Malignant risk stratification of thyroid FNA specimens with indeterminate cytology based on molecular testing.

BACKGROUND Fine-needle aspiration (FNA) has been employed for many years for examining thyroid nodules, and the cytology of aspirates is the primary determinant for whether thyroidectomy is indicated. Fifteen to thirty percent of thyroid nodules, not being clearly benign or malignant, fall into an indeterminate category. The main goals of molecular diagnostics for thyroid nodules are to prevent unnecessary surgery in patients with benign nodules and to stop patients with malignant nodules from being subjected to repeated operations. METHODS This study was designed to evaluate the diagnostic utility of 4 markers in thyroid FNA cytology via testing for the BRAF V600E mutation and the expression of microRNA-221, microRNA-222, and galectin-3 protein in FNA samples with indeterminate cytology. RESULTS A predictor model distinguishing benign samples from malignant samples on the basis of the 4 aforementioned markers was formulated. This decision model provided a sensitivity of 73.5%, a specificity of 89.8%, and a diagnostic accuracy of 75.7%. The positive predictive value was 80.6%, and the negative predictive value was 85.5%; this suggested that the prediction had good reliability. CONCLUSIONS One hundred twenty FNA samples were examined, and 62 nodules were classified as benign with the proposed diagnostic algorithm. This resulted in a reduction of the initial 120 patients to 58 and thus decreased by half the number of persons undergoing surgery. Cancer (Cancer Cytopathol) 2015;123:471–9. © 2015 American Cancer Society.