Validation of Ultrasound Risk Stratification Systems for Cervical Lymph Node Metastasis in Patients with Thyroid Cancer

SIMPLE SUMMARY: Ultrasound (US) malignancy risk stratification systems (RSS) for cervical lymph nodes (LNs) have not been fully established in patients with thyroid cancer. In this study, we assessed the malignancy risks of each US feature and risk category from the Korean Society of Thyroid Radiology (KSThR) and the European Thyroid Association (ETA). Both systems effectively classified malignancy risks; however, 15.1% of LNs were unclassifiable in ETA RSS. Suspicious US features of hyperechogenicity, cystic change, echogenic foci, and abnormal vascularity were independently associated with metastasis. When the primary tumor characteristics were assessed, tumor multiplicity was associated with metastasis in the indeterminate LN group. We refined this system and proposed an RSS based on the KSThR system for cervical LNs in patients with thyroid cancer. ABSTRACT: A malignancy risk stratification system (RSS) for cervical lymph nodes (LNs) has not been fully established. This study aimed to validate the current RSS for the diagnosis of cervical LN metastasis in thyroid cancer. In total, 346 LNs from 282 consecutive patients between December 2006 and June 2015 were included. We determined the malignancy risk of each ultrasound (US) feature and performed univariable and multivariable logistic regression analyses. Each risk category from the Korean Society of Thyroid Radiology (KSThR) and the European Thyroid Association (ETA) was applied to calculate malignancy risks. The effects of size, number of suspicious features, and primary tumor characteristics were analyzed to refine the current RSS. Suspicious features including echogenic foci, cystic change, hyperechogenicity, and abnormal vascularity were independently predictive of malignancy (p ≤ 0.045). The malignancy risks of probably benign, indeterminate, and suspicious categories were 2.2–2.5%, 26.8–29.0%, and 85.8–87.4%, respectively, according to the KSThR and ETA criteria. According to the ETA criteria, 15.1% of LNs were unclassifiable. In indeterminate LNs, multiplicity of the primary tumor was significantly associated with malignancy (odds ratio, 6.53; p = 0.004). We refined the KSThR system and proposed a US RSS for LNs in patients with thyroid cancer.