Diagnostic accuracy of high-resolution neck ultrasonography in the follow-up of differentiated thyroid cancer: a prospective study.

OBJECTIVE

To conduct a prospective study of the diagnostic value of features of cervical lymph nodes (large size, central location, abnormal shape, cystic changes, calcifications, and loss of echogenic hilum), assessed by neck ultrasonography (US), in patients scheduled for surgical treatment of persistent or recurrent differentiated thyroid cancer.

METHODS

We studied 152 US abnormalities in 42 patients (median age, 38.5 years) who had undergone one or more neck operations, with or without radioiodine therapy, but continued to have persistent or recurrent disease, which was confirmed by fine-needle aspiration. Another surgical procedure was planned for these patients. On the day of operation, patients underwent a detailed US neck examination by an experienced radiologist. US abnormalities were plotted on a standard diagram of the neck and given specific numbers to help track them during surgical intervention and histopathologic examinations. The US features were compared with the final histopathologic diagnosis.

RESULTS

Of 152 US abnormalities, 127 involved cervical lymph nodes and 25 involved other types of tissue. In univariate analysis, size, absent echogenic hilum, cystic changes, calcifications, and central location (medial to the sternomastoid muscle) of cervical lymph nodes were significantly associated with the presence of metastatic involvement. In multivariate analysis, only central location (odds ratio, 4.07; 95% confidence interval [CI], 1.64 to 10.10) and size (odds ratio, 5.14; 95% CI, 1.64 to 16.06) remained significant. The receiver operating characteristic curve for the size of lymph nodes showed a large area under the curve of 0.77 (95% CI, 0.68 to 0.85), and a size of 7.5 mm showed the highest sensitivity and specificity.

CONCLUSION

Size and central location of cervical lymph nodes assessed by US during follow-up of patients with differentiated thyroid cancer were the most important predictors of presence of metastatic disease.