Utility of on-site evaluation of endobronchial ultrasound-guided transbronchial needle aspiration specimens

Background: Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is an integral tool in the diagnosis and staging of malignant tumors of the lung. Rapid on-site evaluation (ROSE) of fine needle aspiration (FNA) samples has been advocated for as a guide for assessing the accuracy and adequacy of biopsy samples. Although ROSE has proven useful for numerous sites and procedures, few studies have specifically investigated its utility in the assessment of EBUS-TBNA specimens. The intention of this study was to explore the utility of ROSE for EBUS-TBNA specimens. Materials and Methods: The pathology files at our institution were searched for all EBUS-TBNA cases performed between January 2010 and June 2010. The data points included number of sites sampled per patient, location of site(s) sampled, on-site evaluation performed, preliminary on-site diagnosis rendered, final cytologic diagnosis, surgical pathology follow-up, cell blocks, and ancillary studies performed. Results: A total of 294 EBUS-TBNA specimens were reviewed and included in the study; 264 of 294 (90%) were lymph nodes and 30 of 294 (10%) were lung mass lesions. ROSE was performed for 140 of 294 (48%) specimens. The on-site and final diagnoses were concordant in 104 (74%) and discordant in 36 (26%) cases. Diagnostic specimens were obtained in 132 of 140 (94%) cases with on-site evaluation and 138 of 154 (90%) without on-site evaluation. The final cytologic diagnosis was malignant in 60 of 132 (45%) cases with ROSE and 46 of 138 (33%) cases without ROSE, and the final diagnosis was benign in 57 of 132 (47%) with ROSE and 82 of 138 (59%) without ROSE. A cell block was obtained in 129 of 140 (92%) cases with ROSE and 136 of 154 (88%) cases without ROSE. Conclusions: The data demonstrate no remarkable difference in diagnostic yield, the number of sites sampled per patient, or clinical decision making between specimens collected via EBUS-TBNA with or without ROSE. As a result, this study challenges the notion that ROSE is beneficial for the evaluation of EBUS-TBNA specimens.