Endocrine Abstracts

Introduction: Fine-needle aspiration (FNA) is the most accurate diagnostic approach for determining thyroid nodule malignancy. Most nodules are benign, therefore, only suspicious ones require FNA. In 2017 American College of Radiology proposed a scoring system – Thyroid Imaging, Reporting and Data System (TI-RADS) for identifying clinically significant malignancies. Whether a nodule requires FNA depends on various criteria, one of which is the size. The aim of this study was to determine whether size is an important factor in deciding the necessity for FNA.

Methods: A total of 288 ultrasound images of patients with thyroid nodules were analysed. The nodules were scored, measured and assigned to one of five TI-RADS levels (TR): TR1 – benign, TR2 – not suspicious, TR3 – mildly suspicious, TR4 – moderately suspicious, TR5 – highly suspicious. The results were compared with histology findings.

Results: 219/288 (76%) benign and 69/288 (24%) malignant thyroid nodules were verified histologically. Nodules were distributed as follows: TR1 – 17 (5.9%), TR2 – 27 (9.4%), TR3 – 72 (25.0%), TR4 – 126 (43.8%), TR5 – 46 (16.0%). The mean size of measured nodules was 2.05±1.02 cm. In categories TR1 and TR2 100% of nodules were benign according to FNA. In TR3 68/72 (94.4%) of nodules were benign and 4/72 (5.6%) malignant, 38/72 (52.7%) <2.5 cm in size and 34/72 (47.22%) – ≥2.5 cm. None of the malignant nodules in TR3 were ≥2.5 cm. Negative correlation (r_s =−0.298, P=0.011) was found between size and malignancy. In TR4 93/126 (73.8%) of nodules were benign and 33/126 (26.2%) malignant, 59/126 (46.8%) <1.5 cm in size and 67/126 (53.2%) – ≥1.5 cm. Size of ≥1.5 cm had sensitivity of 39.39%, specificity of 41.94%, positive predictive value (PPV) of 19.40%, negative predictive value (NPV) of 66.10% and accuracy of 41.27%. No significant correlations between size and malignancy in TR4 were found. In TR5 14/46 (30.4%) of the nodules were benign and 32/46 (69.6%) malignant, 11/46 (23.9%) <1 cm in size and 35/46 (76.1%) – ≥1 cm. Size of ≥1 cm had sensitivity of 68.75%, specificity of 7.14%, PPV of 62.68%, NPV of 9.09% and accuracy of 50%. Negative correlation (r_s =–0.304, P=0.04) was found between size and malignancy.

Conclusion: Our study shows that thyroid nodule size threshold suggested in TI-RADS levels TR3 – TR5 is not reliable in predicting malignancy risk. Therefore, we think that nodule size is neither a good predictor of malignancy nor a good indicator for FNA.