MON-497 Evaluation and Comparison of Ultrasonographic Features According to Histology of Thyroid Nodules

Abstract Introduction: The prevalence of thyroid nodules identified by ultrasonography (US) may be up to 70%, with the majority being incidentalomas. Less than 15% of thyroid nodules will be malignant. Ultrasonography plays a key role in the diagnosis and management of thyroid diseases and US features have been described in way to predict the risk of malignancy. Aim: Evaluate and correlate US features with histological results, in our hospital, from January 2017 to December 2018. Materials and methods: Cross-sectional study. A total of 117 patients were evaluated, 77 had histology of nodular hyperplasia (benign) and 40 had histology of thyroid carcinoma (malign) after partial or total thyroidectomy. We described the size of the nodule (the largest diameter), the US features detailed in the ultrasound report (shape, margins, echogenicity, composition and microcalcifications), the EU-TIRADS score of each nodule and the histology of thyroid. The data were analysed using nonparametric statistical tests. The results are presented as median (interquartile range). The level of significance accepted was p<0.05. Results: Of the 117 patients, 82.1% were female, with a median age of 59.7 (25.0) years, nodule size in benign histology group was 35.0 (14.0) mm and in malign histology group was 20.0 (28.0) mm. Of malignant nodules, 90.0% (n=36) were papillary carcinomas. The prevalence of US features: 76.9% were solid, 35.0% had microcalcifications, 25.6% had irregular margins, 17.9% were markedly hypoechoic and 6.8% were taller-than-wide. Three characteristics were significantly related to malignancy: irregular margins (p=0.001), markedly hypoechoic (p=0.001) and microcalcifications (p=0.015). The taller-than-wide feature had the highest specificity of 94.8%; the marked hypoechogenicity with a specificity of 92.2%, had the highest positive likelihood ratio (LHR+: 4.81 (95% CI: 2.02–11.44)) and the highest diagnostic odds ratio (DOR: 7.10 (95% CI: 2.48–20.30)). The EU-TIRADS 5 category had a significant association with malignancy (p=0.001) and 80.0% of malignant nodules were classified in this category. The specificity was 62.3%, LHR+ was 2.12 (95% CI: 1.53–2.94) and DOR was 6.62 (95% CI: 2.69–16.31). Conclusion: In the evaluation of the US features, marked hypoechogenicity had the highest LHR+ and DOR. It was possible to verify that the EU-TIRADS 5 category was a good predictor of risk to malignancy, which reinforces the importance of categorizing the nodules for selective perform of fine needle aspiration.