Endobronchial ultrasound (EBUS) with transbronchial needle aspiration (TBNA) is becoming widely used for mediastinal lymph node staging in patients with known or suspected lung cancer. While a substantial number of case series have evaluated test performance of this investigation, the small sample sizes limited the ability to accurately evaluate the precision of EBUS-TBNA as a staging modality. A systematic review was performed of published studies evaluating EBUS-TBNA for mediastinal lymph node staging to ascertain the pooled sensitivity and specificity of this investigation. A literature search was constructed and performed by a professional medical librarian to identify the literature from 1960 to February 2008. Pooled specificity and sensitivity was estimated from the extracted data with an exact binomial rendition of the bivariate mixed-effects regression model. Of 365 publications, 25 were identified in which EBUS-TBNA was specifically focused on mediastinal node staging. Of these, only 10 had data suitable for extraction and analysis. The overall test performance was excellent with an area under the summary receiver operating characteristics curve of 0.99 (95% CI 0.96 to 1.00); similarly, EBUS-TBNA had excellent pooled specificity of 1.00 (95% CI 0.92 to 1.00) and good pooled sensitivity of 0.88 (95% CI 0.79 to 0.94). EBUS-TBNA has excellent overall test performance and specificity for mediastinal lymph node staging in patients with lung cancer. The results compare favourably with published results for computed tomography and positron emission tomography.
