Objective. To determine the role of radiological–pathological correlation in reducing false negative results of fine–needle aspiration (FNA) at thyroid nodules. Methods. This retrospective study included 307 nodules 1 cm or larger in 302 patients diagnosed as benign at initial cytologic evaluation and that underwent follow–up ultrasound (US) or FNA. We compared the risk of malignancy of nodules classified into subgroups according to the initial US features and imaging–cytology correlation. Results. Among included nodules, 7 nodules were proven to be malignant (2.3%) in follow–up FNA or surgery. When a thyroid nodule had discordant US findings on image review after having benign FNA results, malignancy rate increased to 23.3%, significantly higher than that of nodules with suspicious US features (P < 0.001). Conclusions. Repeat FNA can be effectively limited to patients with cytologically benign thyroid nodules showing discordance in imaging–cytology correlation after initial biopsy, which reduces unnecessary repeat aspirations.