Distinguishing benign from malignant thyroid nodules by using thyroid ultrasonography: utility of adding superb microvascular imaging

Objective Thyroid nodulesare very common, however malignancy are lesser than 7%. Therefore, additional techniques such as microvascular imaging is attempted to improve the diagnostic performance of conventional ultrasonography. We want to investigate the utility of additional superb microvascular imaging (SMI) to B-mode ultrasound in thyroid nodulesin distinguishing benign from malignant thyroid nodules. Methods We analyzed 36 thyroid nodules(malignant = 14, benign = 22) and a reviewer scored the likelihood of malignancy for three data sets before versus after the additional Color Doppler imageor SMI using 4-scale visual analysis (i.e., B-mode ultrasonography alone, B-mode ultrasonography + Color Doppler image, and B-mode ultrasonography + SMI). In addition, the SMI pixel count showing flow signal was measured in the region of interest including whole nodule on the longitudinal view. The area under the receiver operating characteristiccurve (Az) values, sensitivities, and specificities were compared. Results A comparison of the data sets showed that the AUC values were similar without a statistical difference in B-mode (0.924), B-mode + Color Doppler (0.860), and B-mode + SMI (0.909) (p > 0.05). However, the SMI pixel count was significantly higher in malignancy than in benign (19.38 vs. 10.58, p  =  0.04). The optimal SMI cutoff pixel count for predicting malignant thyroid nodulesobtained by receiver operating characteristiccurve was 14.25 (63.6% sensitivity, 78.6% specificity). Conclusions Additional use of SMI to conventional ultrasonography could be helpful to differentiate to malignant thyroid nodulesfrom benign ones.