Sonographic Features of Endobronchial Ultrasound in Differentiation of Benign Lymph Nodes

Abstract

Endobronchial ultrasound (EBUS) features have been found to be a useful tool in differentiating malignant from benign lymph nodes, but the use of these features to distinguish benign intrathoracic lymphadenopathies, including tuberculosis, sarcoidosis and reactive lymphadenitis, has not been established. The goal of this study was to evaluate the use of EBUS features in predicting tuberculosis, sarcoidosis, and non-specific inflammation. One hundred eighty-eight patients with suspected benign lymphadenopathy were included in the study. The EBUS features studied were short axis, shape, calcification, central hilar structure, necrosis sign, margins, echogenicity, clustered formation and vascular patterns. The sonographic findings were confirmed by clinicopathologic results. EBUS-Guided transbronchial needle aspiration was performed on 452 lymph nodes. Thirty-seven tuberculous nodes, 193 sarcoid nodes and 150 reactive nodes were retrospectively analyzed. Excluded were 72 nodes. Multivariate analysis revealed that presence of the necrosis sign and absence of the clustered formation are independent factors predictive of tuberculous nodes. Short axis >1 cm, absence of central hilar structure, distinct margins, presence of clustered formation and non-hilar perfusion were predictive of sarcoid nodes. Presence of central hilar structure, absence of clustered formation and vascular pattern (hilar perfusion or avascularity) were predictive of reactive lymphadenitis. The sum score model of these combined predictive factors indicated that the best diagnostic accuracies for predicting tuberculous nodes, sarcoid nodes and reactive lymphadenitis were 77.1%, 89.2% and 87.1%, respectively. Sonographic features could be helpful in differentiating the type of benign intrathoracic lymphadenopathy during EBUS examination.

Keywords: Endobronchial ultrasound; Inflammation; Lymph nodes; Sarcoidosis; Sonographic features; Tuberculosis.