Rapid on-site pathologic evaluation does not increase the efficacy of endobronchial ultrasonographic biopsy for mediastinal staging.

BACKGROUND Endobronchial ultrasonography with transbronchial needle aspiration (EBUS-TBNA) has been shown to be equivalent to mediastinoscopy in lung cancer staging for mediastinal node involvement. Rapid on-site evaluation (ROSE) to determine the adequacy of nodal sampling has been claimed to be beneficial. METHODS A retrospective evaluation was performed in 170 patients who underwent EBUS-TBNA from July 2008 to May 2011. The patients were classified as having either high or low pretest probability for mediastinal disease based on history and radiographic imaging. ROSE was compared with the final pathology reports based on slides and cell blocks. RESULTS One hundred thirty-one (77%) patients were classified as being in the high pretest cohort based on clinical staging. Of these, 101 (77%) patients had adequate tissue sampling based on ROSE, with 70 (69%) patients having positive mediastinal disease. In the 30 (23%) patients who had inadequate tissue by ROSE, the final analysis of all the prepared slides and cell blocks allowed for a diagnosis in all but 8 patients. The sensitivity and specificity of ROSE in the high pretest probability cohort were 89.5% and 96.4%, respectively, whereas the overall sensitivity and specificity of EBUS-TBNA was 92.1% and 100%, respectively. Despite having inadequate tissue on ROSE in 30 of 131 patients, sufficient tissue was available on final analysis for diagnosis in 22 of 30 patients. CONCLUSIONS ROSE does not impact clinical decision making if a thorough mediastinal staging using EBUS is performed. Despite inadequate tissue sampling assessment by ROSE, a final diagnosis was made in most patients, potentially avoiding an additional surgical procedure to prove mediastinal disease.