The utility of rapid on-site evaluation (ROSE) during endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) for lymph node staging in lung cancer is still controversial. The aim of this study was to assess the role of ROSE during EBUS-TBNA and the interpretation of its results. We performed a retrospective chart review of patients with suspected or diagnosed lung cancer who underwent EBUS-TBNA for lymph node staging. The slides were air-dried and Diff-Quik (American Scientific Products, McGaw Park, IL) staining was used for ROSE. Additional smears were prepared for Papanicolaou staining and any remaining sample was placed in 10% formalin for histologic evaluation. The results of ROSE were compared with the results of the final pathologic diagnosis. EBUS-TBNA was performed in 438 patients on 965 lymph nodes. Eighty-four lymph nodes (8.7%) were determined insufficient for definitive diagnosis by final cytologic evaluation. However 45 of the 84 lymph nodes were able to be diagnosed by histologic examination. The non-diagnostic sampling rate was 4.0%. There were no false-positive results on ROSE; however 25 cases (5.7%) were falsely evaluated as negative on ROSE. The concordance rate for staging between ROSE and final pathologic diagnosis was 94.3%. The sensitivity, specificity, negative predictive value, and diagnostic accuracy rate of EBUS-TBNA for correct lymph node staging was 96.5%, 100%, 89.8%, and 98.2%, respectively. ROSE during EBUS-TBNA for material adequacy showed a low rate of non-diagnostic sampling. There was a high agreement between the on-site and final pathologic evaluation during EBUS-TBNA; however immediate diagnosis should be approached with caution. Copyright © 2013 The Society of Thoracic Surgeons. Published by Elsevier Inc. All rights reserved.
