Rapid On-Site Evaluation in Detection of Granulomas in the Mediastinal Lymph Nodes.

ABSTRACT

RATIONALE Rapid On-Site Evaluation (ROSE) of specimens collected by endobronchial ultrasound (EBUS)-guided-transbronchial needle aspiration (TBNA) ensures sample adequacy and triages subsequent biopsy procedures. EBUS-TBNA allows sampling of lymph nodes in granulomatous diseases; however, the ability of ROSE to predict the final diagnosis in this setting has not been well characterized. OBJECTIVES: We performed a retrospective evaluation to study the utility of ROSE in the diagnosis of granulomatous diseases as well as to establish the procedure characteristics that would optimize the concordance between ROSE and final diagnosis. METHODS: Charts of patients with a cytological diagnosis of granuloma by EBUS-TBNA between June 2008 and May 2013 were reviewed. Preliminary ROSE findings and final cytological diagnosis were compared. Patient demographics and procedure variables were assessed using mean (±SD). The variables collected were considered in a logistic regression analysis using concordance as the outcome. MEASUREMENTS AND MAIN RESULTS: In our study, 255 procedures were performed to sample 625 lymph nodes that contained granulomas. An average of 2.4 (±1.2) lymph nodes were biopsied per procedure, with a mean size of 14.4 (±7.9) mm. The concordance between ROSE and the final diagnosis was 81.6%. The concordance rate was not impacted by needle size, lymph nodes size or station, number of stations biopsied, or passes per lymph node. The concordance did improve with the experience of the bronchoscopist (P < 0001). CONCLUSIONS: In this single-center study, there was a high concordance between ROSE and the final cytological diagnosis for mediastinal lymph nodes containing granulomas that were sampled by EBUS-TBNA. ROSE may serve to reduce procedure time, enhance sample triaging, and obviate the need for further invasive testing. The only variable associated with increased concordance was the experience of the operator.