Endocrine Abstracts

Introduction: Various thyroid imaging reporting and data systems (TIRADS) have been proposed in recent years, such as the EU-TIRADS proposed by the European Thyroid Association (ETA). These classification systems intend to estimate the malignancy risk of thyroid nodules, however they are very much based on suspicious ultrasound (US) features typical of papillary thyroid cancer. As such, follicular carcinomas may be reported as low suspicion nodules. Our work aims to evaluate the performance of the EU-TIRADS in follicular neoplasms and how accurate are their ultrasound high suspicion features in distinguishing follicular and Hürthle cell adenomas from carcinomas.

Methods: Retrospective study including patients followed-up in a Portuguese central hospital for the past 10 years who underwent to thyroid surgery and received subsequent histological diagnosis of follicular carcinoma, follicular adenoma, Hürthle cell carcinoma or Hürthle cell adenoma – 151 patients. We identified every patient who had US imagological records from before surgery available for review. The US features were retrospectively evaluated and the nodules were classified according to their EU-TIRADS score.

Results: We included 39 patients with histological diagnosis of follicular neoplasm (18 malignant) – seven follicular carcinomas, 16 follicular adenomas, 11 Hürthle cell carcinomas and five Hürthle cell adenomas. The mean age at surgery was 55±14 years and 74% of patients were female. The mean nodule diameter was 31 mm (adenomas: 29 mm and carcinomas: 34 mm, P=0.260). The presence of a hypoechoic halo was associated (P=0.034) with increased odds of benign nodule etiology. No statistically significant differences were identified regarding nodular composition, shape, margins or existence of microcalcifications between carcinomas and adenomas. Both adenomas and carcinomas had EU-TIRADS score of 3 (benign: 8, malignant: 4), 4 (benign: 6, malignant: 8) or 5 (benign: 7, malignant: 6) and this distribution did not reach a statistically significant difference (P=0.479). EU-TIRADS category four had the highest sensitivity for detecting malignant lesions – 44.4%. Most malignant (n=12) and benign (n=14) nodules did not present any EU-TIRADS high suspicion US features and the number of features was not statistically significantly different between carcinomas and adenomas (P=0.208). Only 1 malignant nodule did not have an EU-TIRADS score and diameter that would imply fine-needle aspiration (FNA) according to ETA EU-TIRADS guidelines.

Conclusion: EU-TIRADS based high suspicion US features do not seem valuable in distinguishing follicular adenomas from carcinomas. Nevertheless, the majority of carcinomas scored as EU-TIRADS category 4 and 5 – intermediate and high-risk categories which should warrant FNA in most cases.