Contrast-enhanced ultrasound–based nomogram for predicting malignant involvements among sonographically indeterminate/suspicious cervical lymph nodes in patients with differentiated thyroid carcinoma

Abstract

This prospective study was aimed at assessing the value of nomograms based on conventional and contrast-enhanced ultrasound (CEUS) features in the pre-operative diagnosis of sonographically indeterminate/suspicious lymph node metastasis (LNM) in patients with differentiated thyroid carcinoma (DTC). A total of 72 cervical LNs from 47 patients with DTC from January to June 2018 were included in the primary data set, and 30 LNs from 15 patients with DTC from July to August 2018 were included in the external validation data set. The LNs of the included patients were preoperatively evaluated by conventional ultrasound (US) and CEUS. Each included LN was labeled by puncture localization with carbon nanoparticle suspension injection (Canalin) under US guidance and dissected separately to ensure the one-to-one correspondence between ultrasonic features and pathology status. Univariate logistic regression analysis was used to identify risk factors for LNM. A nomogram was used to construct a prediction model for cervical metastatic LNs. Round shape, absence of hilar structure, peripheral or mixed blood flow and centripetal or mass enhancement were risk factors for lymph node metastases. The area under the receiver operating characteristic curve of the nomogram model based on conventional US and CEUS features was 0.93 (95% confidence interval: 0.872-0.985), which was superior to that of the nomogram based on conventional US features(0.85, 95% confidence interval: 0.707-0.989). CEUS features can provide incremental benefit in the diagnosis of LNM among DTC cohorts. Nomograms based on conventional US and CEUS features can predict LN status with high accuracy.