Methodology: It will be studied cytological material from thyroid nodules, obtained from FNA between 2009 and 2015 in the Botucatu Clinics Hospital, and diagnosed as III and IV categories according to Bethesda System. Follicular nuclei will be analyzed by a software named ImageJ, that will consider morphometric and textural parameters. Eventually, it will be observed if these nuclear parameters are able to predict the lesion malignancy, according to the final histological diagnose.