Diagnostic performance of elastography in cytologically indeterminate thyroid nodules

Cytological examination of material from fine-needle aspiration biopsy is the mainstay of diagnosis of thyroid nodules, thanks to its remarkable accuracy and scarcity of complications. However, follicular lesions (also called indeterminate lesions or Thy3 in the current classification), a heterogeneous group of lesions in which cytology is unable to give a definitive diagnosis to, represent its main limit. Elastography has been proposed as a potential diagnostic tool to define the risk of malignancy in the aforementioned nodules, but at present there is no conclusive data due to the small number of specifically addressed studies and the lack of concordance among them. The objective of our study was to evaluate the role of real-time elastography (RTE) for refining diagnosis of Thy3 nodules, by integrating diagnostic information provided by traditional ultrasound (US). The study included 108 patients with Thy3 nodules awaiting for surgery, which were evaluated by US (considering hypoecogenicity, irregular margins, microcalcifications, halo sign, and intranodular vascularization) and RTE. Nodules were classified at RTE using a four-class color scale. At histologic examination, 75 nodules were benign and 33 malignant. As expected, none of the ultrasound parameters alone was adequate in predicting malignancy or benignity of the nodules; in the presence of at least two US risk factors, we obtained 61 % sensitivity, 83 % specificity, and 77 % accuracy with 6.8 OR (95 % CI 2.4–20.4). RTE scores 3 and 4 showed 76 % sensitivity, 88 % specificity, 74 % PPV, and 89 % NPV with diagnostic accuracy of 84 %; the data are statistically significant ( p < 0.0001) with a OR of 21.9 (95 % CI 7.1–76). By combining RTE with US parameters, the presence of at least 2 characters of suspicion had 88 % sensitivity and 94 % NPV with 23.8 OR (95 % CI 7–106.3). The use of combined RTE and US leads to the identification of two patients subpopulations which have a significantly different malignancy risk (6 vs. 63 %); further studies are needed to verify if it is possible to send only the first group to thyroidectomy and the other to follow-up.