Clinician-performed ultrasound in assessing potentially malignant thyroid nodules

Abstract

Background: Ultrasound (US) is used in the workup of thyroid nodules. Ultrasonographic characteristics, such as an ill-defined margin, hypoechoicity or fine calcifications, are known to be associated with malignant thyroid lesions. The association between these characteristics and the risk of malignancy has been reported predominantly from series published where US is performed in radiology departments. Clinician-performed ultrasound (CPU) is increasingly being used as a modality, although there is little published literature validating this practice. Method: A prospectively collected database of known ultrasonographic characteristics of malignancy as determined by CPU on thyroid nodules is reported and correlated against adequate cytology or operative histopathology. Results: In total, 157 thyroid nodules (28 malignant, 129 benign) were included and characteristics of poorly defined capsule (sensitivity 46%, specificity 91%), absence of halo (sensitivity 54%, specificity 80%), hypoechoicity (sensitivity 79%, specificity 54%), heterogeneity (sensitivity 64%, specificity 68%), fine calcifications (sensitivity 36%, specificity 95%) and central blood supply (sensitivity 71%, specificity 69%) were found to be associated with malignant thyroid nodules. Negative-predictive values (NPVs) for these characteristics were consistently high (89%, 89%, 92%, 90%, 87% and 94%, respectively). Discussion: These results are consistent with the previously published datasets of ultrasonographic characteristics of malignancy and validate the use of CPU. The consistently high NPV suggests that the absence of ultrasonographic characteristics of malignancy correlates well with benign lesions. CPU is a reliable and useful tool in the hands of surgeons assessing and following potentially malignancy thyroid nodules.