Endocrine Abstracts

Introduction: Thyroid nodules have a high prevalence in the adult population (20-76%). In 2018, the American College of Radiology proposed an ultrasound classification score of thyroid nodules - Thyroid Imaging Reporting and Data System-TIRADS, with 88% of sensitivity and 49% of specificity to predict malignancy. Since then, several studies report sensitivity values of 70.6-97% and specificity of 30-98%. We pretend to compared TIRADS score with cytologic results based on Bethesda classification.

Methods: This is a retrospective study of patients with thyroid nodules submitted to fine-needle aspiration (FNA) in an endocrinology department, during 16 months. Cytologic results (Bethesda classification) and TIRADS score (TR) were evaluated. TR score was calculated at the time ultrasound-guided FNA was performed. All Bethesda I (non-diagnostic) results were excluded.

Results: A total of 143 FNA of thyroid nodules were analysed, from 123 patients (57.7 % female) with a mean age of 64 ± 14.3 years-old. The nodules’ mean dimension was 24.6 ±9 mm. 61.5% of nodules were classified as TR 4, 27% TR3, 7% TR5, 2.1% TR1 and 1.4% TR2. 90.2% of FNA were Bethesda II (benign) and 9.8% were Bethesda III (follicular lesion of undetermined significance). No results of Bethesda IV-VI were diagnosed (malignancy). Analysing all TR1 and TR2 nodules, 80% were Bethesda II and 20% were Bethesda III. From nodules of TR3 to TR5, Bethesda II results were obtained in 92.3% of TR3, in 92% of TR4 and 72.7% of TR5, and Bethesda III were obtained in 7.7% of TR3, 8% of TR4 and 27.3% of TR5. No statistically significant difference was verified between categories (P=0.392). A 91.9% specificity was achieved for TIRADS score to identify benign thyroid nodules.

Conclusion: Our data demonstrated an elevated TIRADs score specificity that is in line with that reported in the published literature. Our results showed that TIRADs score seems to be particularly useful in identifying benign thyroid nodules. Higher risk TIRADS categories did not reflect higher risk for malignancy in cytologic diagnosis. More studies with greater samples are needed.