RESUMO
To investigate confounding factors of real-time ultrasound elastography (RTE) and to evaluate the diagnostic performance of ultrasound (US)-guided FNA for thyroid nodules with indeterminate elastography compared with conventional US. This study included 244 nodules with indeterminate elastography caused by several confounding factors (large or small size, deep location, isthmic or paratracheal location, calcification, thyroiditis, conflicting results between conventional US and RTE), and corresponding prevalences of malignancy were calculated. Additionally, conventional US and US-FNA data were collected and compared. The prevalences of malignancy of confounding factors were 74.1%, 75.0%, 73.3%, 46.2%, 27.3%, and 53.2%, respectively. Sonographic features (border, margin, echogenicity, echohomogeneity, and microcalcification) were significantly different between benign and malignant thyroid nodules (p < 0.05), and most of them exhibited good sensitivity but unsatisfactory specificity and accuracy. While US-FNA exhibited better performance with a sensitivity of 96.9%, a specificity of 99.1% and an accuracy of 98.0% in the diagnosis of malignancy. Given that indeterminate RTE is inevitable with a rather high malignant risk due to several confounding factors, our study revealed that US-FNA was a valuable tool in nodules with indeterminate elastography by increasing the detection rate of thyroid malignancy.